The concept of a ‘school’ that has been evolving in my mind as I pursue this work has something to do with the notion of a ‘distinct totality’: a group of Sangha who see themselves as in some sense distinct from other Sangha, and who view their own system as complete, adequate for a full spiritual life. This would involve a textual tradition, devotional centres, lineage of masters, institutional support, etc. When these factors are there to a sufficient degree for a particular portion of the Sangha to agree that they themselves constitute such a ‘distinct totality’, we can speak of a school.

Let us consider the main evidence for sectarian formation, dividing our sources into two groups, those before and those after the Common Era (about 400-500 an), and see where such a distinct totality can be observed. Within each group I shall consider the archeological evidence first, as that can clearly be fixed in time. The dates of all of the textual sources are questionable, and most of them probably straddle our divide. Nevertheless, I try to assign a place as best I can.


The Early Period (Before the Common Era)

Here our main sources are the archaeological evidence of the Aśokan inscriptions and the Vedisa stupas and inscriptions, the doxographical literature (Kathāvatthu and Vijñānakāya), and the Sinhalese Vinaya Commentary (which by its definite links with the archaeological evidence is proved to have roots in this period). We might also include the Aśoka legends which, while lacking such distinct archaeological confirmation as the Vinaya Commentary, nevertheless may have at least some origins in this period.

The Aśokan inscriptions do not mention any schools or any explicit occurrence of schism. When the edicts say the Sangha has been ‘made unified’, this suggests that there has been some conflict, but it falls short of establishing that a schism had occurred. In any case, even if there had been a schism, the edicts assert that it had been resolved. Nor do the Aśokan edicts mention any doctrines, texts, or anything else that might even hint at the existence of schools. The main sect-formative factor at work here would appear to be the geographical spread of the Sangha, which was to become a powerful force in the evolution of distinct sectarian identities.

The inscriptions on reliquaries retreived from the stupas in Vedisa mention several sectarian-formative factors, such as local saints, local institutions, and the name Hemavata, which at least at some time was taken to be the name of a school. But there is no clear and definitive evidence for the existence of a school. Hemavata may be purely a geographical term here. As Cousins observes, no unambiguous evidence for any Hemavata texts has survived, so the status of this school is doubtful in any case. The emergence of a local identity is a natural progression from the geographical spread under Aśoka, and we have no evidence that the Vedisa community saw itself as distinct from other Buddhist communities.

The doxographical literature likewise evidences sectarian-formative factors, particularly the articulation of controversial doctrines that characterized certain schools. But there is no explicit acknowledgement of the existence of schools, with the sole exception of the mention of the Puggalavāda in the Vijñānakāya.

The Sinhalese Vinaya Commentary was finalized much later, but there is definite archaeological evidence that proves the relevant portions must stem from genuine historical records. This is particularly true in the case of the Sudassanavinayavibhāsā, which was evidently taken to China and translated from a text predating Buddhaghosa’s revision of the commentaries in the the 5th Century ce. This details an extensive account of the period in question, and finds no reason to mention even in passing the existence of any schools.

Likewise the Aśokavadāna, Aśokarājasūtra, Divyavadāna, etc., give many elaborate stories of Aśoka without involving the schools. Of course these legendary works were much augmented over time, but if anything this strengthens our argument: since these texts were doubtless finalized in the sectarian period, there must have been a temptation to explicitly associate Aśoka with their own school. But this was not done, at least so far as I have seen.

Summing up this period, there is no evidence unambiguously belonging to the early period that mentions or implies the existence of schools. We find only the mention of various forces that lead to sectarian formation, never to the actual schools that resulted from these forces. This remains true even if we allow texts that are actually finalized later, but which probably have roots in this period.


The Middle Period (After the Common Era)

For this period our primary sources are the inscriptional evidence, the various schism accounts, and the śāstra/commentarial literature.

The inscriptions, starting in Mathura around 100 ce, regularly mention the names of schools.

The śāstras (e.g. Abhidharmakośa, etc.) and commentaries (e.g. Kathāvatthu-aṭṭhakathā, Mahāvibhāṣā, etc.) regularly mention schools by name, and discuss their doctrines. The textual sources agree fairly well with each other, and also with the inscriptions.

The schism accounts again mention similar names and sometimes similar doctrines as the other sources.

