Articles by alphabetic order
 Ā Ī Ñ Ś Ū Ö Ō
1 2 3 4 5 6 7 8 9 0

The antarabhava dispute among Abhidharma traditions and the list of anagamins

From Chinese Buddhist Encyclopedia
Jump to navigation Jump to search

by Qian Lin

1. Introduction

Indian Buddhists would appear to have had a passion for history that is somewhat unusual in the milieu of ancient Indian culture. Early Indian Buddhist histories recount not only the spread of Buddhism throughout the Indian subcontinent but also the relationships among the various Buddhist schools that came to be recognized by the tradition. Early Buddhist scholiasts, as the compiler suggests in the opening verses of Vasumitra’s Samayahhedoparacanacakra (translated by Xuanzang in the 7th c. CE), state that the purpose of

such histories that chronicle the tradition’s past is “to distinguish the gold of the true teachings of the Buddha from the sands of wrong teachings of those sectarian teachers.”1 Several early his- torical texts preserved in Pali, Chinese, and Tibetan affiliated with different Buddhist schools offer sectarian maps of early Indian Buddhism. However, when comparing these accounts, as Etienne Lamotte has done in his survey of all available historical texts, we find that they do not match and, in fact, often contradict each other regarding the origin, development and even the distinctive character of Buddhist schools. One potential source of knowledge concerning Buddhist schools are inscriptions discovered at archae­ological sites with references to schools’ names. Although these

1 Taisho no. 2031 (^1) T49, 15al6: .. 1WMM

Volume 34 • Number 1-2 • 2011 (2012) pp. 149-186 inscriptions may provide evidence of school names connected to specific areas, they do not contain the details required for a clear picture of the character of the schools cited or of their interrela­tionships.

The most detailed sources are the early Indian Buddhist texts themselves, which, although not historical documents in the con­ventional sense, nonetheless can, if carefully examined, yield im­portant historical information. Like all documents, such texts are themselves historical products whose similarities and differences reveal connections and lineages of transmission that suggest re­lationships among early Buddhist groups. These relationships are best uncovered not through a mapping of general doctrinal posi­tions but through the careful analysis of specific text passages and subtle terminological and syntactic comparisons. In addition to the early Indian Buddhist textual materials preserved in Pali and in Chinese translation, we now have Gandhari manuscripts that pro­vide invaluable textual evidence by being sources that were not redacted after the lst-2nd centuries.

The present article is offered as a case study that will attempt to find clues about text-historical relationships among Buddhist texts and Buddhist schools by examining one doctrinal point of dispute recorded in sutra and Abhidharma texts from different schools: specifically, the dispute concerning an intermediate state (cmtardbhava) between death and rebirth and the related list of five types of non-returners (andgdmiri).

To explain how a sentient being’s karma is carried from one life to the next, Buddhist theorists found it necessary to provide a detailed account of the mechanism of death and rebirth. Since the general descriptions found in the sutras were considered in­sufficiently detailed to withstand critical examination, some later Buddhists developed the notion of an intermediate state or antara­bhava between death and rebirth. They suggested that a being at death enters the antarabhava for seven days or more until attain­ing the next life.

In this way, causal continuity between the past life and the new life was preserved, and, along with it, the causal efficacy of past karma. The notion of antarabhava also explained how a being who dies in one place at one time can be reborn in another place at another time. However, the antarabhava was re­jected by several Buddhist schools. Their main argument against it was that the antarabhava was not mentioned in the sutras, and that the classification of sentient beings and their realms did not include a realm of antarabhava. For example, Pali sources reject the antarabhava theory, and instead develop a sophisticated theory of mind-process to explain the process of death and rebirth without resorting to a notion of antarabhava

In order to support their positions, both opponents and propo­nents of the doctrine of antarabhava developed different argu­ments that are, however, both traditionally based upon either scrip­ture (agama) or reasoning (yukti). In the case of the antarabhava dispute, since there is no explicit reference to it in the sutras, there is no agama argument to directly support it. But, within the list of non-returners or andgdmins, which is found within the sutras, certain Buddhists understood the first member of the list - an- tara-parinirvayin - as someone in the state of antarabhava and, as a result, used this list of andgdmins as a proper agama argument for the proponents of antarabhava. So, our investigation starts with the anagamin list in the sutras.

2. The list of anagamins in the sutras

The word antarabhava does not occur in the Pali Nikayas, nor do the corresponding Chinese translations +W or cf3?# occur in the Chinese Agamas. But its two components antard and bhava are very common words in Indic languages and are encountered very frequently. Antard can act as an adverb, a preposition, or a prefix added to other words carrying the meaning “in between” or “in the middle.” And bhava is mostly used in the sense of “form of rebirth,” “state of existence,” or “life” (PED s.v). The list of five kinds of anagamins is especially interesting regarding the dispute over antarabhava because the first member of the list is antara- parinirvayin (Pali antara-parinibbayin), which contains the pre­fix antard. An anagamin is one variety of Buddhist practitioner or noble person (arya/ariya) who will never return to this realm of sensual desires (kdma-dhdtu).w

The word antard-parinirvdyin/ antara-parinibbayin literally means “one who attains complete nirvana in between.” Like many other terms from early Buddhist literature, there is no explanation of its meaning in the sutras. Proponents of antarabhava interpret the phrase “in between” as in between the death moment in the realm of sensual desires and the rebirth moment in the realm of form (rupa-dhatu); hence, it can be considered a type of intermediate state of being or antarabhava. Therefore, they use the list of andgdmins as a scriptural justifi­cation (i.e. an argument grounded in scripture or dgama) for the notion of antarabhava. Others, who interpret antard-parinirvdyin as someone attains nirvana in the middle of his life span in the rupa-dhatu, instead of in the state of antarabhava between one lifetime and the next, would reject this list as a justification for the antarabhava.

