R O C Z N I K O R I E N T A L I S T Y C Z N Y, T. LXVII, Z. 1, 2014, (s.

255–268)

CHARLES WILLEMEN

Remarks about the History of the Sarvāstivāda Buddhism

Abstract

Study about the history of a specific Buddhist monastic lineage known as “Sarvāstivāda”
based on an overview of the history of its literature.

Keywords: Sarvāstivāda, Buddhism, schism, Mahāyāna, Abhidharma, India, Gandhāra

All scholars agree that the Sarvāstivāda (“Proclaiming that Everything Exists”)
Buddhism was strong in India’s north-western cultural area. All agree that there was the
first and seminal schism between the Sthaviravāda and the Mahāsāṅghika. However, many
questions still remain to be answered. For instance, when did the first schism take place?
Where exactly in India’s north-western area? We know what the Theravāda tradition has
to say, but this is the voice of just one Buddhist tradition.

Jibin 罽賓

The Chinese term Jibin is used to designate the north-western cultural area of India.
For many years it has been maintained by Buddhist scholars that it is a phonetic rendering
of a Prakrit word for Kaśmīra. In 2009 Seishi Karashima wrote that Jibin is a Chinese
phonetic rendering of Kaśpīr, a Gāndhārī form of Kaśmīra.1 In 1993 Fumio Enomoto
postulated that Jibin is a phonetic rendering of Kapiśa (Kāpiśī, Bagram).2 Historians
have long held a different view. In his article of 1996 János Harmatta said that in the
seventh century Jibin denoted the Kapiśa-Gandhāra area.3 For this opinion he relied on

1 Karashima 2009: 56–57.

2 Enomoto 1993: 265–266.
3 Harmatta (1996) 1999: 371, 373–379.
256 CHARLES WILLEMEN

Édouard Chavannes’s work published in 1903. In this work Chavannes refers to the Song
encyclopaedia Ce fu yuan gui (冊府元龜) of 1005–1013 A.D., which contains a list of
areas subjected to the Türk yabghu of Tokharistan. One of those areas, by the way, is
Fanyan guo (帆延國), which is the Chinese name for Bamiyan. In this encyclopedia Jibin
is also mentioned.4 In A. Herrmann’s widely used historical atlas of China, published in
1935 (new edition 1966), Jibin is identified with Kapiśa on the map of China in Central
Asia in 660 A.D.5 In my publications6 I maintain that Jibin originally was the Gandhāran
cultural area, which included Bactria. The term is not a phonetic rendering, but actually
refers to the area of Uḍḍiyāna (the Swat Valley) from where foreigners (bin, 賓) came
with their esteemed cloths (blankets, ji, 罽).7 The Chinese used the term to refer to
foreigners who came from across the Karakorum Mountains. The term is found in the
official History of the Former Han (206 B.C. – 8 A.D.), which was completed around
120 A.D. The term ji (blanket) occurs in the explanation of the area of Jibin.8 Thus
the term Jibin precedes the introduction of Buddhism in China. The area of Uḍḍiyāna
probably encompassed the region in which the so-called Gāndhārī Dharmapada was
composed in the second century A.D. This text was found in Hotan (和田), and it is of
Dharmaguptaka affiliation.9 Even Xuanzang (玄奘, 602–664 A.D.), who travelled in India
and Central Asia circa 629–645 A.D., still mentions the presence of Dharmaguptakas in
Uḍḍiyāna. He also mentions the presence of Mahāsāṅghikas, Mahīśāsakas, Kāśyapīyas,
and Sarvāstivādins.10
The Mahāvaṇija Jātaka (Book XIV, jātaka 493) says: kāsikāni ca vatthāni uddiyāne
ca kambale – “While Kāśī was famous for its silk, Uḍḍiyāna was famous for its blankets
(kambala).” The term ji means kambala. In Buddhist texts translated into Chinese at
a later period, the term kambala is represented by the term he (褐) or hezi (褐子). The
area stretching beyond Uḍḍiyāna was Gandhāra. From the Gandhāran area there was
relatively easy access to Hotan. During the Kuṣāṇa period, first to third century A.D.,11
the route between Uḍḍiyāna and Hotan was frequently travelled. In the third century,
Chinese (e.g. Zhu Shixing, 朱士行) went to Hotan to study Indian literature on the
perfection of wisdom (prajñāpāramitā), which is of Mahāsāṅghika affiliation. Chinese
were interested in this literature. Gandhāra was largely dominated by the Mahāsāṅghikas.
The area to the west of the Khyber Pass was also part of the Gandhāran cultural area.
Chinese know that area as Daxia (大夏). Chinese were inclined to think that their first

4 Harmatta (1996) 1999: 371 n. 12. The author refers to É. Chavannes, Documents sur les Tou-kiue (Turcs)

occidentaux. Recueillis et commentés, suivis de notes additionelles. (Présenté à l’Académie Impériale des Sciences
de St.-Pétersbourg le 23 Août 1900.) Vol. 6. St. Petersburg, 1903.
5 Herrmann 1966: 32.
6 Willemen 1998: 83; 2008: 39; 2012a: 483.
7 Willemen 2012a: 483.
8 Hulsewé 1979: 106, esp. n. 218.
9 Edited by J. Brough in 1962. See Willemen 2013a, Introduction and n. 19.
10 Li Rongxi 1996: 84.
11 Golzio 2008: 89 presents a chronological table for the Kuṣāṇa period, from the first century to the middle

of the third century A.D.

