11 December 2015

Pāḷi Ur-Text: Some Recent Observations by Alex Wynne

Dr Alexander Wynne
(Online pics are rare)
These are some notes from a talk by Alex Wynne given at the Oxford Centre for Buddhist Studies, along with some of my own comments. Wynne is currently Academic Head of the Dhammachai Tipiṭaka Project, at Wat Phra Dhammakaya, preparing a critical edition of the Pali canon.Wynne makes a number of interesting points that were news to me.

Wynne has been working as part of a team in Thailand since 2012, working on a critical edition of the Pāḷi Canon. In fact this will be the first ever critical edition of the Pali texts. A critical edition is more than simply another "version", but attempts to use existing manuscripts to reconstruct the original manuscript that all current copies were copied from. And

Those who dabble in Pāḷi might wonder why such a thing is needed. After all we have two widely available editions: the Pali Text Society (PTS) and the Chaṭṭhasaṅgāyana (6th Council) Burmese Ed. (1954-56), available as the Vipassana Research Institute CD/Online and as part of the Digital Pāḷi Reader (a plug in for Mozilla Firefox). The trouble is that while the Burmese Ed. compares with other editions but mainly used Burmese manuscripts. Neither the Burmese nor the PTS made extensive use of the Pāḷi manuscript tradition. The PTS Ed. is based on only a handful of manuscripts.Both made critical decisions, but did not tell us what they were.It turns out that from Sri Lanka, through Burma and across to Thailand (and places in-between) there are hundreds of manuscripts of the Pāḷi Canon that have largely been ignored when editing the Pāḷi Canon. The Dhammachai deals with five regional traditions of Pāḷi: Sinhalese, Burmese, Khom (Laos, Cambodia aka Khmer), Mon (Western Thailand), Tham (Laos, Northern Thailand).

These manuscripts are mainly written by scratching the surface of a prepared talipot palm leaf and rubbing ink into the scratches. The verb for writing in pāḷi, lipi is a loan word from Old Persian (via Sanskrit) from a verb meaning "to scratch". The scripts used are all related and derive from South India variants of Brahmī. Most of these palm leaf manuscripts are no older than the 18th century and are copies of copies over many generations. As such they are of varying quality. The oldest Pāḷi manuscript is from Burma. It was scratched into gold leaf and dates from around the 4th - 6th century (see Stargardt 1990 and below)

To produce a critical edition from manuscripts, one typically selects the best manuscript and uses it as the basis for the Ed., noting and repairing obvious errors (such as scribal misspellings, omissions and additions) and noting any variants amongst other manuscripts. To date there is no fully featured critical edition of the Pāḷi Canon. However, when there are 100's of manuscripts, one must be more selective. The project selects the 20 best manuscripts and uses them for the critical edition. This still requires that every one of the hundreds of manuscripts must be read and compared, but it eases the burden on the editor in the final stages. Some progress is being made. A first volume of the Dīghanikāya is finished and the rest of the Nikāya will be finished soon. The entire Canon will take a decades to complete. From what Wynne says many of the manuscripts have been digitised and will be made available online, along with the edition itself. This will be a great boon to serious Pāḷi studies.

One of the vicious tendencies of editors is to silently emend their text. Changes are often desirable, but all changes should be noted. As we will see below, this is because errors can supply us with a good deal of information. In the case of Pāḷi manuscripts there has been a tendency to correct the Pāḷi with Pāninian Sanskrit in mind, to regularise the forms. But Pāḷi, as a Middle-Indic language has a lot more flexibility regards certain aspects of grammar and there is no a priori reason to "correct" manuscripts that might reflect the nature of the imperfect language. The critical edition is making every effort to preserve the Middle-Indic features of the manuscript tradition. Sanskritisation is less obvious in manuscripts from Thailand compared with Burma and Sri Lanka where the Indian Sanskrit traditions were more influential.

The aim of the project is to reproduce the text that Buddhaghosa had in ca 5th Century. We know what Pāḷi looked like at this time from inscriptions all over South East Asia. Wynne says that trying to recreate the "original text" would be difficult, partly because we don't know what the language of that time looked like. Wynne mentions that Lance Cousins had distinguished between Classical Pāḷi (roughly what Buddhaghosa used) and Old Pāḷi which is the language of the discourses before being written down.

One benefit of scribal errors is that they become markers for the philologist interested in the provenance of manuscripts and relationships amongst them. A scribal error is a spelling error, which may include one or many wrong akṣaras (roughly "letters"), a diacritic added or left off, or changed word order. Scribes sometimes omit words, lines and even whole leaves. If a single manuscript has a mistake that none of the others share, then that mistake is just a localised problem caused by a particular scribe. If two manuscripts share exactly the same error in then the chances are that the scribes who produced them have both been copying a source manuscript which contains that error.

Gold Pali Manuscript ca 4th Century.
See Stargardt (1990)
Apart from being error prone, traditional manuscript copying practices are very conservative and there is a great reluctance to correct the texts. This is presumably related to their religious nature, but in any case it leaves a trail. One of the most interesting stories the scribal errors tell is that manuscripts copied in South-East Asia did not come from the same source as manuscripts in Sri Lanka. We can take this to mean that the Pāḷi texts in SE Asia were originally from Mainland India rather than Sri Lanka. Also the gold Pāḷi manuscript found in Burma and dated to the 4th Century is in an Indian Brahmī script rather than a Sri Lanka script, according to Harry Falk (See Stargardt 1990). Indeed at this point Wynne says a curious thing. There is no solid evidence for the Sri Lankan Theravāda outside of Sri Lanka, i.e. no explicit reference to them in any text or inscription, until the 11th Century CE. had I attended the lecture, I would like to have asked him to expand on this point.

