The Transmission of Buddhist texts

to Tocharian Buddhism

Ogihara Hirotoshi

Introduction

Previous studies have already revealed that the school affiliation of

­Tocharian Buddhism should belong to the (Mūla-)Sarvāstivādins. Among
others, the late Prof. Kudara Kogi indicated that the tradition of the
Sarvāstivādins of the Kaśmīr region is reflected in the Udānālaṅkāra, a
commentary in Tocharian on the Skt. Udānavarga which is said to have
been compiled by Dharmasoma. In addition, my research on the Vinaya
texts in Tocharian A and B alludes to some different textual traditions of
the (Mūla-)Sarvāstivādins having been accepted by Tocharian monks at
different periods from the 5th century onwards. However, a close affinity
with the Mūla-Sarvāstivādins has rarely been detected in the fragments
found in Kucha.1‌ Yet, those fragments found in Kizil, which should have
belonged to one and the same manuscript, contain five stories that can be
assigned to the tradition of the Mūla-Sarvāstivādins, a fact which sug-
gests that the tradition of the Mūla-Sarvāstivādins would have also been
transmitted to Tocharian Buddhism in Kucha. In this paper, the following
two aspects will be discussed, on which the transmission of Buddhist
texts to Tocharian Buddhism can shed light: [1] the relationship between
the Sarvāstivādins and the Mūla-Sarvāstivādins and [2] the complexity
of the origin of Tocharian Buddhism.

1
 In this article, I use the term ‘the Mūla-Sarvāstivādins’ to refer to the Buddhist school
which held as their Vinayas the texts entitled ‘the Vinayas of Mūla-Sarvāstivādins’ known
to us. On the other hand, ‘the (Mūla-)Sarvāstivādins’ is used in a broad sense which
includes both ‘the Sarvāstivādins’ and ‘the Mūla-Sarvāstivādins’.

Journal of the International Association of Buddhist Studies

Volume 38 • 2015 • 295–312 • doi: 10.2143/JIABS.38.0.3134545

98643_JIABS.indb 295 22/03/16 07:54

 296 ogihara hirotoshi

1. Sarvāstivādins and Mūla-Sarvāstivādins in Tocharian Buddhist texts

In Chinese literature, especially the records of the journeys of those Chi-

nese monks who travelled to India, it is said that Kuchean monks in the
Tang dynasty learned the doctrine of the Sarvāstivādins and Chinese monks
learned Mahāyāna Buddhism.2‌ Chinese Buddhism was transported by the
Chinese who had been dispatched to Kucha, which was under the control
of the Tang dynasty. Considering these records and the fragments unearthed
in this region, Buddhism in Kucha can be summarized as follows:

Kuchean Buddhism Chinese Buddhism

Language(s) Sanskrit and Tocharian B Chinese
Doctrine Sarvāstivādins Mahāyāna Buddhism

The Buddhist literature in Tocharian is now housed in England, Germany,

France, Russia, China and Japan. Given that the research on Tocharian
Buddhist texts is still being performed, I am obliged to give a provisional
classification of Tocharian Buddhist literature below.3‌
[1] Āgama
[2] Vinaya
[3] Abhidharma
[4] Udānavarga, its commentary
[5] Avadāna, Jātaka, legends of Buddha
[6] Buddhastotra
[7] Buddhist drama
[8] Medical texts

2
 Cf. Da Tang xiyu ji 大唐西域記 (T.51 no. 2087, 870a24–27) and Youfangji chao 遊
方記抄 (T.51 no. 2089, 979a26–29). The Sarvāstivādins were also assigned as the school
affiliation of the Buddhism in Shorchuk in Da Tang xiyu ji (T.51 no. 2087, 870a11–13).
On the historical background of Tocharian B material under the Tang dynasty, see Ching
2013.
3
 As an introduction to Tocharian texts, see also Pinault (1989: 14–16). Basically the
same genres as those given in this classification are also observable in the Sanskrit Buddhist
literature unearthed in the Northern rim of the Tarim Basin, cf. Hartmann and Wille (1992,
1997, 2014) and Wille (2014a, 2014b). According to Hartmann (2004: 126), the texts repre-
sented by the largest number of manuscripts are the Udānavarga, the Bhikṣuprātimokṣasūtra
and the Buddhastotras of Mātṛceṭa.

98643_JIABS.indb 296 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 297

[9] Deśanā
[10] Sanskrit grammar
[11] Tantric texts
[12] Inscriptions

The school affiliation of Tocharian Buddhism was first determined by

Lévi (1912), according to which affiliation can be assigned to the Sarvās-
tivādins. His study is based on the comparison between a story appearing
in one Tocharian B Vinayavibhaṅga fragment that is said to have been
discovered at Jigdalik near Kucha with parallel texts in other languages.
This school affiliation accords with the depiction given in Da Tang xiyu ji.
Since then, Lévi’s view has been supported by other studies.4‌ Among
others, an article published by Kudara (1972) deserves mention. His
study focuses on the elucidation of the school affiliation of an Abhidharmic
discussion found in the Udānālaṅkāra [below UA], which is a commen-
tary on the Skt. Udānavarga compiled in Tocharian Buddhism [below
Udv.]. A comparison with Chinese parallels revealed that the didactic
commentary on Skt. ānāpānasmṛti- “Atmungsachtsamkeit” (SWTF I:
258a) that appears in the 15th chapter shows an affinity with the theory
held by the Sarvāstivādins in Kaśmīr. The connection between Tocharian
Buddhism and the Sarvāstivādins in Kaśmīr pointed out by Kudara is of great
importance for research on Tocharian Buddhism, because he succeeded in
elucidating to which tradition in the Sarvāstivādins, that in Kaśmīr or that
in Gandhāra, Tocharian Buddhism would have been related.5

4
 Concerning the research on the school affiliation of Tocharian Buddhism up to 2009,
see Ogihara 2009b: 14–15. Schmidt (1985) should be mentioned as a representative
among studies in which the Sarvāstivādins are assigned as the school affiliation of Toch-
arian Buddhism.
5
Most of the Buddhist terms used in Tocharian Buddhist texts come from Sanskrit
and Gāndhārī, which suggests that Tocharian Buddhism might have been under the strong
influence of Indian Buddhism, which is in contrast to Sogdian and Old Uighur Buddhism
in which the strong influence of Chinese Buddhism is observable through loan words or
the translation style. In contrast, Tamai (2013: 331) supposes that part of this commen-
tary might have been translated from a Chinese original, based on the consistency of the
content between the Tocharian B text and the Chinese text. However, to my knowledge,
none of the Tocharian Buddhist texts seem to have been translated from the Chinese
originals.