It is the schism accounts we must discuss in more detail, as they are the main sources from which the idea of an early schism was derived. The main four texts are closely related and must hark back to the same original in certain respects. But in the form we have them today they represent the perspectives of the four main groups of schools. Certain other lists are disregarded here (such as Bhavya I & II) but I believe they will not change matters significantly. These four main texts are:


  • Śāriputraparipṛcchā (Mahāsaṅghika)
  • Vasumitra’s Samayabhedoparacanacakra (Sarvāstivāda: this should be interpreted together with the Mahāvibhāṣā)
  • Dīpavaṁsa (Mahāvihāra/Vibhajjavāda/Sthavira)
  • Bhavya III (Puggalavāda)


These accounts can be further divided into two pairs by date. The Śāriputraparipṛcchā and Vasumitra are earlier, and probably date around 200 ce. The Dīpavaṁsa and Bhavya III are more like 400 ce (although the text of Bhavya III is later still, 600 ce +).

The Śāriputraparipṛcchā, which is the earliest or second-earliest of the schism accounts, stems from the Mahāsaṅghika. This account, which attributes the schism to an attempt on the part of the Sthaviras to expand the ancient Vinaya, dates the schism about a century after Aśoka. As we have seen, this is in perfect accord with all the inscriptural evidence, and with all the early textual evidence. It has been discounted by scholars who have asserted the text is corrupt and chronologically confused. However, a close examination of the text does not support this. The text is, admittedly, a poor and difficult translation, but the chronology of the period in question fits coherently into an overall narrative. The schism cannot be arbitrarily moved back before Aśoka without destroying this context. Indeed, one of the main purposes of the narrative is to claim for the Mahāsaṅghika school the mythic authority of Upagupta, a figure closely associated with Aśoka.

Vasumitra places the schism at the time of Aśoka, which for his short chronology is 100+ an. This version, which attributes the schism to a dispute on the ‘five points’ at Pāṭaliputta, is closely related to the Mahāvibhāṣā and Bhavya III. But we note that, while these three sources describe the same event, only Vasumitra connects this explicitly with Aśoka. Due to different ways of counting the years between the Buddha and Aśoka, the dating is hoplessly confused: Vasumitra places the events at Aśoka, which it says is 100+ an; Bhavya III places the same events before Aśoka, but the date is 137 an. The Mahāvibhāṣā does not name the king, so provides no support for any particular dating. In addition, the story, which is an outrageously polemical attack on ‘Mahādeva’, is only found in the larger and presumably later Mahāvibhāṣā, which dates at least half a millenium after the event. From the Mahāvibhāṣā we can see how the Sarvāstivāda school used these events to develop a distinctive mythos explaining how they came to be established in Kaśmīr. This would provide ample motivation for the Sarvāstivādins to associate the schism with Aśoka, regardless of any actual historical facts.

The Dīpavaṁsa was compiled shortly before Buddhaghosa, and is therefore significantly later than the Śāriputraparipṛcchā or Vasumitra. Dating 700 years after the events, it is the first text that claims that the schism was pre-Aśokan, placing it just after the Second Council in 100 an. The account of the schisms has been inserted from a Vasumitra-style text. However, the cause of the schism (textual corruption), the date, and the place (Vesālī) are all completely different. It has been crudely interpolated into a retelling of the story of the Councils otherwise preserved in the Sinhala Vinaya Commentaries. There is no need to assume that the original context of the interpolated schism account placed the events in this particular historical context; on the contrary, the setting is obviously incongruous. The Dīpavaṁsa’s dating of the schism just after the Second Council was probably an invention of the author(s) of the Dīpavaṁsa itself, whose aim was to establish an exclusivist mythos for the Mahāvihāra. The historical credibility of this account approaches zero.

Finally, like the Dīpavaṁsa, Bhavya III places the schism before Aśoka. But the events have nothing to do with the account of the Dīpavaṁsa. Rather it attributes the schism to the ‘five points’ as does Vasumitra, with dating inconsistencies as I mention above. The lack of mythic context makes this account harder to assess, but no doubt it was pressed into service to authorize the Puggalavāda school. We note that it is the two latest sources (Bhavya III and Dīpavaṁsa) that place the schism pre-Aśoka. It seems that the schism date is gradually getting earlier, a natural feature of the mythic process.