 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 257

royal dynasty called Xia, moved westward when it came to en end in China in the eleventh
century B.C.12 Daxia, Greater Xia, may be translated as Bactria. It formed the western
part of traditional Jibin, and it was mainly a Sarvāstivāda territory. But Pudgalavāda,13
Dharmaguptaka, Mahāsāṅghika, and others were also present there. From this area the
Indians crossed over the Pamirs, called Onion Range, Cong Ling (蔥嶺) in Chinese, to
Kashi (喀什, Kaxgar), and further to Kuqa (庫車). This road was much travelled e.g.
in the fourth century. Such scholars as Saṅghadeva, Saṅghabhadra,14 Kumārajīva, and
Buddhabhadra may have travelled east along this road. Traditional Jibin was the cultural
area of Gandhāra. Its westernmost part, Bactria, was also the westernmost part of Xiyu
(西域), the Western Regions, i.e. Central Asia, the area west of Dunhuang (敦煌). An
Indian said to come from Jibin and from Xiyu, was a Bactrian.
During the reign of Kaniṣka (155 – ca. 179 A.D.),15 about 160–170 A.D., a Sarvāstivāda
synod was held in Kaśmīra. This synod established a shorter vinaya and a new abhidharma
in Sanskrit. In the Gandhāran cultural area was used Gāndhārī language, written in
Kharoṣṭhī script. Traditional Sarvāstivādins had a long vinaya, transmitted by Upagupta
from Mathurā.16 It contained many stories, dṛṣṭāntas, illustrating the rules. That is why the
traditional Sarvāstivādins could be called Dārṣṭāntikas. Their new version of vinaya, devoid
of most of the stories and much briefer, composed in Sanskrit, was called Daśabhāṇavāra,
the Vinaya in Ten Recitations (十誦律 Shi song lü, T. XXIII 1435). The rules however
remained unchanged. The main Gandhāran abhidharma text, the Aṣṭagranthaśāstra of
Kātyāyanīputra, Treatise with Eight Compositions (八犍度論 Ba jiandu lun, T. XXVI
1543), translated into Chinese by Saṅghadeva, was rewritten in Sanskrit and called
Jñānaprasthānaśāstra, Treatise: Development of Knowledge (發智論 Fa zhi lun, T. XXVI
1544); it was translated into Chinese by Xuanzang in 657–660 A.D. So, a new vibhāṣā,
or a commentary, was needed. This was the Sanskrit Mahāvibhāṣāśāstra, or Treatise:
Great Commentary (大毗婆沙論 Da piposha lun, T. XXVII 1545). Therefore from now
on, following the name of this text the new “orthodoxy” in Kaśmīra became known as

12 One may also think of the Western Xia, Xi Xia, the Tanguts (1032–1227 A.D.). Not only the ancient Xia
are supposed to have migrated westward, also the Qin (秦, 221–206 B.C.) were supposed to have done so.
13 The “Personalists” are the Vātsīputrīyas, later known as Sāṃmitīyas. Saṅghadeva translated their San fa

du lun (三法度論, Tridharmakhaṇḍa(ka)śāstra(?), Segments of the Three Factors), into Chinese (T. XXV 1506).
The three factors, dharmas, are: guṇa (qualities), doṣa (evil), āśraya (basis, i.e. the body). Each factor consists
of three parts, forming an abhidharma with nine characteristics. The text was translated on Mt. Lu in 391 A.D.,
and inspired Sengyou’s (僧祐) Hong ming ji (弘明集, T. LII 2102, 34bc), where we read Huiyuan’s (慧遠) San
bao lun (三報論), about the three kinds of retribution (Willemen 2006 a:7; 2008: 37–38). Pudgalavāda was quite
strong in Bactria, with close links to Sarvāstivāda there. It was later even thought necessary to add a refutation of
the pudgala (person) as the ninth chapter of the Abhidharmakośabhāṣya, which was an elaboration of the Bactrian
Abhidharmahṛdaya.
14 Sengqie Bacheng or Badeng (僧伽跋澄). The name means Zhongxian (眾賢), and it is not equivalent to

Saṅghabhūti. He probably came to Chang’an from the area of Bactria about 380 A.D. He translated the Udāna
(出曜經 Chuyao jing, T. IV 212) in 399 A.D., and the Vibhāṣā(-śāstra) (鞞婆沙論 Biposha(lun), T. XXVIII 1547),
a commentary on the Aṣṭagrantha, in 383 A.D.
15 Golzio 2008: 89.
16 Lamotte 1988: 174.
258 CHARLES WILLEMEN

Vaibhāṣika.17 Six more existing texts, rewritten and eventually enlarged, were added to
the Jñānaprasthāna, forming as it were its parts, padas or pādas. The term pāda or pada
just means “part”, usually one fourth, but Xuanzang uses the term zu 足, which usually
means “foot”, but which may mean also “part” (usually ”one fourth”).18 Because the
Mahāvibhāṣā does not quote the Dhātukāyaśāstra, Treatise: Corpus of Elements (界身論
Jie shen lun, T. XXVI 1540), which belongs to those six parts, it seems that this text
may have been completed later, perhaps in early third century(?).19 The new “orthodoxy”
of seven texts claimed that they were proclaimed by the Buddha (buddhabhāṣita) in
heaven. The traditional Sarvāstivādins did not believe that. For them only the sūtras of
the āgamas (traditions) were proclaimed by Buddha. That is why the term Sautrāntikas
is used for them. The terms Sautrāntika and Dārṣṭāntika go together like the compound
dharmavinaya.20 Throughout the history, however, the term Sarvāstivāda may be used,
because their heterogeneous groups all agreed on sarvāstitva. But they did not agree
what sarvam (‘everything’) and even asti (‘is’) really meant.21 Thus, Sarvāstivādins were
split into two main groups at the end of the second century A.D.: 1. new “orthodoxy”
in Kaśmīra; and 2. heterogeneous groups of traditional Sarvāstivādins. The traditional
groups were located in Gandhāra (Bactria) and in northern India (Magadha). Because
Kaśmīra was the area of the “orthodoxy” in the north-west, it became a very prominent
part of Jibin.22 In China Sautrāntikas were the Sarvāstivādins. Their abhidharmaśāstras
offered manuals instructing monks how to become an arhat. They followed the brief
Daśabhāṇavāra. That is why the term Jingliang (經量), Sautrāntika, is well known
in China, whereas the Chinese term for Dārṣṭāntika, Piyuzhe (譬喻者), is much less
known. The Vaibhāṣika abhidharma was brought to China by Xuanzang in the seventh
century, just before it lost its position to the non-Vaibhāṣikas in India. Nālandā was
under a non‑Vaibhāṣika influence not only in the time of Xuanzang. About 700 A.D.
Sarvāstivādins were united again, now they were also called Mūlasarvāstivādins. Their
vinaya was very extensive as it contained numerous stories.23 This vinaya had by then
a very long development. Vaibhāṣikas now looked like one more group of traditionally