98643_JIABS.indb 297 22/03/16 07:54

 298 ogihara hirotoshi

However, an affinity with the Mūla-Sarvāstivādins can be also detected

in the narratives given in other strophes in the UA. As I have discussed
elsewhere (Ogihara 2011a, 2011b and 2012: 203), two narratives given
in B3 (Udv. I 13, 14, 16) and B21–22 (Udv. IX 9) have parallel texts in
the Bhaiṣajyavastu of the Mūla-Sarvāstivādins and the Avadānaśataka
respectively. This fact suggests that the UA may be an original work com-
piled in Tocharian Buddhism.6‌ Although it is not impossible to suppose
that these stories would have been transmitted to Tocharian Buddhism
without any connection with the Mūla-Sarvāstivādins, the fact that these
narratives that appear in the UA show a close relationship with this
school leads to the supposition that they might have been extracted
from works belonging to it . In fact, quotations from other works of this
school are also found in the Puṇyavantajātaka in Tocharian A, in which
the Āgamasūtra, corresponding to the Saṃyuktāgama (T.02 no. 99–1283,
353a21–b24) in Chinese, is quoted (Ogihara 2009a).
It is also worth mentioning that an Abhidharmic commentary in B41
that was studied by Kudara, and a narrative that is entitled Rājā in the
Avadānaśataka are attested in the same manuscript of the UA found in
Shorchuk (ms. B).7‌ Considering that those strophes of the Udv. translated
into Tocharian B in the UA are generally classified as ‘Recension 1’
– the textual tradition of the Sarvāstivādins (Nakatani 1988: 156)8‌ – the
background of the UA may not be as simple as scholars have supposed
so far.
The fact that some different traditions of the (Mūla-)Sarvāstivādins were
accepted by Tocharian monks at different periods from the 5th century
onwards can be supported by my research on the Vinaya texts in Tocha-
rian. According to my studies, most of these texts can be classified as
those of the Sarvāstivādins, while some fragments clearly show an affinity
with the tradition of the Mūla-Sarvāstivādins. Among them, those frag-
ments found in Turfan have a close affinity with the Mūla-Sarvāstivādins

6
See also Hackstein, Habata, and Bross 2014.
7
Other fragments in Tocharian A and B mentioned in the preceding paragraph were
also found in Shorchuk.
8
 Concerning the difference between ‘Recension 1’ and ‘Recension 2’ of the Udāna­
varga, see Schmithausen 1970.

98643_JIABS.indb 298 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 299

(Ogihara 2009b, 2011c).9‌ In contrast to those fragments found in Kizil and

Shorchuk, they show late linguistic and palaeographic features (Peyrot
2008: 221).10‌ The fact that the Vinaya texts of the late stage of Toch-
arian B found in Turfan are assigned to the Mūla-Sarvāstivādins can be
also confirmed by the Sanskrit fragments discovered in the same region;
many Sanskrit Vinaya texts from this region also show a close affinity
with this school. Hence, it is possible to suppose that these Vinaya texts
in Sanskrit and Tocharian B belong to the later stage.11‌ It is also note-
worthy that one can find many glosses in Old Uighur in these Tocharian
B Vinaya fragments from Turfan (Maue 2009 and 2010), which suggests
that the tradition(s) of the (Mūla-)Sarvāstivādins were transmitted to Old
Uighur Buddhism.12‌

2. Affinity with Mūla-Sarvāstivādins in Archaic Tocharian B

As mentioned above, an affinity with the Mūla-Sarvāstivādins can be

detected in Tocharian B fragments unearthed in Shorchuk and Turfan,
which reflect the classical and the late stage of Tocharian B, respectively
(Peyrot 2008: 219). Given that one Sanskrit Udv. fragment found in
Kumtura near Kucha of which the palaeographic features can be classified
as North Turkestan Brāhmī, Type A (Sander 1968: 182, 2005 [2009])
contains the text belonging to ‘Recension 2’ – the tradition of the Mūla-
Sarvāstivādins –, it is reasonable to think that among Tocharian B material
found in Kucha, there should exist fragments that show an affinity with
this school (Ogihara 2013a). However, Tocharian B fragments unearthed

9
 Christiane Schaefer (1997) first indicated the affinity between Tocharian B Vinaya
fragments found in Turfan and the Mūla-Sarvāstivādins.
10
 For the chronology of the Tocharian B language from a linguistic viewpoint, see
Peyrot 2008 : 204–206. Those fragments written in Late Tocharian B are dated 7th century
onwards. However, the written variants in Archaic Tocharian B and Classical Tocharian
B are assigned to the 5th and 6th century onwards, respectively. Concerning the palaeo-
graphical classification of Tocharian B fragments, see Malzahn 2007 and Tamai 2011.
11
 Those Sanskrit fragments unearthed in Turfan are dated to the 9th–14th centuries, see
Sander 1968: 23 and 2005 [2009].
12
 Among the Old Uighur manuscripts in Brāhmī and the Sogdian manuscripts found in
Turfan, some fragments were identified as belonging to the Vinaya of the Mūla-Sarvāsti­
vādins, see Maue 1996: 18–64 and Yoshida 2007: B51–B54, respectively.