The sutra materials we have today consist of two major collec­tions in Pali and Chinese translations, with some separate texts or fragments preserved in Sanskrit and Gandhari, and a few sutras in Tibetan translations. These various materials have been passed down by different traditions. Table 1 is a list of sutra materials that contain the anagamin list and their school affiliations as generally accepted by Buddhist scholars. It should be noted that here I use the term “affiliated with” in the sense that a text is closely connected to Buddhist groups such as the Dharmaguptaka or Sarvastivada and not to indicate that it originated within a particular sectarian context. �

Gandhari Sangitisutra (Sang-G) Dharmaguptaka
Dirghagama (DA) Taisho no. 1 AIAA® Dharmaguptaka

  • SangTtiparyayasutra (DJFMJ) Taisho no. 12 AftAAA unknown

Madhyamagama (MA) Taisho no. 26 17-: AfA Sarvastivada
Samyuktagama (SA) Taishono. 99 JtHA® Sarvastivada

Pali Nikayas Theravada

Table 1: Sutra texts containing the anagamin list and their school affiliations

Table 2 lists the occurrences of the anagamin list in the sutras. The first row indicates where the list occurs, and rows 2-6 contain the forms in which the five members are listed.
Sang-G DA (Tl,

51C13-14) DJFMJ (Tl,

231bl4-15) MA (Tl, 427al3-c24)
1 amtaraparinivai MW AA MW

2 uvahacapariniva'i MW AA MW
3 asamkharapariniva't AASAW AAA ASAW
4 sasamkharapariniva't WASAW AAA AASAW

5 ubhrasodo akinithakami ±;®A SAW

... continuation of table 2

MA (Tl,

616al5-16) SA (T2,

196cll-20, 197a21-28;


211bl7—24) SA (T2, 219cl5-23, 220a5-16) Pali Nikayas (D III 237; A I 233; A II155; A IV 13-14, 71-74, 146, 380; A V 120; S

V 69-70)

1 ASAS!? ASAW ASAW antaraparinibbayT

2 am ASAW ASAW upahaccaparinibbayT

3 assw WSAW asamkharaparinibbayT

4 AAfSAW WSAW AAJSAW sasamkharaparinibbayT

5 aaissw AAtSAW uddhamsoto akanitthagamT

Table 2: Occurrences of the anagamin list in sutra texts

In the six Chinese translations listed in table 2, some of the varia­tion merely reflects different Chinese translations for the same Indic terms. For example, -A in the DJFMJ and in other Chinese versions are obviously translations of the same term upapadya-parinirvayin. If we reconstruct the Indic language list from the Chinese texts, there are actually only three versions, which are shown in table 3.

Version I Version II Version III

Pali Nikayas, Sang-G DA, SA MA, SA, DJFMJ
1 antard-parinibbdyin antara-parinirvayin

ASAW antard-parinirvdyin


2 upahacca- parinibbayin upapadya-parinirvayin

ASAW upapadya-parinirvayin


3 asarikhara-

parinibbayin asamskara-parinirvdyin

StSAW sasamskara- parirurvaxiti kasaw

4 sasahkhdra- parinibbayin sasamskara-parinirvayin

WASAW asamskara-parinirvdyin


5 uddhamsota urdhvasrota tTut urdhvasrota Ai'A

Table 3: Variations in the list of the five anagamins in the sutrasThese three versions record two specific differences in the ana- gamin list:

(1) The wording of the second item: In the Pali and Gandhan lists, the second item is upahacca, which seems to be a gerund form from the root han - “to smite,” thus the meaning of this term could be “after hitting, damaging” or “after reaching” (CPD s.v.) In the other texts, this item is upapadya, which is from the root -J~pad “to go” with the added prefix upa- “near,” in this context most likely under­stood as “being reborn” (^), as confirmed in the Chinese transla­tions.

(2) The order of the third and the fourth items: Version II of the list has asamskara as the third and sasamskara as the fourth item; while in Version III the order of these two items is reversed.

As emphasized earlier, there is no explanation of the list in sutra materials, except for the spark simile discussed in section 4.1, which, unfortunately, still does not give any explicit clarification of these five kinds of persons. As a result, there is no internal evi­dence within the sutra texts themselves to indicate which list is the “original” or “correct” one. Confronted with this problem, ancient commentators tried their best to justify their own version of the list, as becomes clear in the interpretations presented within the commentaries and Abhidharma texts.

3. The lists of anagamins in early commentaries and Abhi­dharma texts With the term “early commentaries and Abhidharma texts,” I refer to the commentarial texts from the earliest days of Buddhism up to around the 5th or 6th century CE, when the Abhidharma systems of Buddhist schools developed to maturity both in the Northern and the Southern traditions. Table 4 lists the relevant texts and their school affiliations. It also fists their positions as supporting or re­�jeering the notion of antarabhava, which will be discussed later in detail. The last column of the table shows the matching version of the andgdmin list in the sutras from table 3.

