7. Vibhajjavada vs. Sarvastivada?



In non-Pali sources, vibhajjavādin is sometimes contrasted with sarvāstivādin. Cousins makes it clear that he sees Sarvāstivāda as distinct from vibhajjavāda, but does not explain why.[1] It is problematic to assume that the Mahāvihāra tradition meant to imply this contrast, since it is not found in the Pali sources.

Indeed, Cousins’ article occasionally hints at the problems when he tries to imply a distinction between the Sarvāstivāda and the Vibhajjavādin schools in the narrow sense (Dharmaguptakas, Kaśyapīyas, Mahīśāsakas, Mahāvihāravāsins). For example, he remarks that the Abhidhamma-piṭaka of the Pali school is distinct, but ‘no doubt closely related to the Abhidhamma literature of other Vibhajjavādin schools’.[2] This is true, but slightly obscures the situation. Frauwallner has shown decisively that the Pali Abhidhamma Vibhaṅga is very closely related to the Sarvāstivādin Dharmaskandha. Both of these are also connected with the Dharmaguptaka’s Śāriputrābhidharma, but it seems, somewhat more distantly, at least in the form of the works, if not the doctrinal content. So, yes, the Vibhajjavādins probably had closely related Abhidhammas, but no closer than the Sarvāstivādins (with the probable exception of the Jñānaprasthāna).

Similarly, Cousins argues that the epigraphic evidence supports the idea that the Vibhajjavādins were the main missionary schools. But of course the Sarvāstivādins are well-attested in the north-west, and the lack of inscriptions to the south merely confirms the mission account that the Sarvāstivādin patriarch Majjhantika went to Kaśmīr.

Classic and influential contexts for the view that vibhajjavāda is specifically meant to contrast with sarvāstivāda include Vasubandhu’s Abhidharmakośa,[3] and part of the explanation for the sectarian list of Bhavya I. We should not remember that this explanation is expanding on the basic list of schools in Bhavya I; but in that list vibhajjavāda is a synonym of sarvāstivāda. Such inconsistency within a single section of a text should warn us against expecting consistency across the vast schools, lands, and times of ancient Buddhism. Here I quote from Bhavya in Rockhill’s translation:


‘Those who say that all exists - the past, the future, and the present - are called in consequence “They who say that all exists” or Sarvāstivādins.

‘Those who say that some things exist, (such as) past actions of which the result has not matured, and that some do not exist, (such as) those deeds of which the consequences have occurred and the things of the future; making categories (or divisions), they are called in consequence “They who speak of divisions” or Vibhajjavādins.’[4]


Now this view is discussed in the Kathāvatthu itself, where the opponent says that some of the past and future exists and some does not.[5] The commentary ascribes the heretical view to the Kassapīyas, who are generally held to be one of the vibhajjavādin schools (although from the Mahāvihāravāsin viewpoint the Kaśyapīyas were descended from the Sarvāstivādins). Vasumitra agrees in ascribing such a view to the Kaśyapīyas.[6] In any case, the view in question is specifically refuted by the Mahāvihāravāsins, in the book which was supposed by them to be written by Moggaliputtatissa at the very same Third Council we are considering.

It is unsurprising that the northern texts would have referred to a usage of vibhajjavādin that was actually followed by a school represented in the north, rather than the remote Sinhalese. But there is no particular reason to think that these passages refer to a clearly defined school; in fact such a view may well have been held by different groups or individuals. Rather, it seems that the northern sources treat at least one meaning of vibhajjavāda as being a doctrine specifically opposed to the sarvāstivāda doctrine. The Mahāvihāravāsin sources, however, never use the term in that way, nor do they actually hold the view that is ascribed to the vibhajjavādins in those contexts.

This is not the only case where the northern sources attribute views to the vibhajjavādins that differ from the Mahāvihāravāsin perspective. The Vibhāṣā discusses the view that time is eternal, while conditioned dhammas are not eternal; conditioned dhammas migrate like fruits being taken out of one basket and placed in another.[7] This view is ascribed to the Dārṣṭāntikas and Vibhajjavādins, but is not a position held by the Mahāvihāravāsins.

Of course, there may well be other contexts where the northern sources describe vibhajjavādin views that are in fact held by the Mahāvihāravāsins. But we must clearly differentiate between how the term vibhajjavādin is used in the different sources.