98643_JIABS.indb 299 22/03/16 07:54

 300 ogihara hirotoshi

in Kucha have rarely shown an affinity with the Mūla-Sarvāstivādins,

although Schmidt (1998 [2000], 2010) recently identified most of the
captions that explain the mural paintings in Cave No. 110 of the Kizil
grottoes, as reflecting the tradition of the Mūla-Sarvāstivādins.
Among the Tocharian Vinaya texts found in Kucha, some fragments
show an affinity with this school. Nevertheless, none has been identified
as the Vinaya text of this school. For instance, B333, which I identified as
the Upasaṃpadā of the Karmavācanā, contains one passage that is pecu-
liar to the textual tradition of the Mūla-Sarvāstivādins (Ogihara 2013b).
B333b1–2: ku se no ṣamāne mā ce warñai kca klyomots wāki yainmu sakwä
täkä[lyñe] (wṣīl)[yñ](e pālo)i tär aiykemar wat te maṃt lkāskewat te maṃnt …
A monk who, not yet having reached any excellence of the noble ones,
praises [falsely] his condition of dwelling with things that are pleasant
to touch, [saying] either thus, “I know” or thus, “I see” …
Skt: yaḥ punar bhikṣur anabhijānann aparijānann asantam asaṃvidya­
mānam anuttaramanuṣyadharmam alamāryaviśeṣādhigamaṃ jñānaṃ vā
darśanaṃ vā sparśavihāratāṃ vā pratijānīyād idaṃ jānāmīdaṃ paśyāmīti …
(Banerjee 1977: 15)
Whatever monk, unknowing and not understanding, should boast of having
superhuman faculties, sufficient knowledge and the specific spiritual realiza-
tion of the nobles and insight and a state of comfort which are inexistent and
unobtained [by him], saying, “I know this, I see this” … (Prebish 1975: 53)13‌

Here, Toch.B sakwä täkälyñe wṣīlyñe, a calque of Skt. sukhasaṃsparśa-

vihāratā- “state of dwelling in pleasant touch” (BHSD 597a), could
translate Skt. sparśa-vihāratā- “state of comfort” (BHSD 612a), which
is attested to only in the Prātimokṣasūtra of the Mūla-Sarvāstivādins.
Although the text of B333 accords well with that of the Sarvāstivādins,
we have a phraseology that can be found only in the Vinaya text of the
Mūla-Sarvāstivādins in the prescription itself. It is noteworthy that B333
can be classified as Archaic Tocharian B, which is datable to sometime
before the 6th century AD.

13
 The tradition of the Sarvāstivādins runs as follows: yaḥ punar bhikṣur anabhijāna(n)n
aparijānann uttarimanuṣyadharmam alamāryaviśeṣādhigamaṃ jñānaṃ vā darśanaṃ vā
pratijānīyāj jānāmīti paśyāmīti … “Wenn ein Mönch, ohne tiefgehende und genaue
Kenntnisse zu haben, mit den Worten “ich erkenne, ich sehe” behauptet, übermenschliches
Wissen oder übermenschliche Einsicht zu haben, …” (Simson 2000: 164–165, 271).

98643_JIABS.indb 300 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 301

In my view, this can be explained in the following way. There might

have existed some traditions with slight differences among the Sarvāstivā­
dins, one of which could have developed into the so-called Mūla-Sarvās-
tivādins, and the Tocharian Vinaya texts could reflect this complexity
within this school. Enomoto (1998, 2000, 2004) assumes that basically
the Sarvāstivādins and the Mūla-Sarvāstivādins would have stemmed
from the same school, and there would have existed two or more different
Vinaya texts in this school.14‌ Tocharian Vinaya texts could confirm this
theory.15‌
In genres other than the Vinaya texts, fragments of Sūtras, Jātakas,
verses and medical texts are at our disposal in archaic Tocharian B, but
none of them has been identified as belonging to the literature of the
Mūla-Sarvāstivādins.

3. ‘Avadāna Manuscript’ in Tocharian B

The fragments introduced so far show that Tocharian Buddhism should

have been composed of some different traditions. At least, different tra-
ditions in the (Mūla-)Sarvāstivādins can be detected. However, these
fragments belong to various genres and came from various areas ranging
from Kucha to Turfan. In addition, linguistic and palaeographic stages
also range from the archaic to the late stage of Tocharian B. To show the
complexity in Tocharian Buddhism more clearly, I would like to intro-
duce another example in which the same story was transmitted to Toch-
arian Buddhism twice in different periods. Here, one Avadāna appeared
in Sanskrit and Tocharian B fragments, both of which were discovered
in Kizil, represent two different linguistic and palaeographic stages.

14
 This has been already indicated on the basis of the Sanskrit fragments discovered in
Central Asia, cf. Simson 2000: 2–13. See also Sasaki 2000 and Clarke 2001.
15
 In fact, there are other fragments that show the Mūla-Sarvāstivādin affiliation in the
Tocharian Karmavācanā texts (= IOL Toch 139 and IOL Toch 1148), see Ogihara 2011c:
124, n. 21 and 128–130. However, Tocharian B sthūlāñcana, the plural form of *sthūlāñca,
which was borrowed from *(s)thūlāñca reconstructed on the basis of 偷蘭遮 tou lan zhe
(= EMC *thəw lan tɕia), is attested in IOL Toch 139a7. Yet, the term 偷蘭遮 is used in
the Vinayavibhaṅga that belongs to the Sarvāstivādins. The provenance of IOL Toch 139
is thought to be around Kucha, cf. Skjærvø 2003: xli–xlii. It has been examined in Ogihara
2009b: 158 and 465–467. On IOL Toch 1148b3, see Ogihara 2011c: 128–130.

98643_JIABS.indb 301 22/03/16 07:54

 302 ogihara hirotoshi

Among Tocharian B material found in Kizil, 47 fragments can be

c­ onstrued as belonging to the same manuscript on linguistic and palaeo-
graphic grounds. Given that five identified stories contain Avadānas
or Jātakas, I have named this manuscript the ‘Avadāna manuscript’ (Ogi-
hara 2012). Actually, one finds the following passages at the end of the
story:
THT1166+2976a4: jotiṣ[k]e[n]tsa apadāṃ āra, “The Avadāna by Jyotiṣka
has finished.”
THT1683b3: ṣanmīrśkaṃntsa dṛṣṭānt āra, “The paragon by young novices
has finished.”

The former is given at the end of the Jyotiṣkāvadāna, while the latter remains
unidentified. Below is the list of this manuscript as reconstructed by me.
[Avadāna manuscript]
− THT1165+1548 [folio <43>], THT1166+2976 recto: Jyotiṣkāvadāna
− THT1166+2976 verso, THT1556: Śroṇakoṭiviṃśāvadāna
− THT1253+3056: Dhanikāvadāna
− THT1551 [folio<187>], THT1683 [folio<190>]: (?)
− THT1285+1507+1680+2981.frg.1+3054: Sāraṇa
− THT1249+1681+2981.frg.6 [folio <287>]: Sāraṇa
− THT1168+3034: Padma
− THT1554+3112: Kacchapajātaka (?)
− THT1245+2996+2999: (?)