17 Willemen 2008a: 39–41.

18 Willemen 2008a: 47.
19 Willemen 2008a: 46–47.
20 Willemen 2008a: 45.
21 Willemen 2008a: 41. Does “everything” mean all factors [dharma]? If so, how many? Or does it mean all

aggregates (skandha)? And what precisely means ”is”: When? Present and past are relatively easy to understand
but the future?
22 Willemen 2008a: 71. Xuanzang uses Jibin for Kaśmīra, the area of Sarvāstivāda “orthodoxy” since the third

century.
23 The term Mūlasarvāstivāda appears after Xuanzang left India and before the arrival of Yijing (義淨), i.e.

towards the end of the seventh century. The extensive Mūlasarvāstivāda vinaya, as it was preserved in e.g. Tibetan
version, may be regarded as a continuation of the vinaya practiced by Dārṣṭāntikas. A long time has passed from
Upagupta to the eighth century A.D. The bulky vinaya certainly was subject to a very long development. It should
be reminded that the non-Vaibhāṣikas were by no means uniform.
 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 259

heterogeneous Sarvāstivādins. This kind of Sarvāstivāda Buddhism spread to Tibet and

then to Mongolia.

The Basic Schism

The Theravāda tradition recognizes the second synod, saṅgīti, which took place in
Vaiśālī. In fact, there was discord, but at the end there was no schism. The Theravāda
tradition then mentions the third synod during Aśoka’s reign (ca. 264–227 B.C.) in
Pāṭaliputra. Only Sthaviras seem to have participated. The basic schism seems to have
taken place before Aśoka and after the Vaiśālī synod. The schism seems to have taken
place in Pāṭaliputra during the reign of Mahāpadma Nanda (ca. 340 B.C.).24 A schism can
only occur due to a disagreement about vinaya rules. But it is quite possible that there
were considerable doctrinal differences in the saṅgha, which have lead also to a vinaya
disagreement. We do not know what kind of the doctrinal disagreements were at that
time. We only have an idea of what happened later. One group, the Sthaviravāda, claimed
the factors (dharmas) to be existent. This group had as its religious goal to become an
arhat, and their texts taught a long road to arhatship, based on the four noble truths.
Their teachings they said to come from Śāriputra. The other group, the Mahāsāṅghikas,
emphasized the six perfections, and especially the perfection of wisdom, prajñāpāramitā.
They proclaimed emptiness of all things and they strove for the perfect enlightenment of
a Buddha. Their texts outlined even a longer path of realization to become a bodhisattva.25
This group was in the majority. It seems that the sthaviras, or elders, wanted to keep
the Order in unity by adding some minor rules to the vinaya, but the majority did not
agree. In result there was a schism. W. Pachow’s comparative study of the prātimokṣas,
the basic set of rules for the monks, confirms that the vinaya of the Mahāsāṅghikas, as
we have it in Chinese translation, is very old.26 The Theravāda vinaya is ancient too.
The Śāriputraparipṛcchāsūtra, or Canonical Text about the Questions of Śāriputra
(舍利弗問經 Shelifu wen jing, T. XXIV 1465), which exists only in a Chinese version
of the Eastern Jin (東晉, 317–420 A.D.), a text of Mahāsāṅghika affiliation, confirms
that the Mahāsāṅghika vinaya is ancient. Thus, a vinaya disagreement directly led to the
first schism, even though it may have been also the result of other disagreements. The
Buddhist Theravāda tradition mentions that five points raised by a certain Mahādeva were
the cause of the split, but it seems likely that a Mahādeva played a role in the further
splitting up of Mahāsāṅghikas later in Andhra. Aśoka’s synod was a Sthaviravāda synod.
The winners during the synod called themselves Vibhajyavādins, analysts, or distinctionists.
The Pudgalavāda Vātsīputrīyas had left the main group somewhat earlier. A. Bareau
defined Vibhajyavādins as the non-Vātsīputrīya sthaviras who opposed Sarvāstivāda
24 Willemen 2012b: 1.
25 Willemen 2012b: 2–3. There is definitely a link between the path of an arhat (Sthaviravāda) and the path
of a bodhisattva (originally Mahāsāṅghika).
26 Willemen 2012b: 1, referring to research of Pachow 1955.
260 CHARLES WILLEMEN

ideas.27 Accordingly, meant are the later Mahīśāsakas, Dharmaguptakas, Kāśyapīyas,

and the Śrī Laṅkan Theravādins. I have proposed that the Vibhajyavādins were called
Mahīśāsakas, “converting or instructing the earth”, i.e. India, when they spread all over
India, from Gandhāra to Śrī Laṅkā. Theravāda Buddhism was a conservative movement
in Śrī Laṅkā, which reacted against the Mahīśāsakas, who were present on the island
but who kept stronger position in south India. It also reacted against the views of the
continental Mahāsāṅghikas.28
The Mahāsāṅghikas used the term Mahāyāna with regard to their form of Buddhism.29
They regarded Mahākātyāyana as their ultimate source of wisdom.30 The Mahāsāṅghika
versus Sthaviravāda rivalry throughout the history of Buddhism is more important than
the so-called Mahāyāna(Great Vehicle) – Hīnayāna(Lesser Vehicle) division. Both groups
used successful ideas of their antagonists, a fact which can be observed in the meditation
manuals, texts about the practice of yoga, yogācāra.31 Asaṅga (second half of the fourth
century) was a Mahīśāsaka monk who continued Gandhāran Sautrāntika yogācāra,
adopting the Mahāsāṅghika Madhyamaka conception of emptiness. Sino-Japanese “Pure
Land” Buddhism seems to have originated in Bactrian Sautrāntika circles, in reaction to
the Lokottaravāda Mahāsāṅghika idea of a Buddha land, Buddhakṣetra.32 Therefore there
is also a Sarvāstivāda Mahāyāna. No wonder that the term Hīnayāna is far less frequent.