Now as well as the Pāli texts we have a few smaller collections of texts in Chinese, and a selection of texts in Sanskrit translation and Gāndhārī. In many cases it is possible to compare the Chinese versions with the Pāḷi and there is now a steady stream of article which do just this (including one by yours truly). Wynne recounts that Anālayo, a German scholar/monk who has been at the forefront of this comparative effort, pointed out a scribal error in Pāḷi. At the beginning of the Brahmajāla Sutta (§1.5-6; PTS DN i.4 Walshe 1995: 68) there is statement about disparaging the three jewels. At §1.5 The Buddha exhorts the monks that if the Three Jewels are disparaged they should not become angry or resentful because this would be a hindrance. After all the monks can recognise true from false, and if they hear something false they should just say "this is false" (etaṃ abhūtaṃ). Now, at §1.6 is a mirror passage dealing with someone praising the Three Jewels and the monks not getting carried away with pleasure. The section in which the Buddha asks the monks if they can tell true from false in the case of praise is missing in Pāli. So while the mirror each other in every other respect, §1.5 and §1.6 are different in this, and it looks like a scribal omission (rather than an oral error).

What is really interesting is that this error exists in all known Pāḷi manuscripts from all over the world, but that the mirror passage is present in the Chinese (T21 I.264.b13). Wynne draws the obvious conclusion which is that all known Pāḷi manuscripts come from a common source. At some point there must have been one written source -- an ur-text -- from which all the surviving manuscripts are generational copies.

I would sound a note of caution here. Not all manuscripts survive. The talipot palm leaf is not a long lived medium in the tropics. It is subject to insect and microbial attacks of various kinds. That all surviving manuscripts can be traced to a single source text it does not mean that manuscripts which did not belong to this lineage did not once exist. There is always the possibility of the Black Swan Effect - it only takes one manuscript which does not have this shared error for the ur-text theory to fall down. Of course the writing down of the Canon probably did create a choke point in it's history. Previous to the first written text the situation would have been quite complex with different groups of monks preserving different collections and versions of the texts. Creating a written version required many editorial decisions which probably resulted in considerable lose of information. The fact of the differences in the Chinese collections suggests that a separate by related written edition was extensively edited in the North-West. On the other hand the Pāḷi collections were all edited with Sanskrit in mind in medieval times. It is also entirely possible that a Chinese scribe has "fixed" the text by filling in the gap.

A project of this kind takes a huge amount of time, resources, and organisation. It does not happen very often and once completed it is likely to become the definitive version of the Pāli Canon for generations to come. A great many of the differences will be minor, but there is also the possibility of major discoveries. It shows that as far as Buddhist texts goes, when one is studying an existing translation one is very far removed from the original text. We may never really know what the original texts looked like, since the Edition being produced is aiming to have a text such as Buddhaghosa might have read in the 5th century rather than an earlier era.

It's not related to the topic of Pāḷi texts, but in looking for pictures of Wynne I got drawn into look at the venue for the project: Wat Phra Dhammakaya. This is a newish monastery, founded in the 1970s on the basis of a new dispensation, and regularly attracts tens and hundreds of thousands of followers to activities. They teach that the Buddha equated nibbāna with attā (i.e. ātman), and that the goal of a Buddhist is to find their "true self" or dhammakāya (See Laohavanich 2012). Now this kind of thinking is quite well known to anyone interested in Indian religions - it is the view of the early Upaniṣads, especially the Bṛhadāraṇyaka. Such eternalistic thinking is explicitly criticised in any number of Buddhist discourses, though bares more than a passing resemblance to the eternalistic Tathāgatagarbha Doctrine. But not only are they explicitly teaching wrong-view as right-view, but the leader, Phra Dhammachayo, has been embroiled in a number of financial scandals (See e.g. Bangkok Post 30 Oct 2015). The organisation can afford to finance such a large and costly project as a critical edition of the Pāḷi Canon because they received billions of Baht in donations from their followers, though some of this money is allegedly being embezzled by the leader. I suspect that in years to come the project will be tainted by this association. Even the involvement of a bonafide scholar like Alex Wynne is not going to save the new edition from suspicion that it is aimed at furthering the idiosyncratic aims of the frankly weird Dhammakāya Movement.



Fernquest, Jon (2015) Phra Dhammachayo faces theft, money laundering charges.Bangkok Post, 30 Oct 2015.

Laohavanich, Mano Mettanando. (2012) Esoteric Teaching of Wat Phra Dhammakāya. Journal of Buddhist Ethics, 19, 483-513. 

Stargardt, Janice (1990). The Ancient Pyu of Burma: Early Pyu cities in a man-made landscape (illustrated ed.). PACSEA. https://www.academia.edu/7784094/The_Pyu_Civilisation_of_Myanmar_and_the_city_of_Sri_Ksetra
Related Posts with Thumbnails