This list only enumerates those fragments for which the content or its
position in the manuscript was identified. Although 20 fragments still
remain unclear, five identified stories have their parallel texts in the works
of the Mūla-Sarvāstivādins: Vinayakṣudraka or Divyāvadāna [No. 19]
(Jyotiṣkāvadāna), Saṅghabhedavastu (Śroṇakoṭiviṃśāvadāna), Bhaiṣa­
jyavastu (Dhanikāvadāna), Kalpanāmaṇḍitikā [No. 65] or Karmaśataka
[No. 89] (Sāraṇa) and Avadānaśataka [No. 7] (Padma).16‌ This identification

16
 The ‘Avadāna manuscript’ shows that monks in Tocharian Buddhism could have
compiled their own Buddhist works. However, it still remains unclear why these stories
were selected to be compiled in one work and how they were arranged. For instance, the
Jyotiṣkāvadāna (THT1165+1548 [folio <43>], THT1166+2976 recto) and the Śroṇako­
ṭiviṃśāvadāna (THT1166+2976 verso, THT1556) are given as one unit. This structure
may be because of the fact that they are the stories of Buddha’s disciples. However, the

98643_JIABS.indb 302 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 303

clearly shows that the tradition of this school was also known to Tocharian
monks in Kizil. Because this manuscript can be classified as classical-late
Tocharian B on linguistic and palaeographic grounds, it would belong to a
relatively later stage of Tocharian B material found in Kizil, probably dat-
able to the 7th–8th century (Peyrot 2008: 204–206).
The style of these stories is simpler than their parallels in other lan-
guages, which indicates that they could be an adaptation of these stories
in Tocharian Buddhism. It is also noteworthy that no compilation in San-
skrit of Avadānas or Jātakas is known that has the same stories as the
‘Avadāna manuscript’ in Tocharian B, which suggests the possibility that
the ‘Avadāna manuscript’ was compiled within Tocharian Buddhism or
was based on a lost compilation in Sanskrit.
Among these stories, one entitled Sāraṇa will be introduced below.
This story exists in Sanskrit, Chinese and Tibetan versions, among which
the two latter versions give us the entire story.
[Sanskrit]
SHT21.80, Bl. 214 (Kalpanāmaṇḍitikā): Lüders (1926: 182–183 [1979:
300–301])
Avadāna Anthology from Merv, Fol. 32: Karashima and Vorobyova-Desya-
tovskaya (2015: 226–229)
[Chinese]
Da zhuangyan lun jing 大莊嚴論經: vol. 12, Chap. 65 (T.04 no. 201,
323c4–326b13)17‌
Za baozang jing 雜寶藏經: vol. 2, Chap. 24 (T.04 no. 203, 459a21–c23)18‌
[Tibetan]
Karmaśataka: Chap. 89 Sa-ra-na (D:a22b4–26b6)19‌

passage tū mäkte nesäṃ śāstärne aklyilñe […] “How is it (to) … the teaching in (sacred)
book?” appears as an introductory phrase three times (THT1253+3056a1, 1551b2 and
1683b4) in the ‘Avadāna manuscript.’ The leaf number of the latter two fragments sug-
gests that they were also arranged as one unit, to which the other two fragments
(THT1253+3056) that were identified as the Dhanikāvadāna were added. However, it still
remains unclear what role these stories played in this work. In short, the arrangement of
Avadānas or Jātakas in these works would have been based on an interpretation of these
stories that could reflect the peculiarity of Tocharian Buddhism.
17
 For the French translation of this story, see Huber 1908: 342–355.
18
 For a summary of this story, see Chavannes 1911: 23.
19
 For a summary of this story, see Feer 1901: 439–441.

98643_JIABS.indb 303 22/03/16 07:54

 304 ogihara hirotoshi

In fact, the Sanskrit version, of which palm leaf fragments were also
discovered in Kizil by the German Turfan Expedition, is the original of
one of the Chinese versions mentioned above, which is Da zhuangyan lun
jing 大荘厳論経 translated by Kumārajīva. The palaeographic feature,
Schrifttypus I–II, Kuṣāṇa/Indische Gupta-Schrift, points to the 4th century
as the date of this Sanskrit manuscript.20‌ Lévi (1908) indicated that par-
allel texts of this work are found in literature belonging to the (Mūla-)
Sarvāstivādins.
The outline of this story is as follows: Prince Sāraṇa, son of King
Suvīra, became a monk without ascending the throne.21‌ One day, he went
to a garden in the country of King Pradyota, where he sat in meditation.
Pradyota also went there with women of his court. Having met Sāraṇa in
search of fruits, these women liked him and sat around him. On seeing
Sāraṇa sitting with them under a tree, Pradyota became angry and asked
him his rank. When Sāraṇa told him that he had not achieved any rank,
Pradyota ordered his subordinates to beat him with sticks bitterly. When
Sāraṇa returned to the place where Kātyāyana,22‌ his master, was staying,
he decided to abandon monkhood to avenge himself against King
Pradyota. That night, Kātyāyana used his supernatural power to make
Sāraṇa dream a terrible dream, in which he was defeated by Pradyota and
was being led to an execution ground. On wakening up from this dream,
Sāraṇa abandoned his wish for revenge.
Although the Tocharian B version restored by me lacks the beginning
and the concluding part of this story, we can still compare it with other
versions. Such a comparison reveals that the Tocharian B version tells
the same story as the other versions in spite of it being much simpler than
other versions. However, the following passage deserves mention:
THT1507+1680+2981.frg.1a3: āklä(sk)[au](ne) ṣ lānte kercyeṃnne am[p]-
lākätte mā yänmaṣle.
“And I teach him: one should not enter into the royal palace without per-
mission.”

 Cf. SHT I: 14–15 and Sander 1968: 51–136, 2005 [2009].

20

Sāraṇa became a monk before his father’s death in the Karmaśataka.