Sarvāstivāda Abhidharma

During the synod in Kaśmīra (second half of the second century) the new Vaibhāṣika
“orthodoxy” established a Sanskrit abhidharma, proclaimed by the Buddha, and consisting
of seven (i.e. six plus one) texts.33
Puguang (普光, fl. 645–664 A.D.), a disciple of Xuanzang, gave the following
information about the six parts (“feet”) of the “orthodox” Sarvāstivāda abhidharma in
his commentary on Vasubandhu’s Abhidharmakośa (俱舍論記 Jushe lun ji, T. XLI 1821,
p. 8b26–c6). Pugang’s information probably comes from his master Xuanzang, who
translated the Vaibhāṣika texts into Chinese:
1. Saṅgītiparyāyapāda-śāstra by Śāriputra (T. XXVI 1536, from 663 A.D.).
2. Dharmaskandha° by Mahāmaudgalyāyana (T. XXVI 1537, from 659 A.D.).

27 Bareau 1955: 169, 206.

28 Willemen 2012a: 481, 491.
29 Willemen 2012b: 3, 10.
30 Willemen 2008a: 67.
31 Willemen 2012b: 8–11.
32 The Lokottaravāda Mahāsāṅghikas may have developed the idea of Buddha fields, Buddha not being restricted

to one earthly existence. In Gandhāra, the eastern part of traditional Jibin, the concept of Akṣobhya’s paradise Abhirati
was developed. Immediately after that the Sarvāstivādins (and Pudgalavādins) in Bactria, the western part of traditional
Jibin, developed the concept of intermediate existence (antarābhava) as their paradise, Sukhāvatī, the so-called
Pure Land. Already in Kuṣāṇa times both paradises occur side by side in the same texts. See Willemen 2013b.
33 Lamotte 1988: 184–185.
 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 261

3. Prajñapti° by Mahākātyāyana (T. XXVI 1538, not translated by Xuanzang).

4. Vijñānakāya° by Devaśarman (提婆設摩 Tipo Shemo. This is translated as Tianji
天寂, Devaśama?) (T. XXVI 1539, from 649 A.D.).
5. Prakaraṇa° by Vasumitra (T. XXVI 1542, from 660 A.D.).
6. Dhātukāya° by Vasumitra (T. XXVI 1540, from 663 A.D.).
According to Yaśomitra’s (eighth century) Abhidharmakośa-vyākhyā ad I.11 the
abhidharma works were composed by the following authors: no. 1 by Mahākauṣṭhila,
no. 2 by Śāriputra, no. 3 by Maudgalyāyana, no. 4 by Devaśarman, no. 5 by Vasumitra,
and no. 6 by Pūrṇa.34
Xuanzang may not have had an Indian text of the Prajñapti°. The text catalogued
in Taishō vol. XXVI no. 1538 is not a Vaibhāṣika text but the second chapter entitled
Kāraṇaprajñapti of a text translated by Dharmapāla (963–1058 A.D.) and Weijing (惟淨)
shortly after 1004 A.D. The Chinese text is equally of Mahāsāṅghika as of Sarvāstivāda
affiliation.35 The original Indian text supposedly had eight chapters, three of which exist
in Tibetan translation, the rest is lost. Therefore, exceptionally T. 1538 is not a Vaibhāṣika
text.36 Puguang’s information follows the Vaibhāṣika tradition (except for the Prajñapti°)
whereas Yaśomitra follows the Mūlasarvāstivāda tradition.
The basic text of the Vaibhāṣika “orthodoxy” is called Jñānaprasthāna, which is
a rewritten Gandhāran Aṣṭagrantha.37 The Gandhāran text had many commentaries
called Vibhāṣā. Two of them exist in Chinese translation, viz. 1. Biposha lun (鞞婆沙論,
T. XXVIII 1547), translated in 383 A.D. by Saṅghabhadra;38 and 2. Apitan piposha
lun (阿毗曇毗婆沙論,T. XXVIII 1546), translated in 439 A.D. by Buddhavarman.
The new commentary on the Jñānaprasthāna, the Vaibhāṣika Mahāvibhāṣā – Apidamo
da piposha lun (阿毗達磨大毘婆沙論, T. XXVII 1545), was translated by Xuanzang
in 659 A.D. It may be added that the Indian non-Vaibhāṣika texts gradually have
adapted to the “orthodoxy” from the third century on. This process may be noticed
in the Vibhāṣās as well as in a non-Vaibhāṣika Prakaraṇa° (眾事分阿毗曇論 Zhong
shi fen apitan lun, T. XXVI 1541), translated by Guṇabhadra and Bodhiyaśas
in 443 A.D.
It is very likely that the non-Vaibhāṣikas also established a corpus comparable with
the seven Vaibhāṣika abhidharma texts.39 Harivarman’s Chengshi lun (成實論, T. XXXII
1646) makes this clear.40 Harivarman (ca.300 A.D.) was a brahmin from Central India

34 Wogihara 1971: 11.

35 This was established by Lin Li-kouang 1949: 137–144.
36 Willemen 2008a: 60–62.
37 Willemen 2008a: 43.
38 See note 14; Willemen 2008a: 59.
39 Willemen 2008a: 54–57.
40 Jñānakāya, Corpus of Knowledge, (i.e. the Aṣṭagrantha and its additional parts), Prodbhūtopadeśa, The

Explanatory Discourse: (Corpus of Knowledge) Realized. Willemen 2006b; 2008: 54–55. The Sanskrit title of the
Chengshi lun is not Tattvasiddhi-śāstra or Satyasiddhi°.
262 CHARLES WILLEMEN

who – as so many brahmins41 – was converted to Sautrāntika Buddhism. He was a disciple