Text School affili­ation antara­bhava ancigamin list version

Pali Commentaries and Canonical Abhidhamma Theravada No I

  • Sdriputrabhidharma (SAS)

Taisho no. 1548 Dharmagup- taka (?) Uncertain, apparently No I
Sahgitiparydyapada (SP) Taisho
no. 1536 Sarvastivada Yes III
Jhanaprasthdna (JP) Taisho no.
1544 Sarvastivada Yes III

  • Aryavasumitrasamgrhrta

(AVBS) Taisho no. 1549'
SWO Sarvastivada Yes Uncertain, apparently III
Mahdvibhdsdsdstra (MVS)
Taisho no. 1545 Sarvastivada Yes III

  • Sammatiyanikdyasdstra (SM)

Taisho no. 1649 Sammatiya Yes III

  • Samyuktdbhidharmahrdaya

(SAH) Taisho no. 1552
>L'bW Sarvastivada Yes III
Abhidharmakosa (AKBh) Sarvastivada/
Sautrantika Yes II, III

1646 fzicWgiH Darstantika (?) No II
Yogdcdrabhumi (YBh) Taisho no.
1579 JtftWO Yogacara Yes II
(AS) no. 1605 Yogacara Yes II

Table 4: Early commentary and Abhidharma texts

From tables 3 and 4 we can see obvious uniformity among the texts within each school and the anagamin list that each text con­tains. The three versions of the anagamin list roughly correspond to the Theravada (version I), the Sarvastivada (version III), and the Yogacara (version II) lineages. However, it should be noted that in the Yogdcdrabhumi list the third and the fourth items are spelled as anabhisamskara-p° and sabhisamskara-p° - with an extra prefix abhi- attached. This spelling appears also in the Abhidharmakosa.

4. Interpretations of the anagamin list

The sutra and commentarial texts contain three explanations or interpretations of the anagamin list: (1) the spark simile found in the surras; (2) the Sarvastivada interpretation found in the SangTti- parydya (SP) and strictly followed by most Sarvastivada texts; (3) the Theravada interpretation found in the Pali commentaries and Abhidhamma texts.

The spark simile occurs in the Pali Anguttara Nikaya and Chinese Madhyamagama. It is also mentioned in the later Sanskrit and Chinese Abhidharma texts. So, it is certain that this simile is shared by most of the traditions or schools and is not associated with any particular version of the anagamin list in table 2. But the Theravada interpretation of the simile is limited to the Theravada version of the list, just as the Sarvastivada interpretation is limited to the Sarvastivada version. I will discuss the three interpretations in the following three sub-sections.

4.1. The spark simile in the sutras

In the Pali Anguttara Nikaya (A) IV 70-4, the Purisagati (“Going of man”) sutta has the same content as the Chinese (“Going of good man”) siltra in the Chinese Madhyamagama (MA) (T1 427al3-c24). It discusses the seven ways in which one can enter nirvana. In other words, when a monk practices in a certain way and abandons the fetters, he may attain nirvana in seven ways, like the extinction of seven kinds of sparks: that is to say, when some­one hits a slab of hot metal with a hammer, the sparks that fly off the metal are extinguished in the following seven ways:

Spark Person
  Pali A IV 70-4 Chinese MA T1
427 al3-c24
la a bit which comes off from a hot, beaten iron slab, and then cools down antara-
parinibbayin antard-
lb a bit which comes off, flies up and then cools down antara-
parinibbayin antara-
lc a bit which comes off, flies up, and then cools down before falling on the ground antara-
parinibbayin antara-
2 a bit which cools after falling on the ground upahacca- parinibbayin upapadya- parinirvdyin
3 a bit which flies up and falls on a lit­tle fuel, igniting it, then cools down after the fuel gets used up asahkhara- parinibbayin sasaniskara- parinirvayin
4 a bit which falls on a large heap of fuel, but cools down after the fuel is used up sasahkhara- parinibbayin asamskara- parinirvayin
5 a bit which flies up and falls on a heap of fuel such that a fire spreads, but then goes out when it reaches e.g. water or rock uddhamsota urdhvasrota

Table 5: The spark simile and the anagamins

From the simile, we can see that this list of andgdmins is under­stood as an ordered sequence, and just as the first spark within the list is extinguished faster than the following ones, an anagamin mentioned first in the list attains nirvana faster than the following ones. In other words, an antard-parinirvdyin should enter nirvana faster than an upapadya-parinirvayin / upahacca-parinibbayin, etc. Or, in Buddhist terms, an anagamin at the beginning of the list is superior to the following ones because he has fewer remaining defilements.

It should be noted that the anagamin list given in the Pali text is obviously different from the one in the Chinese MA. Item 2 in the Pah list is upahacca while in the Chinese is upapadya. and the positions of item 3 and 4 are switched in the Chinese list, relative to the Pali one. Obviously, these two texts share the same tradition of the spark simile but they represent different traditions or lineages in the interpretation of the anagamin list. The Yogacarabhumi (178,10-182,5; T30 425all-b5) explicitly quotes this simile to explain three kinds of antard-parinirvayins', and it also follows the sutra in presenting the list as an ordered sequence reflecting the relative superiority of different andgdmins. The interpretation in the *Tattvasiddhi (T32 246a27-b25) fol­lows a pattern similar to the Yogacarabhumi, but, significantly, the *Tattvasiddhi uses the term antarabhava to explain the three kinds of antard-parinirvayins, which contradicts its own position, as given in chapter 25 of the text.

4.2. The Theravada interpretation

Interpretations of the anagamin list occur in several places in the Pali Abhidhamma and commentaries (ptthakatha).73 Among them, the Puggalapahhatti (Pp) is the only canonical text. In the Pp the five anagamins are given in the “Division of Human Types by One” (ekakam). The interpretation can be summarized as follows:

1. antard-parinibbdyin: One having destroyed the five fetters connec­ted to the lower realm, having been reborn in the rupa-dhatu, brings forth the noble path, abandons fetters connected to higher realms, and attains parinibbana before the middle of the life span.

