Asoka, Pāli, and some red herrings.

A few weeks ago I read a reference to a book on Indo-Aryan languages. I found the book in the University library and was immediately struck by the entry on Pāli. Pāli, the book says, is essentially a western Middle Ind0-Aryan dialect. This raised a number of questions for me if only because the Buddha lived in the East of India. So what was going on?

The Indo-Aryan languages are divided roughly into three periods. Old Indo-Aryan (OIA) is represented by the Vedic language of the Ṛgveda, but includes Classical Sanskrit. The Middle Indo-Aryan (MIA) languages began to appear in inscriptions around the time of the Buddha - circa the 5th century BCE. These are often referred to as Prakrits and include Pāli. Although it's often assumed that all Indo-Aryan languages can be traced to Vedic, this is unlikely. Pāli and other Prakrits have antecedents in archaic OIA languages which existed in parallel to Vedic, but which are no longer extant. New Indo-Aryan (NIA) languages are the modern North Indian languages such as Hindi, Gujarati and Bengali.

The claim that Pāli is a western MIA dialect rests on its similarity to an Aśokan rock edict in Girnar (in modern day Gujurat). Girnar incidently was originally Girinagar or "Hill Town". This similarity between the Girnar edict and Pāli seems to have been originally noticed by Ettiene Lamotte. King Aśoka (c 272-231 BCE), whose empire centred on the town of Pāṭaliputra (present day Patna) had a number of these edicts carved into rocks around India, and also made some impressive stone pillars which carried inscriptions. The inscriptions cover a range of subjects including Aśoka's renunciation of violence, and his desire that everyone should live by the "dhamma". History tells us that Aśoka's grandfather, Chandragupta, had become a Jain ascetic late in life and it seems as though Aśoka might have emulated him in becoming a Buddhist.

The edicts are carved in a number of languages. In Bactria and Afghanistan they are in Greek or Aramaic because these were the local languages - legend says that Chandragupta Maurya was inspired in part by glimpsing Alexander on his rampage through Asia and into India, and it was Alexander who left Greek speaking people behind in that region. Aramaic was a relic of the Persians whom Alexander defeated. The other edicts are in forms of MIA. In Gāndhāra, the north-west of what is now Pakistan, the local dialect is now known as Gāndhārī. In the west another unnamed dialect was used. Finally in the east and south edicts were carved in what has come to be called Māgadhī - the language of Māgadha, and presumably of Aśoka himself.

It is usually said that Aśoka himself sent missions to far flung places to spread Buddhism, and especially that he sent one of his sons, Mahinda, down to Sri Lanka. The connection with Sri Lanka is very important because it was there that the texts were written down in Pāli. So at first glimpse it seemed that there was a puzzle here: why would Aśoka send missionaries bearing a collection of texts in a strange dialect? Why weren't the texts in Māgadhī? Theravāda legend says that Pāli is Māgadhī, but this idea has not been taken seriously by scholars for some decades.

In 1978 the great philologist K. R. Norman wrote:

"One thing is certain. The Pāli canon is not in the language of Aśoka's own capital; for the basic features of Māgadhī (s for all three sibilants, la for ra, and nominative singular in -e) are either lacking completely, or merely sporadic. (p.30)

The view finds authoritative support in Salomon's Indian Epigraphy where he reviews the writing in the Aśokan edicts. So until quite recently there seemed to be a consensus that Pāli was not Māgadhī and that it most closely resembles the language of the Girnar edict which is in the west, in what is present day Gujarat. So I began to construct a scenario which would account for this. Firstly I wondered whether the texts were in Pāli when they arrived in Sri Lanka - there is no account of their having been translated however. Next I discovered the traditional account in the Mahāvaṃsa, chapter thirteen of which tells that Asoka spent time in Avanti - a western province - where he was governor before taking up the throne in Māgadha. Here he had his son Mahinda with a local woman. Could this account for the language? I later discovered that Mahinda's origins in Avanti had already been considered as an answer to the language problem by Ettiene Lamotte (this is cited in one of K.R. Norman's many articles on the subject).

I also began to explore the history of the time. Asoka was not the last emperor of the Mauryan dynasty, but it did not long survive his death. And inscriptions in Māgadhī cease around this time. The Mauryan dynasty continued for a few decades until the last Mauryan king was assassinated by one of his generals who then founded the Śunga Dynasty. The Śungas are obscure but thought to be Brahmins who came from none other than the Avanti area in western India. It began to seem likely that the Śungas might have spread a western dialect into the East, and helped it to become the lingua franca in the last two centuries BCE. In fact the western dialect is found as far afield as Orissa and Amarāvati.