We saw above that in describing the use of vibhajjavādin, the later Pali sources do speak of doctrines that are held by Sarvāstivādins, but other schools may well have held such views as well, and the Sarvāstivādins’ main tenet is not mentioned. Such contexts are clearly aimed at other Buddhist schools in general and do not specifically define vibhajjavāda as an alternative to the Sarvāstivādin theory of existence in the three times. In other words, there is no reason to think that in using the term vibhajjavādin, the Mahāvihāravāsins meant to distinguish themselves from the Sarvāstivādins in particular.


The Early Controversies

This conclusion is reinforced by examining the doctrinal sources for the discussion of the Sarvāstivāda controversy. This is found in two early canonical Abhidhamma works, the already mentioned Kathāvatthu of the Mahāvihāravāsins, and the Vijñānakāya of the Sarvāstivādin Devaśarman.

The Kathāvatthu is ascribed by the Mahāvihāravāsins to Moggaliputtatissa, said by the commentaries to have been composed at the Third Council. The work as a whole cannot have been composed at that time, for it is the outcome of a long period of elaboration, and discusses many views of schools that did not emerge until long after the time of Aśoka. In addition, we have seen that the ascription of the work to Moggaliputtatissa at the Third Council is likely to be a late Mahāvihāra modification.

Nevertheless, there is no reason why the core of the book should not have been started in Aśoka’s time, and indeed K. R. Norman has shown that particularly the early chapters have a fair number of Magadhin grammatical forms, which are suggestive of an Aśokan provenance. In addition, the place names mentioned in the text are consistent with such an early dating.[8] So it is possible that the main arguments concerning the important doctrinal issues, which tend to be gathered at the start of the book, were developed by Moggaliputtatissa and the work was elaborated later.

Strong supporting evidence for this comes from the Vijñānakāya. This work starts off with extensive discussions, not of hundreds of points like the Kathāvatthu, but just two: the thesis that all exists, and the thesis of the ‘person’. The Sarvāstivādins agreed with the Mahāvihāravāsins that there was no ‘person’ in the ultimate sense, so their refutations of the views of the Puggalavādins share much in common. But on the proposition that ‘all exists’ they held opposing views. Whereas for the Mahāvihāravāsins this view was the sixth view discussed, the Sarvāstivādins made it number one.

The first chapter of their work is titled ‘Moggallāna section.’[9] This is a debate with a monk who in the title is called目乾連 (mu-gan-lian), and in the body of the text is called 沙門目連 (sha-men mu-lian = Samaṇa Moggallāna). Given the closeness of the two discussions of the ‘person’, and given that Moggaliputtatissa is said by texts of both schools to have discussed this view, there seems little doubt that this is referring to the same Elder.[10]

The Vijñānakāya discussion is simpler than the Kathāvatthu. Each paragraph begins with Moggallāna repeating his thesis: ‘The past and future are not; the present and the unconditioned exist.’[11]The straightforwardness of this view agrees with the Kathāvatthu and disagrees with the compromise position ascribed to the Vibhajjavādins by Bhavya and Vasubandhu (as discussed earlier). Moggallāna, unfortunately, does not get much of a chance to defend his thesis, but is simply countered with a barrage of arguments based on Sutta quotes. The basic form of the argument is that in order to abandon, say, greed, one must directly ‘see’ it with the mind. But the seeing of the greed must be distinct from the greed itself. One therefore must be ‘seeing’ past occasions of greed. But one can only ‘see’ what really exists. Hence the past exists.[12]

Strangely, while every paragraph repeats this phrase, after eleven repetitions we find a different thesis, with no explanation for this difference. The remaining eight paragraphs of this section return to the original thesis, again with no explanation. The aberrant thesis is  有無所心[13] which appears to be equivalent to the Pali: atthi anārammaṇaṁ cittaṁ (there is mind with no object). This rather cryptic phrase seems incongruous, as it appears to have nothing to do with the question of existence in the three periods of time. But in fact it clearly partakes in the basic abhidhamma debates: for example, the threes of the Dhammasaṅgaṇī mātikā include ‘dhammas with past object, dhammas with future object, dhammas with present object…’.

Related issues are discussed in several places in the Kathāvatthu, but the most relevant appears to be the heretical assertion that: atītārammaṇaṁ cittaṁ anārammaṇanti (mind with past object is without object).[14] This, while seemingly self-contradictory, addresses an important question: if the past and the future do not exist, what are we aware of when recollecting the past or predicting the future? Given that the non-Sarvāstivādin schools denied the existence of the past and future, they must come up with another account of this. Thus this assertion, given that it appears right in the middle of a debate on the three periods of time, would seem to be addressing the question of what the object of consciousness is when we think of the past and the future.