2. upahacca-parinibbdyin: One having destroyed the five fetters con­nected to the lower realm, having been reborn in the rupa-dhatu, brings forth the noble path, abandons fetters connected to higher realms, and attains parinibbana after the middle of the life span.

3. asahkhdra-parinibbayin: One having destroyed the five fetters con­nected to the lower realm, having been reborn in the rilpa-dhdtu, brings forth the noble path without sahkhara, abandons fetters con­nected to higher realms, and attains parinibbana.

4. sasahkhdra-parinibbdyin: One having destroyed the five fetters connected to the lower realm, having been reborn in the rupa- dhatu, brings forth the noble path with sahkhara, abandons fetters connected to higher realms, and attains parinibbana.

5. uddhamsota: One having destroyed the five fetters connected to the lower realm, having been reborn in the lower realm of the rupa- dhcitu is then reborn in higher and higher realms until akanittha -the top level of the rupa-dhatu, in which he abandons fetters con­nected to higher realms, and attains parinibbana. It should be noted that the Pp explanation of items 3 (asahkhdra- parinibbayin) and 4 (sasahkhdra-parinibbdyin) does not gloss the words asankhdra and sasahkhdra but merely repeats these terms in the instrumental case to characterize these two kinds of anagamins:

3. asahkhdra-parinibbayin: asankharena ariyamaggam sahjane- ti...

4. sasahkhdra-parinibbdyin: sasankharena ariyamaggam sahja- neti...This does not help to clarify the word sankhara, which is a diffi­cult term with multiple meanings. However, in the commentaries, Puggalapahhatti-atthakatha (Pp-a) Ahguttaranikaya-atthakatha {=ManorathapuranT, A-a), and Samyuttanikaya-atthakatha (=Sa- ratthappakasini, S-a), the word sankhara is glossed with payoga (Skt. prayoga), which gives items 3 and 4 the meanings “without exertion” and “with exertion.” Here we can see a gap in interpreta­tion between the canonical Pp and the atthakathav. the Pp inher­ited the ambiguity from the sutta and continued to use the same word sankhara, while the atthakathas by using the word payoga avoided the ambiguity, but introduced an additional problem in the interpretation of the anagamin list. This issue will be discussed in detail in sections 5.3 and 5.4.

The *Sariputrabhidharmasastra (SAS, Taisho no. 1548 Htb®aiffiT28, 587bl2-588al5) interprets the list in a way some­what similar to the Puggala-pahhatti especially concerning item 1, the antara-parinivayin. In the SAS also, antara-parinivayin is understood to refer to someone reborn in the rupa-dhdtu as a deva, who, in the middle of his lifespan, destroys [[[defilement]]? or life?] dharma and attains antara-parinirvana

+, Obviously, this is contrary to the interpretation of antarahhava as an intermediate state. Also item 2 *upahacca-parinirvayin is translated as (attains parinirvana quickly) and explained as referring to someone reborn as a deva who has a shorter lifespan but more pleasure, who leaves quickly to attain parinirvana (^#^^4712^^ _h,

. The *sasamskara-parinirvayin is one who is born as a deva, attains *anantaryamarga, and then attains parinirvana (j^f W) Here, as in the Pp, the SAS retains the term asamskara from the sutra in its explanation. But for *sasamskdra-parinirvayin, it speaks of someone reborn as a deva, who attains *dnantaryamarga with toil (Ott), and then attains parinirvana Ji,

Here, the SAS is closer to the interpretation of the Pali atthakathas. The in­terpretation of the last item *urdhvasrota is virtually the same as the Pp. Thus, the *Sdriputrdbhidharmasdstra is very close to the Theravada interpretation with some differences in minor details. However, it is significantly different from the Sarvastivada inter­pretation, as demonstrated in the next subsection.

4.3. The SahgTtiparydya interpretation

The SahgTtiparydya (SP, Taisho no. 1536 is one of the seven canonical Abhidharma treatises of the Sarvastivada school. It is a commentary on the SahgTtisutra/SahgTtisutta, which appears in the Chinese DTrghagama and the Pali Dighanikaya. The main part of the Sahgitisutra is a list of Buddhist teachings ordered numerically from one to ten. In its chapter on categories of dhar­mas containing five members, it gives the list of five anagamins. The SP explains the list as follows (T26 425c28-426c21):

1. Antara-parinirvdyin-. One has abandoned the five fetters connected with the lower realm (panca dvarabhdgTya samyojananiTL]\\3C3 33 but who still has the five fetters connected with higher realms (pa Pica urdhvabhagTya samyojandni ), has made the re­ sultant karma of coming forth (*utpdda-vipdkakarma but without the resultant karma of rebirth (*upapatti-vipakakarma After the present body perishes and the antarabhava rises, before he is reborn in the rupa-dhatu, he enters parinirvana in the state of antarabhava.

2. Upapadya-parinirvayiir. One has both the resultant karma of com­ing forth and the resultant karma of rebirth. The present body per­ishes, and the antarabhava rises, and then he is reborn in the rupa- dhatu. Shortly after being bom in the rupadhatu, he either (1) enters parinirvana-, or (2) attains andsrava-marga, but lives longer until his life span is finished.