To summarize this period, then, we have consistent and clear evidence of the Buddhist schools dating from the middle period (post-ce). In all of our accounts of Buddhism of this period, the existence and basic nature of the schools is taken for granted and constitutes an essential component. The agreement of the sources as far as the names of the schools, their interrelationships, and their distinctive doctrines is, all things considered, reasonably high, as we would expect since they are describing contemporary conditions. But their accounts of the origins of the schisms, already in the far distant past from their own perspective, are a mass of contradictions. Of the three schism accounts that supply us with sufficient information (Śāriputraparipṛcchā, Vasumitra/Mahāvibhāṣā, Dīpavaṁsa), it is indisputable that the primary function of the accounts was not to record history but to authorize their own school. I believe this provides sufficient reason to explain how the schools came up with their various dating systems.

Of course, this does not prove that the dates in these texts are all wrong. It is quite possible and in fact very common to construct a mythology out of real events. But given the evident contradictions I think it is sheer naïvity to use the dates given in these texts to reach any simple historical conclusions. Like all myths, they are describing the situation in their own time (a situation of sectarian Buddhism) and backdating that in search of archaic authorization.


Comparing pre CE and post CE evidence

Despite the complexities of the situation, which any account including my own must inevitably distort by simplifying, the overall pattern is remarkably consistent. All the evidence of the early period (pre-ce) seems to be quite happy to talk about Buddhism with no mention of the schools. In stark contrast, in the middle period (post-ce) material the existence of the schools is inherent in how Buddhism is conceived. The textual and archaeological evidence is in good agreement here.

I conclude that various separative forces gathered momentum through the early period and manifested in the emergence of ‘schools’ towards the end of the early period, as depicted in the Śāriputraparipṛcchā (and various Chinese and Tibetan works). As the question of sectarian identity became more conscious, mythic accounts of the schisms emerged in the middle period.


The Mahāvihāravāsins

To find a more realistic description of how the schools may have arisen we shall have to look elsewhere. One of the fullest accounts of the origination of any school is found in the Sinhalese Vinaya Commentary, which exists in a Pali version the Samantapāsādikā, and an ancient Chinese translation the Sudassanavinayavibhāsā (T 1462善見律毘婆沙Shan-Jian-Lu-Pi-Po-Sha). The Sinhalese Vinaya Commentary recounts several decisive events that took place in the time of Aśoka. There was a conflict in the Sangha that was resolved by the expulsion of corrupt monks by Aśoka together with the Elder Moggaliputtatissa, following which the ‘Third Council’ was held to reaffirm communal identity. Subsequently Moggaliputtatissa organized the sending out of ‘missionaries’ to various parts of India. The main purpose of this narrative is to establish the credentials of the Sinhalese school.

Today we call this school ‘Theravāda’, but this name invites various forms of confusion. In particular it is a mistake to identify this school with the ‘Sthaviras’ who split from the Mahāsaṅghikas at the first schism. Rather, the Mahāvihāravāsins are just one branch of the Sthaviras who became established in Sri Lanka with their headquarters at the Mahāvihāra in Anuradhapura. In their own texts they refer to themselves as the Mahāvihāravāsins (‘Dwellers in the Great Monastery’) and I will adopt this term. It should be noted that when I refer to texts of this school this does not imply that the school necessarily created the texts in question; I simply mean the texts ‘as accepted by’ or ‘as passed down by’ the Mahāvihāra. In some cases these texts were authored by the school, but many of them are shared in common with other schools, with varying degrees of editorial differences.

There are two major pieces of inscriptional evidence that derive from the early period of Indian Buddhism: the Aśokan edicts and the reliquaries at Vedisa. Strikingly, both of these confirm the evidence found in Sinhalese Vinaya Commentary. The Vedisa inscriptions mention the names of several monks which the Sinhalese Vinaya Commentary says were sent as missionaries to the Himalaya soon after the ‘Third Council’. And Aśoka’s so-called ‘schism edicts’ (which actually state that the Sangha is unified, not schismatic!) mention an expulsion of corrupt bhikkhus, which many scholars have identified with the events prior to the ‘Third Council’. We should also note that Moggaliputtatissa’s sending out of missionaries has often been compared with Aśoka’s sending out of Dhamma-ministers; and that the Sri Lankan archaeological record is in general agreement with the picture of the missions. These two evidences, while not decisive, provide further points of agreement between the Sinhalese Vinaya Commentary and the archaeological record. This correspondence between epigraphic and textual evidence encourages us to take the missions account of the Sinhalese Vinaya Commentary seriously as a source for the origins of the schools.