This seemed quite a likely answer to the question until I received replies to queries which I had sent to Dr Lance Cousins, and Prof. Richard Gombrich. Both said that the scholarly consensus is that Pāli is in fact Māgadhī, albeit heavily Sanskritised, and that Girnar also is best explained as a Sanskritised and partially localised version of Māgadhī. They both recommended reading more of K. R. Norman's recent work on this subject. This is no great hardship as Norman is a fascinating writer.

What Prof. Norman did was to compare in minute detail the language, spelling and writing of the edicts and construct a scenario which could account for the similarities and differences. Norman points out that the features of the Girnar inscription were not confined to the west, and that in fact there are many differences between the Girnar inscription and Pāli. In spoken language there was a general movement away from Sanskrit, with each local dialect or Prakrit changing at different paces. Different features of dialects could change at different rates as well. But at the same time Sanskrit was a powerful literary influence so that written texts were Sanskritised, or made to conform more to Classical Sanskrit. Asoka's edicts were written texts. Most likely they were dictated to a scribe who then made copies to be sent out to the regions to be carved into stone. Some edicts are clearly made from the same original, but show considerable variation in the skill and attention to spelling of the scribe or stone mason. Sometimes a local scribe has modified the original to be more Sanskrit, or more like the local dialect, and sometimes both. Sometimes he was just kak-handed, and sometimes working in a confined space that distorted the letters.

So by 1994 Norman's opinion had changed:

"It is clear that any conclusions about the origin of Pāli which are based upon the regional geography of the Aśokan inscriptions must be examined very carefully, and can be disregarded is they conflict with other evidence." and "If we disregard the evidence of the Aśokan inscriptions and assume that the situation in the second century reflects the pattern of dialect distribution in the third century, then it would be open to us to conclude that the dialect upon which Pāli was based had its home much nearer Magadha. " (p.8)

So it seems that the mystery is solved and ironically that the Theravāda legends were right all the time. Pāli is Māgadhī. Although over the years the written form of Pāli has been Sanskritised. This was partly under the influence of Sanskrit inspired Pāli grammar texts from Burma. We do know that Pāli continued to evolve because Buddhaghosa's Pāli is noticeably different from the language of the suttas.

A further red herring is revealed in Norman's 2004 paper on the Aśokan envoys. It's generally stated that Aśoka sent missionaries to convert people to Buddhism. Norman shows however that although Aśoka did send missions, they were not Buddhist missions but political missions. Aśoka was trying to convince people to give up war, not convert to Buddhism. Buddhist missions were sent, but these mainly went to different places, and sent different kinds of people, namely bhikkhus. And they were sent by Moggaliputta Tissa, not by Aśoka. Norman says:

There is every reason for thinking that the Buddhists did not even know of Aśoka's dhamma and his dūtas [envoys], otherwise we might have expected them to mention Aśoka in connection with the thera [ie religious] missions, in the same way that they mention him in connection with the third saṅgitī [i.e. council]." (2004: p.196)

We seem to have conflated two stories here which when one reads the evidence closely are distinct. Why? Norman speculates that it is because scholars have not read the edicts themselves and so are lead by other writers (who themselves may not have read them). Unless we make the effort to learn the languages then there is no way for us to know for ourselves!

This whole subject is a fascinating case study in the dynamics of scholarship of this type. This has all happened within my lifetime. It shows that just because something is in a book, even what appears to be an authoritative book (such as an impressive university text on languages), and backed up by other authoritative sources, it must still be evaluated. It is possible even for careful and talented scholars, familiar with the languages of the texts, to be in error - how much more so for those with no Pāli, Sanskrit or Prakrit (etc) relying on others to interpret texts for them. It shows how difficult it can be to be well informed on a subject, although writing in 2003 I wonder if Cardona and Jain can really be excused for not being cognisant of Norman's articles. The Pali Text Society have conveniently collected Norman's articles into eight hard bound volumes but these are not likely to be in your local library. Many of the articles are quite readable, but it's better to have some background in Indian history, geography, and languages. Not many of us have the time or inclination to do what needs to be done, so we have a tendency to propagate urban myths and hearsay.