of Kumāralāta. Kumāralāta (ca. 150 A.D.?) is the reputed first master, mūlācārya, of
Sautrāntikas. Harivarman was his disciple in the sense that he has accepted his ideas.
Harivarman later changed his ideas under the influence of a (Bahuśrutīya?) Mahāsāṅghika
master. As a result of this development he composed his Chengshi lun (T. 1646). In it
he mentioned the part named Lokasthāna (Loutan 樓炭), which is one of the six parts
of the so-called abhidharma “with six parts (feet)”.42 Also Dharmapriya (in China ca.
382–390 A.D.), a Sautrāntika who assisted Saṅghadeva to complete his translation of
the Aṣṭagrantha, mentioned six parts (feet), evidently different from the six parts of the
Vaibhāṣikas.43
There was another non-Vaibhāṣika group of abhidharma texts very popular in
southern China. The basic text of this lineage is the Abhidharmahṛdaya-śāstra, Treatise:
Heart of the Abhidharma (阿毘曇心論 Apitan xin lun, T. XXVIII 1550). The text was
composed by Dharmaśreṣṭhin, a man from the area of the river Vakṣu, and translated
into Chinese by Saṅghadeva in 391 A.D.44 The text comes from the Bactrian cultural
area, maybe from the first century B.C., as proposed by E. Frauwallner in 1971.45 This
basic Bactrian text was commented upon more than once in the Gandhāran area. There
is Upaśānta’s (third century) Abhidharmahṛdaya (阿毘曇心論經 Apitan xin lun jing,
T. XXVIII 1551) translated into Chinese by Narendrayaśas in 563 A.D. Then there is
a very influential commentary entitled Miśrakābhidharmahṛdaya-śāstra, Treatise: Heart
of the Abhidharma, With Miscellaneous Additions (雜阿毘曇心論 Za apitan xin lun,
T. XXVIII 1552), composed early in the fourth century by the Gandhāran Dharmatrāta
and translated into Chinese by Saṅghavarman in 435 A.D. in Jiankang (Nanjing).46 This
text together with Saṅghadeva’s translation of the Aṣṭagrantha very rapidly became central
for the Abhidharma School (毗曇宗 Pitan Zong) in southern China.47 Finally there was
Vasubandhu’s Abhidharmakośa-bhāṣya, a work based on the Miśraka°, as T. Kimura has
shown in 1922.48 With the discovery of the Sanskrit original of the Kośa°, Storehouse

41 Even in the time of the Buddha many converts were brahmins. This situation apparently did not change
later (e.g. Aśvaghoṣa). It is even understandable that the Pudgalavāda school came into existence as a result of
this large number of brahmins, who saw in a pudgala an alternative for ātman.
42 Willemen 2008a: 55, 61. The Lokasthānasūtra (樓炭經 Loutan jing, T. I 23) is a cosmological text from

303 A.D., which presents a Mahāsāṅghika and a non-Vaibhāṣika tradition.

43 Willemen 2008a: 55–56.
44 The whole text was translated in Willemen 2006a. The text was first translated and studied in 1975 in

Brussels, but a revised new introduction was offered in the new Indian edition. The 1975 edition is outdated. It
is noteworthy that the name of Dharmaśreṣṭhin is translated as Fasheng 法勝, Dharma-Excellence, while the term
abhidharma is sometimes explained as sheng 勝 (śreṣṭha, excellent) fa 法 (dharma).
45 Willemen 2006a: 3 n. 10, referring to E. Frauwallner’s research from 1971.
46 Willemen 2008a: 48 again explains that the Sanskrit title of Dharmatrāta’s text is Miśraka°, not Samyukta°. This

was established by Kudara Kōgi in 1982 and by Ryose Wataru in 1986. Yaśomitra mentions a Miśrakakāra, most likely
Dharmatrāta; see Wogihara 1971: 251. The text mixes traditional non-Vaibhāṣika views of the Abhidharmahṛdaya
with the “orthodox” Vaibhāṣika views. There are some additional parts in the text.
47 Willemen 2008a: 49.
48 Willemen 2006a: 2 n. 3, referring to research by Kimura Taiken from 1922.
 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 263

[of abhidharma], we possess a complete Sanskrit abhidharmic text since its edition
in 1967.49 For the texts mentioned earlier we hardly have parts of an original Indian
version. The Sanskrit titles and even the names of Indian authors were reconstructed
on the basis of Chinese and other sources, e.g. Yaśomitra’s Vyākhyā. The Kośa° was
written by the Gandhāran Vasubandhu (ca. 350–430 A.D.),50 and translated into Chinese
by Paramārtha in 568 A.D. (阿毘達磨俱舍釋論 Apidamo jushe shi lun, T.XXIX 1559).
This text replaced the Miśraka° in southern China and became the central text of the
Chinese Kośa School. Xuanzang brought a new version of Paramārtha’s text, namely
Apidamo jushe lun (阿毘達磨俱舍論, T. XXIX 1558), in 651-654 A.D. This text became
the central text of a new Kośa School in Chang’an, in the north. Chinese developments
clearly show that the abhidharma there was of a Sautrāntika type, belonging to the
Gandhāran cultural area. Vasubandhu’s Kośa° was strongly opposed by the “orthodox”
Vaibhāṣika master Saṅghabhadra (late fourth century?). He defended the “orthodoxy”
in his treatise Nyāyānusāra(?), Conforming to Correct Principles (阿毘達磨順正理論
Apidamo shun zhengli lun, T. XXIX 1562), translated by Xuanzang in 653–654 A.D. He
did the same in his Abhidharmasamayapradīpikā(?), Obvious Correct Principles of the
Abhidharmapiṭaka (阿毘達磨藏顯宗論 Apidamo zang xian zong lun, T. XXIX 1563),
translated by Xuanzang in 651–652 A.D.51
There were, of course, many more non-Vaibhāṣika abhidharma texts, as for example
Abhidharmāmṛta(rasa)(?)-śāstra, Treatise: The Flavour of Immortality of the Abhidharma
(阿毘曇甘露味論 Apitan ganlu wei lun, T. XXVIII 1553), the work of Ghoṣaka;
Abhidharmāvatāra-śāstra, Treatise: Introduction to Abhidharma (入阿毘達磨論 Ru
apidamo lun, T. XXVIII 1554), the work of Skandhila(?). These texts were ”about”
(abhi°) the teaching, dharma, and therefore could be placed in an Abhidharmapiṭaka.
The Sautrāntikas also possessed this Piṭaka.