The missions account describes how the Sinhalese school was established by Aśoka’s son Mahinda and his daughter the bhikkhuni Saṅghamittā. Several other teachers are described as being sent out to different places. While many of these missions cannot be confirmed, Frauwallner and others have shown that there is a general pattern of plausibility in the account.

In the current context of the revival of the bhikkhuni lineage in Theravāda, it is worth remembering the mission of Soṇa and Uttara to Suvaṇṇabhūmi, which is believed by Burmese to refer to Burma, and by Thais to refer to Thailand. This mission, which to this day forms a crucial narrative of self-identity for Buddhists in these regions, was said to result in the ordination of 1500 women. Thus bhikkhuni ordination is intrinsic to south-east Asian Buddhism from the beginning.


The Dharmaguptakas

One of the other missionaries was Yonaka Dhammarakkhita. He was, as his name indicates, a Greek monk, native of ‘Alasanda’ (Alexandria). One of the major figures in the missions narrative, he features in the Pali tradition as a master of psychic powers as well as an expert on Abhidhamma. He went to the Greek-occupied areas in the west of India. Long ago Pryzluski, followed by Frauwallner, suggested that Dhammarakkhita be identified with the founder of the Dharmaguptaka school, since dhammarakkhita and dhammagutta have identical meaning. Since that time two pieces of evidence have come to light that make this suggestion highly plausible. One is the positive identication of very early manuscripts belonging to the Dharmaguptakas in the Gandhāra region, exactly where we expect to find Yonaka Dhammarakkhita. The second is that the phonetic rendering of his name in the Sudassanavinayavibhāsā (the Chinese version of the Sinhalese Vinaya commentary) evidently renders ‘Dharmagutta’ rather than ‘Dhammarakkhita’. We also note that several texts say that the Dharmaguptaka was started by a certain ‘Moggallāna’. While this is traditionally identified with the great disciple of that name, I think it is more reasonable to see this as a reference to Moggaliputtatissa, the patriarch of the Third Council, who is also regarded by the Mahāvihāravāsins as their founder. We are thus perfectly justified as seeing the Mahāvihāravāsins and the Dharmaguptakas, not as warring schismatic parties, but as long-lost brothers parted only by the accidents of history and the tyranny of distance.


The Mūlasarvāstivādins

With regard to the third of our schools, the Mūlasarvāstivādins, the history is decidedly murky. In my opinion the most persuasive theory for the origin of this school was again provided by Frauwallner, who argued that they were originally based in Mathura. This would align this school closely with the famous arahants of Mathura: Śāṇavāsin and Upagupta. Śāṇavāsin features as a revered Elder and Vinaya master in the Vinaya accounts of the Second Council. He is said to have established a major forest monastery, which is called Urumuṇḍa in the northern sources and Ahogaṅga in the Pali.

Later on, it was to this very monastery that Moggaliputtatissa resorted for retreat. The spiritual power Moggaliputtatissa derived from his time in Śāṇavāsin’s forest monastery was decisive in convincing Aśoka to entrust him with the task of purifying the Saṅgha and organizing the missions. Thus the establishment of the Mahāvihāravāsin and Dharmaguptaka is closely associated with the Śāṇavāsin lineage. It is even possible that Soṇaka, the preceptor of Moggaliputtatissa’s preceptor, is simply a misspelling for Śāṇaka (-vāsin), in which case the Mahāvihāravāsin ordination lineage would be directly descended from Śāṇavāsin and the forest tradition of Mathura.

If Frauwallner’s theory of the distinct Mathuran origins of the Mūlasarvāstivāda school is found to be incorrect, then it would seem inevitable that we should seek the origins of this school as somehow related to the Sarvāstivādins of Kaśmīr. This school originated from one of the other Aśokan missionaries, Majjhantika. After serving as Mahinda’s ordination teacher in Pāṭaliputra, he went to Kaśmir and established the school later known as the Sarvāstivāda. This account, including the association of Majjhantika and Mahinda, agrees with the versions of the northern schools (except they generally place the date earlier).

In conclusion, we find that there is no evidence whatsoever of the origination of schools due to ‘schism’ in the narrowly-defined sense required by the Vinaya. The emergence of Buddhist monastic communities as ‘distinct totalities’ probably occurred gradually after the Aśokan period as a natural consequence of geographical dispersion and consequent differentiation. The accounts of conflicts that we possess today are more profitably read as mythic responses to events at the time the accounts were written, not as genuine histories.