3. Sasamskdra-parinirvayin-. One who is reborn in the rupa-dhatu (1) practices with exertion, further eliminates remaining fetters, and enters parinirvancr, or (2) practices with conditioned (sainskrfa) ob­jects.

4. Asamskara-parinirvayin'. One who is reborn in the rupa-dhatu, (1) practices without exertion, further eliminates fetters, and enters pa- rinirvancr, or (2) practices with unconditioned (asamskrta) objects.

5. Urdhvasrotcr. One is reborn in the rilpa-dhatn among the Subha- krtsna devas (i.e. the upper level of the third dhycina). Then after the life span in this level is finished, he is reborn in the lower level of the fourth dhyana. In this way, in each life, he is reborn in a higher level, until he reaches the top level of the fourth dhyana (i.e., the upper limit of the rupa-dhatu), and attains nirvana there.

The Vibhdsd (T27, 874b21-) and the *Samyuktdbhidharmahrdaya (T28, 912bl6-) follow this interpretation exactly. The Abhidhar­makosa (VI.37) agrees with most of this explanation, but rejects the alternatives concerning conditioned (samskrta) or uncondi­tioned (asamskrta) objects given as 3(2) and 4(2) in the SP.

5. Some observations

These differing versions of the list of andgdmins provide a rich source for the investigation of text-historical relationships among Buddhist texts and Buddhist schools. To this end, in the following section, I will analyze the textual data presented above from dif­ferent perspectives. First, I will try to group the texts according to the versions of the list they contain, with attention to differences among the texts that might point to possible lineage connections. Next, I will analyze the differences in the wording (tipahacca vs. upapadya, samskara vs. abhisamskdra) and ordering (item 3 and 4) of the items, to find what historical information these lists may reveal to us.

5.1. The lineages of texts

If we understand a lineage of Buddhist teachings as a succession of teachers and pupils that pass down texts and doctrines, and if we presume, in addition, that members within a given lineage attempt to maintain the stability of their sacred texts, then the differences in the texts we have may indicate different lineages. As shown in table 3 above, there are three versions of the anagamin list in the sutras. And table 4 shows that each of the later commentarial and Abhidharma texts follows one of these three versions. Based on these three versions of the list, we can divide the texts into three broad groups:

Group 1 (VersionI): Pali Nikayas, Pali commentaries and Abhidhamma, Gandhari Sangitisutra and commentary, *Sdriputrabhidharma Group 2 (Version II): Dtrghagama, Samyuktagama(l), *Tattvasiddhi, Yogacarabhumi, Mahayanabhidharmasamuccaya Group 3 (Version III): Madhyamagama, Samyuktagama(2), SarigTti- paryayapada, Jhanaprasthana, *Aryavasumitrabodhisattvasamgrhi- ta, Mahavibhasa, *Sammatiyanikayasastra, *Samyuktabhidharma- hrdaya

Group 1 is differentiated from group 2 by the wording of item 2 in the list: group 1 (version I) has upahacca, but group 2 (version II) has upapadya. Group 2 is differentiated from group 3 by the order­ing of items 3 and 4 in the list: group 2 has asamskdra preceding sasamskdra, while group 3 (version III) has the opposite order with sasamskdra preceding asamskdra. And finally, the list of group 3 is different from that of group 1 in both the wording of item 2 and the ordering of items 3 and 4. The texts clustered in group 3 conform to certain lineage con­nections given within the historical accounts in ancient Buddhist chronicles. Hence, this grouping of texts is the

most expected among the three. The Sangitiparyayapada, Jhanaprasthana, *Aryavasu- mitrahodhisattvasamgrhita, Mahavibhasa, and the *Samyuktdbhi- dharmahrdaya are all believed to be Sarvastivada texts, so it is no surprise to see they all have the same version of the list. The only exception in this group is the *SammatTyanikdyasdstra, which is a text of the SammatTya school - the major non-Mahayana rival of the Sarvastivada in Northern India. The appearance of the same list of anagamins in their texts may suggest that the

Sarvastivada and the SammatTya shared some common textual lineage. According to the historical accounts of Buddhist schools available to us from a number of sources, it is generally accepted that the SammatTya school descended from the VatsTputriya (Lamotte 1988: 529ff), but how the VatsTputriya is related to the Sarvastivada is unclear. Some contend that they are separate groups that emerged from the origi­nal Sthavira branch, while others postulate that the VatsTputriya descends from the Sarvastivada. But regardless of what might have been the case, it is generally accepted that the VatsTputriya and its descendent, the SammatTya, have many doctrinal common­alities with the Sarvastivada. And regarding the antarabhava is­sue, both the Sarvastivada and the SammatTya accept the interme­diate state as antarabhava’, therefore, it is no surprise to see that they share the same version of the anagamin list.

The texts in group 2 are somewhat mixed in terms of their pos­sible school affiliations. The Chinese DA is probably related to the Dharmaguptaka, the *Tattvasiddhi is likely an eclectic work draw­ing from many sources, but more closely related to the Darstan- tika-Sautrantika, and the Yogdcdrabhumi and the Abhidhar- masamuccaya are Yogacara texts. It is widely accepted that the Yogacara has a close relationship with the group named Sautran- tika, which is believed by some to be a later development from the Darstantikas within the Sarvastivada tradition. This list of anagamins may be seen as one piece of evidence that confirms the possible connection between the Darstantika and Yogacara.