More Sarvāstivāda Literature

In order to be somewhat more complete, I have to say that the Sarvāstivāda

literature extended beyond the collections of abhidharma and vinaya. When the new
Daśabhāṇavāra was established in Kaśmīra one can observe the beginning of independent
extra-canonical avadāna or narrative literature. Kumāralāta’s (middle second century
A.D.?) Kalpanāmaṇḍitikā (大莊嚴論經 Da zhuangyan lun jing, T. IV 201), translated
by Kumārajīva in 384–401 A.D., also known as Dṛṣṭāntapaṅkti, Collection of Similes,
Stories (dṛṣṭāntas), is found at the beginning of this phenomenon.52 Next there is the

49 Pradhan 1967; second ed. 1975.

50 Willemen 2008a: 49, referring to F. Deleanu.
51 Cox 1998: 242–249.
52 As it has been shown by H. Lüders, this text is different from Aśvaghoṣa’s Sūtrālaṃkāraśāstra. Zhuangyan

lun (莊嚴論) is a direct translation of kalpanāmaṇḍitikā. Knowing that Aśvaghoṣa was a Sautrāntika influenced by
Bahuśrutīya(?) Mahāsāṅghika ideas, the erroneous attribution may be understood. In many respects Harivarman,
264 CHARLES WILLEMEN

early Chinese collection of one hundred stories known as Avadānaśataka (撰集百緣經

Zhuanji bai yuan jing, T. IV 200) attributed to Zhi Qian (支謙) from ca. 223–253 A.D.
Next I would like to mention the so-called Udānavarga. This text belongs to
Sarvāstivāda Dharmapada literature. The Dharmapada is probably the best known
Buddhist text in the world. Every Buddhist school, nikāya (Chinese bu) had one, and the
non-Vaibhāṣikas may have had more than one.53 The so-called Udānavarga initially was
a non-Vaibhāṣika Dharmapada, Verses of the Law, compiled by Dharmatrāta, a Gandhāran,
in ca. 150 A.D.54 To this versified Dharmapada stories were added, making it into a real
Udāna. This text exists in Chinese translation by Saṅghabhadra from 399 A.D. (出曜經
Chuyao (Udāna) jing, T. IV 212). This text formed the sixth “member” (aṅga) of the
Buddha’s word (buddha-vacana) as classified into twelve “members”, and later included
into the Kṣudrakapiṭaka, which contained the “minor” texts. The Chuyao jing (Udāna)
was a collection of udānas, or verses expressing intellectual joy,55 put together into a text
called Udāna. This is explained in the prose parts of the Chuyao jing itself. Later the
stanzas, or udānas, were collected from this Udāna again to form a Dharmapada work
known to us as Udānavarga, a collection of udānas from the “member” Udāna. Besides
Sanskrit and Tibetan versions the so-called Udānavarga exists in Chinese version too.
It is the work of Śāntideva56 (Tian Xizai 天息災) from 985 A.D., entitled A Collection
of Important Odes of the Law (法集要頌經 Fa ji yao song jing, T. IV 213). Fa yao
song (法要頌), important odes of the law, means Dharmapada, and ji (集), collected,
translates °varga. The text is of Mūlasarvāstivāda affiliation.57
Other non-Vaibhāṣika literature includes biographies of the Buddha – Lalitavistara
and especially Aśvaghoṣa’s Buddhacarita. Aśvaghoṣa (ca.100 A.D.) may now be called
a Sautrāntika (the word did not exist at that time) influenced by Bahuśrutīya(?) ideas.
His text In Praise of Buddha’s Acts (佛所行讚 Fo suoxing zan, T. IV 192) counts twenty
eight chapters in both the Tibetan and Chinese versions. In the Sanskrit original has been
preserved only half of the text, namely fourteen chapters. Aśvaghoṣa’s Sanskrit poem
belongs to the corpus of world literature. The Chinese version from 421 A.D. is rather
a paraphrase of the contents made by Baoyun (寶雲). This version was translated by the
present author in 2009, offering a readable English version which follows the Chinese
original closer than Ōminami Ryūshō’s (大南龍昇) Japanese translation from 2002.58
Furthermore, some Āgamas belong to non-Vaibhāṣika Sarvāstivāda literature, viz.
Saṅghadeva’s Madhyamāgama (中阿含經 Zhong ahan jing, T. I 26) and most likely also

a ”disciple” of Kumāralāta, may be compared with Aśvaghoṣa. Note that Kumārajīva from Kuqa translated this
non-Vaibhāṣika text before he arrived in Chang’an.
53 Willemen 2012a: 485. Cf. Willemen 2013, Introduction.
54 About Dharmatrāta see Lin Li-kouang 1949: 324–351.
55 Lamotte 1988: 144.
56 Lokesh Chandra established this name in 2010. See Willemen 2013, Introduction, n. 63.
57 Willemen 2013a: Introduction.
58 Ōminami Ryūshō (大南龍昇) has offered the latest Japanese translation of the Chinese Buddhacarita in 2002

in Shin Kokuyaku Daizōkyō (新國譯大藏經), Hon’enbu (本緣部) I: 125–426.