21
22
 The name of Sāraṇa’s master is Kātyāyanīputra (Tib. Kā-tyā-ya-na’i bu) in the Kar­
maśataka.

98643_JIABS.indb 304 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 305

This passage appears in the instruction given by Pradyota to his subor-

dinates after arresting Sāraṇa. It leads us to think that Sāraṇa entered
the royal palace of Pradyota without his permission, where this event
happened. Yet, this instruction is not mentioned in other versions, which
suggests that it would have been added in the course of the compilation
within Tocharian Buddhism. In contrast to the Tocharian B version, the
Sanskrit and two Chinese versions only mention that Sāraṇa entered the
garden in King Pradyota’s country.
Bl.214a1–2: /// sa kadāc(i)t = pra[d]y(ota) …… m = anuprāpta[ḥ] tatra
vivikta u[d]y(āne) ///
One time, he went to the (country?) of Pradyota. There he (was) alone in the
garden …
Da zhuangyan lun jing 大莊嚴論經: 既出家已，隨尊者迦旃延，詣巴樹
提王國，在彼林中住止。
Having become a monk, he went to the country of Pradyota with Kātyāyana
and was staying in the garden … (T.04 no. 201, 323c7–9)
Za baozang jing 雜寶藏經: 名曰娑羅那，心樂佛法，出家學道，頭陀苦
行，山林樹下，坐禪繫念。
(A prince) named Sāraṇa enjoyed Buddha’s dharma(s) and became a monk.
While performing dhūtaguṇas, he was staying under a tree in a garden and
sitting in meditation. (T.04 no. 203, 459a22–23)

However, the Karmaśataka in Tibetan agrees with the Tocharian B version

in this respect.23‌
Karmaśataka: rgyal po Gtum-po-rab-snang gi pho brang gi chab sgor
phyin nas chab sgo ba rnams med pa’i skabs su de rgyal po’i pho brang gi
nang du song nas (D:a24a1)
“Having come to the gate of King Pracaṇḍaprabha’s (= Pradyota’s) castle,
the door-keepers were not there. At that time he entered into the castle of
that king …”

23
 The Avadāna Anthology from Merv Fol. 32 recto 4–5 tells sa ca rājño Pradyotasya
antaḥpuraṃ praviṣṭo rājñā Pradyotena kaśābhiḥ tāḍitaḥ “He entered the women’s quarter
of the palace of King Pradyota and was beaten with a whip” (Karashima and Vorobyo-
va-Desyatovskaya 2015: 226–227). If Skt. antaḥpura- in this passage means “the king’s
palace” (MW 43a), the story appearing in this anthology which belongs to the Sarvāstivāda
tradition and can be dated 5th century agrees well with the Tocharian B version.

98643_JIABS.indb 305 22/03/16 07:54

 306 ogihara hirotoshi

This similarity leads to the supposition that the Tocharian B version would
have been based on the later tradition of this story which is similar to that
in the Karmaśataka, which belongs to the Mūla-Sarvāstivādins, although
the exact date of the compilation of this work cannot be determined.24‌
As shown above, this story was transmitted to Tocharian Buddhism at
least twice in different periods. Given that the parallels of the Karmaśataka
can be confirmed in the literature belonging to the Mūla-Sarvāstivādins,
Tocharian Buddhism in Kizil could have accepted mainly two different
waves from India – the first being that of the Sarvāstivādins and the second
that of the Mūla-Sarvāstivādins. This supposition can be supported by studies
on the mural paintings found in the Kizil grottoes, in which two different
styles are discernible: one reflects the Sarvāstivādins and the other
reflects the Mūla-Sarvāstivādins (Hiyama 2013).
In this connection, the Subhāṣitagaveṣinjātaka [Avadānaśataka No. 38]
should be mentioned. This story is available in Sanskrit, Chinese and Tibetan
versions, and mural paintings depicting this story were also found in some
cave temples in Kucha, which reflect the story’s popularity within Tocharian
Buddhism. These parallels can be classified in two groups, [A]: that of the
Mūla-Sarvāstivādins, namely that given in the Avadānaśataka and the Vinaya
of the Mūla-Sarvāstivādins (in Skt., Chi. and Tib.) and [B]: that available in
two Chinese versions, Xianyu jing 賢愚經 and Zhuanji baiyuan jing 撰集百
緣經. However, the Tocharian B version given in those fragments found in
Shorchuk accords neither with group [A] nor with group [B]. In this story,
the god Indra, incarnated as Yakṣa, informs King Subhāṣitagaveṣin of the
Buddha’s teaching under the condition that he should dive into the fire. The
core of the Tocharian B version accords well with group [A] while it agrees
with group [B] in that Indra taught the king a verse before he dived into the
fire. Thus, the story shows an intermediate status between these two groups
that can also reflect the complexity of Tocharian Buddhism (Ogihara 2013c).
However, Buddhist works belonging to other schools may have been
known to Tocharian Buddhism in view of the fact that some Sanskrit frag-
ments unearthed in Kizil, Duldur-akhur and Murtuq have been identified
as Vinaya texts belonging to the Dharmaguptakas (Chung and Wille 1997,
Wille 1997, and Simson 2000: 153–154).

24
On the textual sources of the Karmaśataka, see Matsumoto 2001.

98643_JIABS.indb 306 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 307

As shown above, Tocharian Buddhism reflected in Tocharian B liter-

ature could have been composed of different textual traditions of the
(Mūla-)Sarvāstivādins at different periods and probably from different
areas. The same situation can also be detected in Tocharian A literature,
which likewise reflects the influence of the Mūla-Sarvāstivādins with
differences of various degrees (Ogihara 2009a, forthcoming). For instance,
the Mūgapakkhajātaka in Tocharian A shows a close relationship with
the tradition of the Mūla-Sarvāstivādins in that Prince Mūgapakkha, a
hero of this story, became a monk with five hundred sons of ministers of
his father’s court (Sieg 1952: 17–26). ‌