The DA is an outlier in this group. Even though it is believed to be a Dharmaguptaka text, item 2 in its list is upapadya and not upahacca as found in the Sang-G, which is also associated with the Dharmaguptakas, but has the version I of the anagamin list. Since the DA was translated into Chinese rather late (413CE) in comparison to the date of the Sang-G (1 c. CE), it is possible that inter-textual influences resulting from interactions among the vari­ous traditions took place in the three hundred intervening years. Could the early date of the Sang-G and its difference from the DA indicate that its list containing upahacca is earlier than the one with upapadyal Unfortunately, these textual

passages alone do not permit a definitive conclusion. Although the use of these different terms may be significant with the DA or with the Dharmaguptaka group in general, there is no way to determine the relative dating of the two lists for texts not affiliated with the Dharmaguptakas. In other words, it is possibile that both upahacca and upapadya ap­peared at an early time and were adopted by different groups. At a later time, especially as the Sarvastivada gained dominance in northwest India, the texts of other schools, like Dharmaguptakas, may have been influenced, consciously or unconsciously, by the Sarvastivada texts and their use of the upapadya list. However, one must always take into account the possibility that texts have been altered or “normalized” by the authors and editors of the later tra­dition without sectarian or doctrinal motivation or explicit recourse to any other justification; for example, changes in texts can result from incomprehension introduced by scribal errors, temporal gaps, etc., which the later tradition then attempts to rationalize.

However, could the fact that the DA and the Yogacara texts have the same anagamin list indicate a close connection between them? It is possible, but again very difficult to prove. Traditional accounts about the Dharmaguptaka, Mahlsasaka, Vibhajyavadin, and Sarvastivada groups are chaotic and do not clarify the relation­ships among them. But I would suggest that the Dharmaguptakasconnection with the Yogacara is likely to be a remote one. It may, however, become a little clearer after we examine the texts of group 1, all of which contain version I of the anagamin list.

Group 1 suggests interesting historical connections among the various textual lineages. The Sang-G, the SAS, and Pali texts share a list that is significantly different from that of both the Sarvastivada and the Yogacara, which may indicate that the texts in group 1 have some special connections with each other. It is believed by some scholars that the SAS we have today may be closely related to the Dharmaguptakas, and the similarity of its anagamin list with that of the Sang-G favors this hypothesis. Furthermore, the two texts connected to the Dharmaguptakas - the SAS and the Sang-G - share the same list with the Pali texts, which may indicate that there is a relatively closer relationship be­tween the texts of the Dharmaguptakas and those of the Southern Theravada. The Southern Theravadins describe themselves as Vibhajyavadins that emerged from the early Sthavira branch. Some scholars think that they are possibly the descendent of a Vibhajyavada branch named Tamrasatlya. Interestingly, in the Mahasanghika account of Buddhist schools given in Taranatha’s work, both the Dharmaguptakas and the Tamrasatiyas are classi­fied as Vibhajyavadins. Moreover, this Mahasanghika account lists the Vibhajyavadins as a branch separate from the Sthavira, which has the Sarvastivadins and the VatsTputriyas as its descendents. The Mahasanghika account about these non-Mahasanghika Buddhist groups is shown in the chart in Figure l.

Our textual data regarding the anagamin list support this ac­count of the possible filiation of relevant Buddhist schools: the SammatTyas, as descendents of the VatsTputriyas, have a closer relation to the Sarvastivadins than the Vibhajyavadins; the Dhar­maguptakas are closely related to the Southern Pali tradition (i.e. the Theravada) as they are both Vibhajyavadins. In other words, the Mahasanghika account makes good sense of the textual data we have; it can explain why the *SammatTyanikayasdstra has ex­actly the same anagamin list as the Sarvastivada texts, and why the Sang-G and the SAS have the same list as the Pali Theravada texts.Mahisasaka

Sthavira -x
Mui sarvastivada
Sautranti ka

Sannagari ka

Figure 1

The Chinese Buddhist scholar Yinshun, after examining all the ac­counts of the chronicles of the Buddhist schools, suggests that the Mahasanghika account might be the most reliable among them, because as a group outside the Sthaviras, the Mahasamghikas can observe and record the development and division of the Sthavira branch of Buddhism in a more detached, unbiased, and therefore relatively, more objective manner (Yinshun 1988a: 144).

Further, concerning the SAS, some scholars like Yinshun pro­pose that it represents a proto-Abhidharma stage that was shared as a foundation for the development of the Abhidharma texts of differ­ent Buddhist groups (1981: 65flf). This view is supported by the ap­pearance of markedly similar Abhidharma lists of topics (matrkd/ matika) that are found within the SAS, the Dharmaskanda, and the Pali Vibhahga. However its explicit Vibhajyavada position re­garding antarabhava and the anagamin list suggests that the SAS available to us now is no longer the proto-Abhidharma text that was shared by many Buddhist groups but rather is a more developed Abhidharma text that contains sectarian contents associated with the Vibhajyavadins.

It is widely accepted that the Yogacara is in some way connect­ed with the Buddhist group referred to as Sautrantika, which some believe developed from the Darstantikas within the Sarvastivada school. The relationships among the Sarvastivadins, Darstantikas, Sautrantikas, and Yogacarins, are still unclear. However, if we look again at the texts grouped according to the anagamin lists, group 2, which represents version II of the list, suggests that the Yogacara texts and the *Tattvasiddhi, which contains positions similar to those of the Darstantikas, may belong to a different textual lineage than the Sarvastivada. Version II of the anagamin list has the same ordering of items 3 and 4 as the Vibhajyavada texts, and, as shown in the Mahdvibhdsd, the authoritative work of the Sarvastivadins, the Vibhajyavadins are opponents of the orthodox Sarvastivadins. The anagamin lists may then be one piece of evidence demonstrat­ing that the Yogacarins and possibly the Darstantikas may not have been “pureSarvastivadins with regard to the texts that they used. They may have absorbed some texts as well as doctrines from op­ponents of the Sarvastivadins such as the Vibhajyavadins.