 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 265

his Ekottarikāgama (增一阿含經 Zengyi ahan jing, T. II 125), completed in 398 A.D.
The Tokharian Dharmanandin had begun the Āgama translations in 384 A.D. in northern
China.59
An important part of the non-Vaibhāṣika literature present meditation manuals. They
were especially popular in Bactrian circles. However, the whole Gandhāran cultural area
has produced such manuals. One of such manuals, preserved in the Sanskrit original
(incomplete) was edited by D. Schlingloff under the title Yogalehrbuch in 1964.60 In
China such manuals are often called “texts about the five gates to dhyāna” (wu men chan
五門禪). The Chinese texts usually add a Mahāsāṅghika, a Mahāyāna part.61 Already An
Shigao (Ashkani, 安世高), active in Luoyang (洛陽) ca. 148–170 A.D., brought out the
Yogācārabhūmi (道地經 Dao di jing, T. XV 607), a treatise about the stages in the practice
of yoga. His text agrees with a part of Saṅgharakṣa’s Yogācārabhūmi.62 Shortly after An
Shigao, Loujia Chen (婁迦讖, usually sanskritized as Lokakṣema)63 translated in 179 A.D.
in Luoyang the earliest known text about the perfection of wisdom (prajñāpāramitā),
viz. the Aṣṭasāhasrikā-prajñāpāramitā. He entitled his work Daoxing (yogācāra) bore
(prajñā) jing (道行般若經, T. VIII 224) to show that this text is wholly devoted to
different kinds of yoga practice.64 The text introduces a (Gandhāran) Mahāsāṅghika,
Mahāyāna type of yoga, as opposed to the Sautrāntika Savāstivāda one. But Sautrāntika
practices were better known in China. Dharmarakṣa’s Xiuxing dao di jing (修行道地經,
T. XV 606) from 284 A.D. offers a translation of Saṅgharakṣa’s text.65 Immediately
upon his arrival in Chang’an (長安) in 402 A.D. Kumārajīva (Jiumo Luoshi 鳩摩羅什)
was asked to bring forward meditation techniques. He accordingly explained the method
of concentration of sitting meditation in his Zuochan sanmei jing (坐禪三昧經, T. XV
614),66 and also in the Essential Explanation of the Way how to Meditate (Chan fa yao
jie 禪法要解, T. XV 616). More meditation manuals were attributed to Kumārajīva in
the unreliable catalogue of Fei Zhangfang (費長房) from 597 A.D. (歷代三寶紀 Lidai
san bao ji, T. XLIX 2034), as e.g. Siwei lüeyao fa (思惟略要法, T. XV 617),67 which is
a text of Mahāsāṅghika affiliation. Dharmamitra’s texts, viz. Basic ways in Sūtras about
59 Willemen 2006a: 6–8.
60 See Schlingloff 2006.
61 Demiéville 1954: 349–351.
62 An Shigao’s text offers chapters 1–5, 22, and 24 of Saṅgharakṣa’s text. Did An Shigao make a selection, or

is this (part of ) the original core of the text?

63 Laukākṣina, as suggested by Lokesh Chandra, is also possible. Cf. Willemen 2004: 10 n. 27.
64 Willemen 2008a: 42. Sthaviravāda term prajñā, wisdom, is always rendered by a “meaningful translation”,

viz. zhihui (智慧), but Mahāsāṅghika “emptiness-wisdom” is rendered by a “sound translation”, viz. panre/banre 般
若), later (certainly in the fifth century in southern China), however, pronounced bo (波) re. For this see Willemen
2011, esp. p. 149.
65 Studied in Demiéville 1954.
66 Immediately upon his arrival Kumārajīva was asked by Sengrui (僧叡) to teach the meditation techniques.

The master explained T. 614 in about one week. For a translation see Yamabe and Sueki 2009.
67 Translated in Willemen 2012b. The text is anonymous, but wrongly attributed to the famous Kumārajīva.

The text is of Mahāsāṅghika affiliation, probably composed by Chinese monks in southern China in 430–440 A.D.,
based on the oral teaching of an Indian master (Dharmamitra?).
266 CHARLES WILLEMEN

the Five Gates of Dhyāna (五門禪經要用法 Wu men chan jing yao yong fa, T. XV 619),
as well as Sūtra(s) about the Secret Essence of Dhyāna (禪祕要（法）經 Chan mi yao
(fa) jing, T. XV 613), actually were composed by Chinese monks in Jiankang (建康),
who were listening to the teaching of Dharmamitra, a Mahāsāṅghika monk.68 There is
also Buddhabhadra’s text about the dhyāna of his master Buddhasena from Bactria, i.e.
Yogācārabhūmi (修行道地 Xiuxing dao di, T. XV 618). This text has been given the
erroneous title Dharmatrāta’s Dhyāna (達摩多羅禪經 Damo Duoluo chan jing) by Fei
Zhangfang.69 Non-Vaibhāṣika meditation manuals deal with yogācāra, the practice of
yoga. Because the Sautrāntikas used ideas of their antagonists, the Mahāsāṅghikas, their
yoga techniques constantly developed into new forms. Asaṅga (second half of the fourth
century) incorporated Mahāsāṅghika Madhyamaka ideas into his Gandhāran yogācāra
work. But the importance of the non-Vaibhāṣikas is even greater when one realizes
that “Pure Land” Buddhism most likely is Sautrāntika too. One can safely say that the
Sarvāstivāda Sthaviravāda and Mahāsāṅghika ideas, often originating in the Gandhāran
cultural area, really shaped East Asian Buddhism and what we know as the Mahāyāna
movement.70 Even Bodhidharma, the patriarch of Chan or Zen (禪), was born into a family
of Gandhāran origin.71

Bibliography

T. = Taishō Shinshū Daizōkyō

Bareau, André. 1955. Les sectes Bouddhiques du Petit Véhicule. Saigon: École Française d’Extrême-
Orient.
Chavannes, Édouard. 1903. Documents sur les Tou-kiue (Turcs) occidentaux. Paris: Adrien Maisonneuve.
Cox, Collett. 1998. “Kaśmīra: Vaibhāṣika Orthodoxy.” In: Sarvāstivāda Buddhist Scholasticism. Charles
Willemen, Bart Dessein, Collett Cox (eds.). Leiden: Brill: 138–254.
Demiéville, Paul. 1954. “Le Yogācārabhūmi de Saṅgharakṣa.” In: Bulletin de l’École Française
d’Extrême-Orient 44: 339–436.
Dhammajoti, Bhikkhu K.L. 2004. Sarvāstivāda Abhidharma. Sri Lanka: Centre for Buddhist Studies.
Enomoto, Fumio (榎本文雄). 1993. “Keihin-Indo Bukkyō no Ichichūshinchi no Shozai (罽賓-印度佛
教の一中心地の所在).” In: Chi no Kaikō-Bukkyō to Kagaku (知の邂逅-佛教と科學). Tsukamoto Keishō
Kyōju Kanreki Kinen Ronbunshū Kankōkai (塚本啟祥教授還暦紀念論文集刊行會). Tōkyō: Kōsei (佼成)
Publishing.