Conclusion

By way of a conclusion to this paper, the following points should be

mentioned.
(1): The transmission of Buddhist texts that is discussed above shows
clearly that some different textual traditions of the (Mūla-)Sarvāstivādins
were transmitted to Tocharian Buddhism at different periods, and they
could have originated from different areas. Xuanzang assigned the
Sarvāstivādins to the school affiliation of Tocharian Buddhism in the
7th century. As elucidated by Enomoto (1998, 2000, 2004), the Sarvās-
tivādins mentioned in some Buddhist works, including Da Tang xiyu ji
by Xuanzang, are actually the (Mūla-)Sarvāstivādins. This supposition
can also be supported by Tocharian Buddhist texts. Thus, Tocharian Bud-
dhism should be investigated from the perspective of the historical and
philosophical development among the (Mūla-)Sarvāstivādins.
In contrast, one seldom finds the Vinaya texts of the Mūla-Sarvāstivā-
dins among the manuscript remains in Sanskrit or in Tocharian that have
been unearthed in Kucha and Shorchuk, although strong influences from
the Mūla-Sarvāstivādins are observable in these areas. This fact leads us
to research the historical and philosophical background under which those
Buddhist works of the Mūla-Sarvāstivādins that are now available to us
were composed by the (Mūla-)Sarvāstivādins and transmitted eastward.
(2): Buddhist works belonging to the (Mūla-)Sarvāstivādins would have
developed some different versions in different periods and different

98643_JIABS.indb 307 22/03/16 07:54

 308 ogihara hirotoshi

areas, which seems to be reflected in Tocharian Buddhist works. In that

case, the manner in which these traditions with different origins were
accepted by Tocharian monks as belonging to the same school should be
questioned. As is well known, there were two large groups within the
(Mūla-)Sarvāstivādins, one in Kaśmīr and the other in Gandhāra. If Toch-
arian Buddhism accepted these two traditions, it should be investigated
how Buddhist monks at that time recognized these different groups as
belonging to the same school. Thus, this leads to another question, that
is, how Buddhist schools can be defined. There still remains much
work to do in considering these questions, and more materials should be
consulted.

Abbreviations
BHSD Edgerton, Franklin. 1953. Buddhist Hybrid Sanskrit Grammar and
Dictionary. 2 vols. New Haven: Yale University Press.
D Derge edition of the Tibetan Kanjur and Tanjur
EMC Early Middle Chinese reconstructed by Pulleyblank, E.G. 1991.
Lexicon of reconstructed pronunciation in Early Middle Chinese,
Late Middle Chinese, and Early Mandarin. Vancouver: UBC Press.
IOL Toch India Office Library, Tocharian manuscripts
MW Monier-Williams, Monier. 1899. Sanskrit-English Dictionary. Oxford:
Clarendon Press.
SHT I Waldschmidt, Ernst. 1965. Sanskrithandschriften aus den Turfan­
funden. Teil I. Wiesbaden: Franz Steiner.
SWTF I Waldschmidt, Ernst, et al. 1973. Sanskrit-Wörterbuch der buddhis­
tischen Texte aus den Turfan-Funden [und der kanonischen Literatur
der Sarvāstivāda-Schule]. Begonnen von Ernst Waldschmidt. Im
Auftrage der Akademie der Wissenschaften zu Göttingen. Göttingen:
Vandenhoeck & Ruprecht.
T. Taishō Tripiṭaka.
THT Tocharian Manuscripts from the Berlin Collections
UA Udānālaṅkāra
Udv. Udānavarga

References
Banerjee, Anukul Chandra. 1977. Two Buddhist Vinaya Texts in Sanskrit: Prāti­
mokṣa Sūtra and Bhikṣukarmavākya. Calcutta: The World Press Private
Limited.

98643_JIABS.indb 308 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 309

Chavannes, Édouard. 1911. Cinq cents contes et apologues: extraits du Tripiṭaka

chinois, et traduits en français. Tome III. Paris: Ernest Leroux.
Ching, Chao-jung. 2013. “The Kuchean kings seen in the inscriptions found in
Kucha Grottoes.” Journal of the Dunhuang and Turfan Studies XIII: 387–
418 [in Chinese].
Chung, Jin-il and Klaus Wille. 1997. “Einige Bhikṣuvinayavibhaṅga-Fragmente
der Dharmaguptakas in der Sammlung Pelliot.” In Heinz Bechert, Sven
Bretfeld, and Petra Kieffer-Pülz, eds., Untersuchungen zur buddhistischen
Literatur II. Sanskrit-Wörterbuch der buddhistischen Texte aus den Turfan-
Funden, Beiheft 8. Göttingen: Vandenhoeck & Ruprecht: 47–94.
Clarke, Shayne. 2001. “The Mūla-sarvāstivāda Vinaya Muktata.” Buddhist Stud­
ies 30: 81–107.
Enomoto, Fumio. 1998. “Mūlasarvāstivādin and Sarvāstivādin.” Journal of
Indian and Buddhist Studies 47/1: 111–119 [in Japanese].
—. 2000. “Mūlasarvāstivādin” and “Sarvāstivādin.” In Ch. Chojnacki, J.-U. Hart-
mann and V.M. Tschannerl, eds., Vividharatnakaraṇḍaka: Festgabe für
Adelheid Mette. Indica et Tibetica, Volume 37. Swisttal-Odendorf: Indica et
Tibetica Verlag: 239–250.
—. 2004. “Emergence of ‘Mūla-Sarvāstivādin’.” In Studies on Indian Philosophy
and Buddhist Thoughts, Volume in Honor of Professor Esho Mikogami.
Kyoto: Nagata-bunshodo: 651–677 [in Japanese].
Feer, M.L. 1901. “Le Karma-çataka.” Journal Asiatique. 9/17: 53–100, 257–315,
410–486.
Hackstein, Olav, Habata Hiromi, and Christoph Bross. 2014. “Tocharische Texte
zur Buddhalegende I Die Geschichte von der Nonne Sundarī B15–18.”
Münchener Studien zur Sprachwissenschaft 68/1: 31–99.
Hartmann, Jens-Uwe. 2004. “On the Relationship between the Buddhist Sanskrit
Texts from Northern Turkestan and those from Afghanistan.” In Desmond
Durkin-Meisterernst, Simone-Christiane Raschmann, et al., eds., Turfan
Revised – The First Century of Research into the Arts and Cultures of the
Silk Road. Berlin: Dietrich Reimer Verlag: 125–128.
Hartmann, Jens-Uwe and Klaus Wille. 1992. “Die nordturkestanischen Sanskrit-
Handschriften der Sammlung Hoernle (Funde buddhistischer Sanskrit-Hand-
schriften, II).” In Jens-Uwe Hartmann, Klaus Wille, Claus Vogel, and Günter
Grönbold, eds., Sanskrit-Texte aus dem buddhistischen Kanon: Neuentdeck­
ungen und Neueditionen II. Sanskrit-Wörterbuch der buddhistischen Texte aus
den Turfan-Funden, Beiheft 4. Göttingen: Vandenhoeck & Ruprecht: 9–63.
—. 1997. “Die nordturkestanischen Sanskrit-Handschriften der Sammlung Pel-
liot (Funde buddhistischer Sanskrit-Handschriften, IV).” In Heinz Bechert,
Sven Bretfeld, and Petra Kieffer-Pülz, eds., Untersuchungen zur buddhisti­
schen Literatur: Zweite Folge: Gustav Roth zum 80. Geburtstag gewidmet.
Sanskrit-Wörterbuch der buddhistischen Texte aus den Turfan-Funden, Bei-
heft 8. Göttingen: Vandenhoeck & Ruprecht: 131–182.