However, the Chinese SA makes the situation even more com­plicated. Versions II and III of the anagamin list are both contained in the SA. As mentioned previously, the SA we have today was very likely used by both the Sarvastivadins and the Yogacarins. If this was the case, the early Sarvastivadins were obviously selective in using their texts; in the case of at least the anagamin list, they sim­ply ignored variations included in the SA. But commentators like Vasubandhu do point out the differences in the list, and he chooses a non-Sarvastivada one as correct precisely because he thinks it is more reasonable. From this we can see that a “lineage” of texts does not indeed constitute a homogeneous thread, but is rather like a growing web influenced by a variety of factors that can only be partially perceived in the current form of the preserved texts.

Here I must emphasize once more that my analysis thus far has been based on data retrieved from ancient texts with the presump­tion that these texts were affiliated with certain Buddhist “tradi­tions” or “schools.” But the way in which the texts were actually associated with historical Buddhist groups is far from certain. The anagamin list has shown complex relations among Buddhist texts, and I would suggest that we should understand Buddhist groups in a similar way. Terms like Buddhisttradition” and “school” should not be understood as representing discrete and static histori­cal entities. Perhaps it would be better for us to take the conceptstradition,” “lineage,” and “school” as convenient umbrella terms representing loose groupings of texts, people, etc., simply for the purpose of discussion, while at the same time conceding that they do not represent the strict, clear-cut categories that ancient sectar­ian historians would have us accept.

5.2. Upahacca vs. upapadya

In all of the Pali texts, the second item of the anagamin list is given as “upahacca,” while in the few Sanskrit extant texts it is upa­padya. Almost all Chinese texts, with the sole exception of the SAS, translate it as ’’which can be understood as the nounbirth” or the verb “to be born.” The SAS translates it as “£8,” which can be understood as an adjective or adverb that means “quick” or “fast.” As it is impossible to interpret upapadya as “quick,” the more likely equivalent of in the SAS is upahacca. Upahacca is a gerund form from the root -Than (to smite), literally meaning “having hit.” As mentioned in section 4.2, the Pali texts interpret the upahacca-parinibbdyin as someone reborn in the rupa-dhdtu, who attains parinirvana after the middle of his life-span but before that life-span ends. In other words, the upahacca-parinibbdyin has a shortened life. In this sense, it seems very likely that the original Indic term in the SAS for the Chinese ® (“quick”) was upahacca. Also this sense of upahacca might be the reason, as Peter Masefield observes, why some modern scholars understand upahacca as “reducing” or “cutting short” (PED s.v.).

The question then is which term - upahacca or upapadya - is the original one? Why and how did the different terms come to be used?

All of the relevant texts are surprisingly silent on the differ­ent wording of upahacca vs. upapadya with the exception of the Kathavatthu-atthakatha (Kv-a). When commenting on the point of controversy concerning whether one can attain arhatship at the very moment of rebirth, the Kv-a states that those of the Northern coun­try (Uttardpathakd) changed upahacca-parinibbdyin to upapajja- parinibbdyin^ in order to support their position that one can attain arhatship at the moment of rebirth. The northern texts indeed con­firm that the position of both the Sarvastivadins and Sammatlyas conforms to that described in the Kv-a, but there is no evidence to prove the contention that the term was intentionally changed to support their position.

Nonetheless, Kv-a’s explanation suggests that the term used is significant in the interpretation of the list. If it is upapajja-pari- nibbayin (Sanskrit upapadya-parinirvayin), the only possible in­terpretation is “one who attains parinirvana after being reborn.” Furthermore, as the simile of the spark shows, the anagamin list is an ordered sequence. Thus, if the item means “one who attains parinirvana after being reborn,” then the prior item should refer to someone who attains nirvana before being reborn, hence justifying the existence of antarabhava. But if it is upahacca- parinihhayin (no corresponding Sanskrit word is attested in this context), it means literally “one attains parinibbana after hitting,” which is rather ambiguous. It may have a similar sense as the upapajja/upapadya, in which “hitting” could be “hitting the new life,” and hence, “being reborn.” However, the Pali commentators want to avoid the possible connection of this term to rebirth and antarabhava. The Pp (p.T7) explains upahacca-parinibbdyin as one “having passed the mid-point of life-span, having gone toward death (upahacca kdlakiriyani), he brings forth the noble path.” It seems that the Pp is trying to explain upahacca as “approaching death.” Obviously this stretches the meaning of this word, and commentators after the Pp appear not to be very comfortable with this explanation. In the atthakathds (except for the Pp-a, of course,) this interpretation, “approaching death,” is never mentioned.