The view in question is ascribed by the commentary to the Uttarapāthakas, an obscure group known to no other text: it seems to be used as a generic term for the northern schools (literally ‘Norwegians’!), although here it must exclude the Sarvāstivādins. It may well include the Kaśyapīyas and the Dharmaguptakas, who are well-attested in the North-west. The view of the vibhajjavādins/Kaśyapīyas that part of the past exists would seem to be related. We bear in mind that, if the account of the missions is to be trusted, all these schools may claim Moggaliputtatissa as a founding teacher.

It is clear that the Vijñānakāya and the Kathāvatthu are ascribing two opposing views to Moggaliputtatissa. Given that the Kathāvatthu is vastly more developed than the Vijñānakāya – this is the 86th view it discusses – and given that only the Vijñānakāya directly attributes this view to Moggaliputtatissa (in the Pali this attribution comes in the commentaries), we might be inclined to trust the Vijñānakāya here. On the other hand, the Sarvāstivādins may have succumbed to the temptation to denigrate their opponents by ascribing to them inconsistent views, attributing to the founder of the school views that were later held by the ‘Uttarapāthakas’, in which case the Kathāvatthu might be more reliable. Other possibilities remain: perhaps Moggaliputtatissa argued for both views on different occasions; or perhaps he held neither. In any case, the two texts agree that Moggaliputtatissa was involved in these discussions, and the difference is in the details of how to work out a successful psychology based on the anti-Sarvāstivāda views, rather than the basic position.

But the most important point for our current purpose is that neither the Vijñānakāya nor the Kathāvatthu with its commentary use the term vibhajjavādin in discussion of this issue. For these texts, the term vibhajjavāda has nothing to do with the debate on the three periods of time.


What schism?

While it is clear that there was debate and disagreement on this issue, it is not at all clear that this had reached a sectarian split at this time. The Kathāvatthu throughout discusses doctrines only, and refrains from referring to specific individuals or schools. Only in its commentary do we find the identification of various views with particular schools. Reading just the Kathāvatthu itself, we couldn’t say whether the discussions were between different schools or merely an ongoing debate among one community. Of course, the lack of reference to specifics of place and time is entirely characteristic of the Pali Abhidhamma, and perhaps we should not read anything into it.

But a similar process is at work in the Vijñānakāya. The first debate, on ‘all exists’, is directed against an individual, Moggallāna. The second debate, on the ‘person’, is directed against a school, the  Puggalavāda.[15] Again, reading straight off the surface of the text, the debate with Moggallāna was a discussion with an individual, while the second topic was a debate between schools. This would be entirely in concordance with a situation where the Puggalavāda schism had already become manifest, so that the followers of that thesis were regarded as a distinct branch of Buddhism, while the Sarvāstivāda schism was still taking shape, still a debate among people who felt they belonged to the same school.



[1] Cousins, On the Vibhajjavādins, 132

[2] Cousins, On the Vibhajjavādins, 166

[3] 若自謂是說一切有宗決定應許實有去來世。以說三世皆定實有故。許是說一切有宗。謂若有人說三世實有。方許彼是說一切有宗。若人唯說有現在世及過去世未與果業。說無未來及過去世已與果業。彼可許為分別說部 (CBETA, T29, no. 1558, p. 104, b22-27)

[4] Rockhill, 184

[5] Kathāvatthu, pg. 151

[6] 其飲光部本宗同義。謂若法已斷已遍知則無。未斷未遍知則有。若業果已熟則無。業果未熟則有 (CBETA, T49, no. 2031, p. 17, a27-29)

[7] Frauwallner, Studies in Abhidharma Literature, 190ff

[8] See Barua

[9] 目乾連蘊 (CBETA, T26, no. 1539, p. 531, c29)

[10] Cf. Cousins, The 'Five Points' and the Origins of the Buddhist Schools, 58

[11] 過去未來無。現在無為有 (CBETA, T26, no. 1539, p. 532, a4-5). I have punctuated to clarify the syntax. The Pali is perhaps: atītānāgataṁ natthi; paccuppannāsaṅkhataṁ atthi. The discussion of the past, future, and present in the Kathāvatthu likewise frequently brings in Nibbana.

[12] For a recent and excellent discussion of this argument, see Bastow.

[13] CBETA, T26, no. 1539, p. 535, a8. Bastow does not notice this variation.

[14] Kathāvatthu 410

[15] 補特伽羅論 (CBETA, T26, no. 1539, p. 537, b2)