98643_JIABS.indb 309 22/03/16 07:54

 310 ogihara hirotoshi

—. 2014. “The Central Asian fragments in the Pelliot collection (Paris).” In

Paul Harrison and Jens-Uwe Hartmann, eds., From Birch Bark to Digital
Data: Recent Advances in Buddhist manuscripts Research. Vienna: Verlag
der Österreichischen Akademie der Wissenschaften: 213–222.
Hiyama, Satomi. 2013. “The Buddha’s sermon scenes in Kizil Cave 207 (‘Paint-
ers’ Cave’).” Paper presented at the workshop Reading Outside the Lines:
A Workshop on the Intersection of Buddhist Art and Text, Ludwig-Maxi-
milians-Universität München, September 15, 2013.
Huber, Édourard. 1908. Açvaghoṣa, Sūtrāṃkāra. Traduit en français sur la ver­
sion chinoise de Kumârajîva. Paris: Ernest Leroux.
Karashima, Seishi and Margarita I. Vorobyova-Desyatovskaya. 2015. Buddhist
Manuscripts from Central Asia: The St. Petersburg Sanskrit Fragments.
Volume I. Tokyo: International Research Institute for Advanced Buddhology,
Soka University: 145–523.
Kudara, Kogi. 1972. “Abhidharma-Annotationen der Udānālaṃkāra im Tocha­
rischen.” Studies in Buddhism 29: 37–62 [in Japanese].
Lévi, Sylvain. 1908. “Açvaghoṣa, Le sûtrâlaṃkâra et ses sources.” Journal Asia­
tique 10/12: 57–184.
—. 1912. “Un fragment tokharien du Vinaya des Sarvāstivādins (Collection
Hoernle N° 149.4).” Journal Asiatique 10/19: 101–111.
Lüders, Heinrich. 1926 Bruchstücke der Kalpanāmaṇḍitikā des Kumaralāta.
Leipzig: Brockhaus.
Lüders, Heinrich. 1979. Bruchstücke buddhistischer Dramen; Bruchstücke der
Kalpanāmaṇḍitikā des Kumaralāta. Wiesbaden: F. Steiner.
Malzahn, Melanie. 2007. “The most archaic manuscripts of Tocharian B and the
varieties of the Tocharian B language.” In Melanie Malzahn, ed., Instru­
menta Tocharica. Heidelberg: Winter: 255–297.
Matsumoto, Junko. 2001. “Karmaśataka no kenkyū – Chibettogoyaku Karmaśa-
taka ni miru Avadāna bunken no sēritsu katē” Karmaśataka の研究–チベ
ット語訳 Karmaśataka にみるアヴァダーナ文献の成立過程. Disserta-
tion, Tōhoku University.
Maue, Dieter. 1996. Alttürkische Handschriften. Teil 1. Dokumente in Brāhmī
und Tibetische Schrift. VOHD, Volume XIII, 9. Stuttgart: Franz Steiner.
—. 2009. “Uigurisches in Brāhmī in nicht-uigurischen Brāhmī-Handschriften.”
Acta Orientalia Academiae Scientarum Hungaricae 62/1: 1–36.
—. 2010. “Uigurisches in Brāhmī in nicht-uigurischen Brāhmī-Handschriften.
Teil II.” Acta Orientalia Academiae Scientarum Hungaricae 63/3: 319–361.
Nakatani, Hideaki. 1988. Subashi shahon no kenkyū スバシ写本の研究. Kyoto:
Jinbun Sho’in.
Ogihara, Hirotoshi. 2009a. “Sur la citation de l’Āgama-sūtra dans le «Puṇyavan­
ta-Jātaka» en tokharien A.” Tokyo University Linguistics Papers 28: 133–171
[in Japanese].
—. 2009b. “Researches about Vinaya-texts in Tocharian A and B.” Dissertation,
École Pratique des Hautes Études).