• Cardona , George and Jain, Dhanesh (eds). 2003. The Indo-Aryan Languages. London : Routledge.
• Norman, K. R. - in Collected Works (dates are the original publication date):
  • Aśokan envoys and Buddhist Missionaries. 2004. (Vol VIII, p.183)
    
  • The Aśokan inscriptions and the Prakrit dialect geography. 1994. (Vol. VI, p.1)
    
  • The Language in which the Buddha taught. (vol.II, p.84):
  • The Role of Pāli in early Sinhalese Buddhism. 1978. (Vol II, p.30)
    
  • The Dialects in which the Buddha preached. 1991. (vol.II, p.128):
  • The Origin of Pali and its place among the Indo-European Languages. 1992. (vol.III, p.225):
• Salomon, Richard. 1998. Indian epigraphy : a guide to the study of inscriptions in Sanskrit, Prakrit, and the other Indo-Aryan languages. Oxford University Press.

Did King Ajātasattu Confess to the Buddha, and did the Buddha Forgive Him?

In lieu of a blog post this week I would like to draw your attention to my recent publication in the Journal of Buddhist Ethics: Did King Ajātasattu Confess to the Buddha, and did the Buddha Forgive Him? In this article I explore a passage from the Sāmaññphala Sutta (DN 2). This is the well known account of the meeting between King Ajātasattu and the Buddha. At the end of the discourse the king becomes a lay disciple and then confesses to the Buddha that he has killed his father, King Bimbisāra. The article begins with a translation problem. Although the action clearly describes a confession, the word often translated as confession - paṭikaroti - means no such thing. In trying to establish what the word does mean I looked at every occurrence in the suttas, and the Pāli commentaries on them, showing that the text uses a stock phrase (or pericope) which is employed in many different settings. This highlighted a feature of Buddhist confession which is distinct from religious confession in the west - that it does not involve reparation or making amends (despite what the translators say!).

In order to better understand what is happening I locate the action in the context of the early Buddhist theory of karma, and in the broader religious context of the day. The latter was deeply concerned with ritual purity, and, having been polluted with the return to ritual purity. The Buddha reinterpreted ritual purity as ethical purity, and confession in early Buddhism is a way of returning to ethical purity. The results of karma cannot be avoided, hence there is no reparation, no requirement to make amends in the confession. However through spiritual practice - including ethical purity - one can avoid creating new karmic results (kamma-vipaka), but crucially one can also reduce the impact of karmic consequences. I believe this is because we become more emotionally robust through spiritual practice, and that we are more able to contain painful vedanā (experience, sensation, feeling). To put it another way we are less likely to be blown off course by the worldly winds. The king however is doomed to rebirth in hell because patricide is an "unforgivable" offence. In fact this fate is undone in later version of the story which are preserved in Chinese translations of the sutta and a Sanskrit frgament. Here the charisma of the Buddha is such that it help Ajātasattu escape his fate. This change is one that deserves more attention but I don't speak Chinese!

Having established what the story is telling us I revisit the phrase 'yathadhamma paṭikaroti' around which the action hinges. I have shown by this point that previous translators (T W R Rhys Davids, Maurice Walsh) have misunderstood this term, and that the Pali-English Dictionary has also misunderstood it. There is no sense of "making amends" in any of the suttas which use this phrase, only of returning to ethical purity. In fact the phrase is difficult to translate into English and I have tentatively suggested that "Dharmically counteract" at least accurately renders the sense of the Pāli. Being an unattractive phrase it is unlikely to catch on, but I couldn't think of anything better.

A subsidiary issue arises in that some translators (Piya Tan, Ñāṇamoli) have understood Ajātasattu to be asking forgiveness and the Buddha to be offering forgiveness. I show that this does not make sense in the context, and it does not make sense in terms of Buddhist doctrine. The king is merely asking the Buddha to acknowledge his resolution to be ethical in the future, and the Buddha acknowledges the intention as an intention. Nothing more.

In early Buddhism confession is mostly associated with the bhikkhu sangha, but as my article shows confession clearly was part of a more general religious landscape with laypeople and even non-Buddhists confessing evil actions. One minor point which I make is that it is the actions which are done foolishly, in confusion, and unskillfully (yathābālaṃ yathāmūḷhaṃ yathā-akusalaṃ). Most translators change the adverb into an adjective describing the person rather than the action. This is consistent with Judeo-Christian ideas of culpability, but not with Buddhist views.

I hope this little precis will encourage people to read the whole article. Those with no Pāli at all may find the first couple of pages a bit daunting, but it soon settles down to discussing the implications, so don't be discouraged!


P.S. some of the ideas that emerged while researching this article have already appeared in blog posts: follow the link to other blog posts on confession.