68 Willemen 2012b: 6–7.

69 Willemen 2013b: 19–20; 2012b: 8.
70 For Sarvāstivāda Mahāyāna see Willemen 2013b (“Pure Land”); Willemen 2008a: 47–48; Willemen 2012b: 3

(Asaṅga’s new yogācāra). One should, however, not forget that there was also a maritime route from east and
south-east India via Funan (扶南) (mainly Cambodia) to southern China, from the third century A.D. until ca.
600 A.D. This route was overwhelmingly Mahāsāṅghika, but also influenced by Mahīśāsaka and Sarvāstivāda
monks. See e.g. Willemen 2008b.
71 Willemen 2013b n. 30. In Jingde chuandeng lu (Keitoku dentō roku) (景德傳燈錄, T. LI 2076, p. 217a9),

from 1004 A.D., it is mentioned that his family comes from Xiangzhi (香至), Gandhavatī.
 REMARKS ABOUT THE HISTORY OF THE SARVĀSTIVĀDA BUDDHISM 267

Golzio, Karl-Heinz. 2008. “Zur Datierung des Kuṣāṇa-Königs Kaniṣka I.” In: Bauddha-sāhitya-
stabakāvalī. Essays and Studies on Buddhist Sanskrit Literature dedicated to Claus Vogel by Colleagues,
Students, and Friends. Dragomir Dimitrov, Michael Hahn, Roland Steiner (eds.). Marburg: Indica et Tibetica
36: 79–91.
Harmatta, János. (1996)1999. “Tokharistan and Gandhara Under Western Türk Rule. Part One: History
of the Regions.” In: History of Civilizations of Central Asia. B.A. Litvinsky, Zhang Guang-da, R. Shabani
Samghabadi (eds.), III: 367–383. Delhi: Motilal Banarsidass.
Herrmann, Albert. (1935)1966. An Historical Atlas of China. Amsterdam: Djambatan N.V. Publishers
& Cartographers.
Hulsewé, Anthony. 1979. China in Central Asia. The Early Stage:125 B.C.–A.D.23. Sinica Leidensia
14. Leiden: Brill.
Karashima, Seishi (幸島靜志). 2009. “Hanyi Fodian de yuyan yanjiu (漢譯佛典的語言研究).” In: Zhu
Qingzhi (朱慶之), (comp.) Fojiao Hanyu yanjiu (佛教漢語研究). Beijing (北京): Shangwu Yinshuguan
(商務印書館): 33–74.
Lamotte, Étienne. 1988. History of Indian Buddhism. Transl. S. Boin-Webb. Louvain-la-Neuve: Institut
Orientaliste.
La Vallée Poussin, Louis de. (1923–1931)1971. L’Abhidharmakośa de Vasubandhu. 6 Vols. Bruxelles:
Institut Belge des Hautes Études Chinoises.
Li Rongxi. 1996. The Great Tang Dynasty Record of the Western Regions. Berkeley: Numata Center
for Buddhist Translation and Research.
Lin, Li-kouang. 1949. L’Aide-Mémoire de la vrai loi. Introduction au compendium de la loi. Paris:
Librairie d’Amérique et d’Orient.
Ono, Genmyō (小野玄妙); Maruyama, Takao (丸山孝雄). 2000. Bussho Kaisetsu Daijiten (佛書解
說大辭典). Compact edition. Tōkyō (東京): Daitō Shuppansha (大東出版社).
Pradhan, Prahlad (ed.). 1967. Abhidharmakośabhāṣya of Vasubandhu. Tibetan Sanskrit Works Series
VIII. Patna: K.P. Jayaswal Research Institute.
Schlingloff, Dieter. 2006. Ein buddhistisches Yogalehrbuch. Unveränderter Nachdruck der Ausgabe von
1964 unter Beigabe aller seither bekannt gewordenen Fragmente. Ed. J-U. Hartmann und H-J. Röllicke.
Düsseldorf: Haus der Japanischen Kultur.
Willemen, Charles. 1998. “New Ideas about Sarvāstivāda Abhidharma.” In: The Indian Journal of
Buddhist Studies 10: 82–94.
Willemen, Charles. 2004. From Where Did Zen Come? Dhyāna in the Early Buddhist Tradition.
Calgary: University of Calgary. Numata Lecture Series.
Willemen, Charles. 2006a. The Essence of Scholasticism. Delhi: Motilal Banarsidass.
Willemen, Charles. 2006b. “The Sanskrit Title of Harivarman’s Chengshi Lun 成實論 (T. 1646).” In:
Journal of Buddhist Studies 4: 244–249.
Willemen, Charles. 2008a. “Kumārajīva’s ‘Explanatory Discourse’ about Abhidharmic Literature.” In:
Journal of the International College for Postgraduate Buddhist Studies 12: 37–83(156–210).
Willemen, Charles. 2008b. “Buddhism’s Maritime Route to China. The Early Phase (Before 589 A.D.).”
In: Journal of Buddhist Studies 6: 83–104.
Willemen, Charles. 2009. Buddhacarita. In Praise of Buddha’s Acts. Berkeley: Numata Center for
Buddhist Translation and Research.
Willemen, Charles. 2011. “Xuanzang and Paramārtha and Prajñāpāramitā’s Origin.” In: Journal of
Buddhist Studies 9: 145–152.
Willemen, Charles. 2012a. “Mahīśāsaka. Some New Ideas.” In: Dharmapravicaya. Aspects of Buddhist
Studies.Essays in Honour of N.H. Samtani. Lalji ‘Shravak’ and Charles Willemen (eds.). Delhi: Buddhist
World Press: 481–493.
Willemen, Charles. 2012b. Outlining the Way to Reflect (T. XV 617: Siwei lüe yao fa 思惟略要法).
Mumbai: Somaiya Publications.
268 CHARLES WILLEMEN

Willemen, Charles. 2013a. A Collection of Important Odes of the Law. Berkeley: Institute of Buddhist
Studies and BDK America.
Willemen, Charles. 2013b. “Early Yogācāra and Visualization (Bhāvanā).” In: Wading into the Stream
of Wisdom: Essays Honoring Leslie Kawamura. Sarah Haynes and Michelle Sorensen (eds.). Berkeley:
Institute of Buddhist Studies and BDK America.
Wogihara, Unrai. 1971. Sphuṭārthā Abhidharmakośavyākhyā by Yaśomitra. Part I. Tokyo: Sankibo
Buddhist Book Store.
Yamabe, Nobuyoshi; Sueki, Fumihiko. 2009. The Sutra on the Contemplation of Sitting Meditation.
Berkeley: Numata Center for Buddhist Translation and Research.