98643_JIABS.indb 310 22/03/16 07:54

 transmission of buddhist texts to tocharian buddhism 311

—. 2011a. “On the quotation of an Avadāna text in the Udānālaṅkāra in Toch-

arian B.” Tokyo University Linguistics Papers 31: 213–233 [in Japanese].
—. 2011b. “Sanskrit fragments identified as the Ālánnàjīng in SHT.” Tokyo
University Linguistics Papers 31: 235–268 [in Japanese].
—. 2011c. “Notes on some Tocharian Vinaya fragments in the London and Paris
collections.” Tocharian and Indo-European Studies 12: 111–144.
—. 2012. “The ‘Avadāna manuscript’ in Tocharian B.” Tokyo University Lin­
guistics Papers 32: 109–243 [in Japanese].
—. 2013a. “A study on the Sanskrit Udānavarga from Kucha.” Xiyu yanjiu
(The Western Regions Studies) 2013/1: 70–77 [in Chinese].
—. 2013b. “Tocharian fragment THT333 in the Berlin collection.” Tokyo Uni­
versity Linguistics Papers 33: 205–217.
—. 2013c. “Liyong hanyi fojing yanjiu chutu huyu fojiao wenxian―yi qiuciyu
wenxian zhong suo jian ‘qiumiaofawang’ gushi wei li, 利用漢譯佛經研究
出土胡語佛教文獻―以龜兹語文獻中所見《求妙法王》故事為例.” In
School of Chinese Classics, Renmin University of China, ed., Guoxue de
chuancheng yu chuangxin, 国学的传承与创新. Shanghai: Shanghai Chinese
Classics Publishing House: 1121–1139.
—. Forthcoming. “Lists of former Buddhas in Tocharian texts – a comparative
study of unearthed manuscripts and captions in Kumtura Cave No. 34.” In
Zhu Yuqi, ed., Xiyu Wenshi (Literature & History of the Western Regions).
Vol. 9 [in Chinese].
Peyrot, Michaël. 2008. Variation and Change in Tocharian B. Amsterdam: Rodopi.
Pinault, Georges-Jean. 1989. “Introduction au tokharien.” Lalies 7. Actes de la ses­
sion de linguistique d’Aussois (27 août-1er septembre 1985). Paris: ENS: 5–224.
Prebish, S. Charles. 1975. Buddhist Monastic Discipline: The Sanskrit Prāti­
mokṣa Sūtra of the Mahāsāṃghika and Mūlasarvāstivādins. London: The
Pennsylvania State University Press.
Sander, Lore. 1968. Paläographisches zu den Sanskrithandschriften der Berliner
Turfansammlung. Wiesbaden: Franz Steiner.
—. 2005 [2009]. “Remarks on the formal brāhmī script from the southern Silk
Route.” BAI 19: 133–144.
Sasaki, Shizuka. 2000. “Vinayas Quoted in the Vibhāṣā.” Journal of Indian and
Buddhist Studies 49/1: 413–421 [in Japanese].
Schaefer, Christiane. 1997. “waṣik kälpaṣṣuki: Zu den westtocharischen Nominal­
bildungen auf -uki.” Tocharian and Indo-European Studies 7: 163–176.
Schmidt, Klaus T. 1985. “Zur Frage der Schulzugehörigkeit des in tocharischer
Sprache überlieferten buddhistischen Schrifttums.” In Heinz Bechert, ed.,
Zur Schulzugehörigkeit von Werken der Hīnayāna-Literatur. Erster Teil
(Symposium zur Buddhismusforschung, III, 1). Göttingen: Vandenhoeck &
Ruprecht: 275–282.
—. 1998 [2000]. “Interdisciplinary research on Central Asia: The decipherment
of the West Tocharian Captions of a cycle of mural paintings of the life of
the Buddha in Cave 110 in Qizil.” Die Sprache 40/1: 72–81.

98643_JIABS.indb 311 22/03/16 07:54

 312 ogihara hirotoshi

—. 2010. “Die Entzifferung der westtocharischen Überschriften zu einem

Bilderzyklus des Buddhalebens in der “Treppenhöhle” (Höhle 110) in
Kizil.” In Eli Franco and Monika Zin, eds., From Turfan to Ajanta, Fest­
schrift for Dieter Schlingloff on the occasion of his Eightieth Birthday.
Vol. II. Lumbini: Lumbini International Research Institute: 835–866.
Schmithausen, Lambert. 1970. “Zu den Rezensionen des Udānavargaḥ.” WZKSO
XIV: 47–124.
Sieg, Emil. 1952. Übersetzungen aus dem Tocharischen II. Aus den Nachlaß
herausgegeben von Werner Thomas. Abhandlungen der Deutschen Akade-
mie der Wissenschaften zu Berlin, Klasse für Sprachen, Literatur und Kunst,
Jahrgang 1951, Nr. 1. Berlin: Akademie-Verlag.
Simson, Georg von. 2000. Prātimokṣasūtra der Sarvāstivādins. Teil II: Kritische
Textausgabe, Übersetzung, Wortindex. Sanskrittexte aus den Turfanfunden,
Volume XI. Göttingen: Vandenhoeck & Ruprecht.
Skjærvø, Prods O. 2003. Khotanese Manuscripts from Chinese Turkestan in the
British Library: a complete catalogue with texts and translations. Reprinted
with corrections. London: British Library.
Tamai, Tatsushi. 2011. Paläographische Untersuchungen zum B-Tocharischen.
Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck.
—. 2013. “The Tocharian Udānālaṃkāra.” Annual Report of the International
Research Institute for Advanced Buddhology at Soka University 16: 315–336.
Wille, Klaus. 1997. “Zwei kleine Fragmente aus dem Bhikṣuṇīprātimokṣasūtra.”
In Heinz Bechert, Sven Bretfeld, and Petra Kieffer-Pülz, eds., Untersuchun­
gen zur buddhistischen Literatur II. Sanskrit-Wörterbuch der buddhistischen
Texte aus den Turfan-Funden, Beiheft 8. Göttingen: Vandenhoeck &
Ruprecht: 307–314.
—. 2014a. “Survey of the Sanskrit manuscripts in the Turfan collection (Berlin).”
In Paul Harrison and Jens-Uwe Hartmann, eds., From Birch Bark to Digital
Data: Recent Advances in Buddhist manuscripts Research. Vienna: Verlag
der Österreichischen Akademie der Wissenschaften: 187–211.
—. 2014b. “Survey of the identified Sanskrit manuscripts in the Hoernle, Stein,
and Skrine collections of the British Library (London).” In Paul Harrison
and Jens-Uwe Hartmann, eds., From Birch Bark to Digital Data: Recent
Advances in Buddhist manuscripts Research. Vienna: Verlag der Österrei­
chischen Akademie der Wissenschaften: 223–246.
Xinjiang Kuche Institute of Grottoes. 2000. Comprehensive Record of Contents
of Kizil Grottoes. Urumqi: Xinjiang Meishu Sheying Chubanshe.
Yoshida, Yutaka. 2007. “Report of the unpublished Buddhist Sogdian texts of
the German Turfan Collection and discussion on the origin of a Sogdian word
for ‘bodhisattva’ (Commemoration of the late Professor Kogi KUDARA).”
Studies in Buddhism 62/63: B46–B87 [in Japanese].

98643_JIABS.indb 312 22/03/16 07:54