It is important to notice that Pali texts, when describing the sparks “not falling on the ground” and “falling on the ground” within the spark simile, use the phrases “anupahacca talam” and “upahacca talam” And the A-a (IV 39) glosses the upahacca with upahanitva, the “normal” BHS gerund form for upa-d~han, which can easily be understood as meaning “having hit.” But interest­ingly, the A-a is silent on the term upahacca within the compound upahacca-parinibbayin. Masefield suggests that the reading of the name of this anagamin might be “infected” by the upahacca talam (“having hit the ground”) in the spark simile, and the original name of the anagamin probably should be upapadya-parinirvayin (Pali upapajja-parinibbdyin).^ And in the few Sanskrit sources we have, “not falling on the ground” is expressed with “prthivyam apcitita” the past participle of Spat “to fall,” instead of the gerund from upa- -fhanr7 Therefore, in contrast to Masefield’s suggestion, it is also possible that the name of the anagamin infected the wording of the simile in the Pali texts. In other words, if the original name of the anagamin were upahacca-parinibbayin, an editor might un­derstand it as “having hit [[[rebirth]]],” and then use the same word to describe the spark in the simile “having hit the ground,” i.e., the gerund from upa-fhan. Thus, given the ambiguity of the meaning of upahacca, I think there is no hard evidence to judge which word is the original one.

5.3. The spelling of items 3 and 4 with and without the prefix abhi-

In the Sanskrit Yogacarabhumi, the third and fourth items in the anagamin list are spelled anabhisamskdra-parinirvdyin and sabhi- samskdra-parinirvdyin. The spellings with the prefix abhi- also appear in the Abhidharmakosabhasya, Table 6 shows the transla­tions of items 3 and 4 of the list in the Chinese translations:�

Texts Item 3 Item 4

DTrghagama (DA) Taisho no. 1 HHA®

  • SahgTtiparyayasutra (DJFMJ) Taisho no. 12 AW® AAA AAA

Madhyamagama (MA) Taisho no. 26 ' •'HA® AASA1®
Samyuktagama (SA) Taisho no. 99 JtHA®
Samyuktagama (SA) Taisho no. 99 JtHA® AASA1®
Sariputrabhidharmasastra (SAS) Taisho no. 1548 WPO »A AASA1 »A

  • Aryavasumitrabodhisattvasamgrhita (AVB S) Taisho no. 1549 iliSF/Tftgiw AASA1®

Mahavibhasa (MVS) Taisho no. 1545 HBbStBI
no. 1547 *Vibhasa ASAW AASA1®
no. 1546 *Abhidharmavibhasa AASA1®

  • SammatTyanikayasastra (SM) Taisho no. 1649 fiAW AAAW
  • Samyuktdbhidharmahrdaya (SAH) Taisho no.

1552®Hg>Q^ ASAW AASA1®
Abhidharmakosa (AK)
Paramartha’s translation Taisho no. 1559 AAM AAM
Abhidharmakosa (AK)
Xuanzang’s translation Taisho no. 1558 S AASA1®

Yogacarabhumi (YBh) Taisho no. 1579

Mahayanabhidharmasamuccaya (AS) no. 1605 AASA1®

Table 6: Chinese translations of items 3 and 4 of the andgdmin listAmong these texts, we are certain that the AKBh and the YBh use the terms with the prefix abhi-, but we see no difference in the translations of these two terms. The fact that no distinction appears to have been made in the Chinese translations might suggest that the Chinese translators did not think the prefix mattered in this context. In that case, why does this prefix appear in the YBh and the AKBh, but not in the Pali texts and the Sang-G? The Sahgitiparyaya may shed some light on this problem. As shown in section 4.3, in addition to the standard “with and without exertioninterpretation for the terms sasamskara and asamskara, the SP proposes an alternative explanation, stating that they may be understood as someone taking samskrta and asamskrta factors as objects of practice.49 50 Here the SP obviously glosses samskdra with samskrta etymologically, which is legitimate since both of them are derived from the root -/kr with the prefix sam-. This also sug­gests that the spelling in the SP is without the prefix ahhi-. The MVS and the SAH also contain this spelling without the prefix abhi-, since they accept the same samskrta interpretation. But the addition of the prefix abhi- would make it impossible to gloss abhi- samskdra with samskrta. Accordingly, the spelling with abhi- in the Ybh and the AKBh appears to be related to a position different from the samskrta interpretation. The AKBh explicitly states that the samskrta interpretation is wrong, but does not give any reason why.

The two interpretations of the term samskdra in the SP show that the early commentators were not certain about the exact mean­ing of this term. I did not find any direct quotation of this list from sutra materials in either the YBh or the AKBh. By using the term abhisamskdra with the prefix abhi-, the YBh and the AKBh restrict the interpretation to “exertion” and rule out the interpretation as different objects of practice based on the term “samskrta” It is then likely that their anti-scms^rta-interpretation position influences their adoption of the term abhisamskdra with the abhi- prefix, which would then unambiguously convey the sense of “exertion.” Firstly, the extant early texts in Indic languages such as Gandhari and Pali all contain the version of the list without “abhi-” The Sanskrit YBh and AKBh are from the 5th century CE or even later, so there is a

greater possibility that the term samskdra alone represents an earli­er version of the list than that containing abhisamskdra. Secondly, as I have shown previously, the interpretations of the SP, itself an early Abhidharma text, are based on the ambiguity within the term samskdra-, therefore we can be certain that the SP list has samskdra instead of abhisamskdra. Finally, the YBh and the AKBh, which employ the spelling abhisamskdra in contrast to the earlier texts, never state that the spelling of the term as used in the earlier texts is a problem, but only challenge the SP’s doctrinal position regarding the samskrta/asamskrta interpretation. In other words, for them, the form of the word in the list as samskdra is not important as a terminological issue but only because of the doctrinal confusion to which it might lead, confusion that could be precluded by “clarify­ing” its sense with the term abhisamskdra. The differences in the terms used in these texts reveals the history of the understanding of the term samskdra in the context of the anagamin list: earlier commentators were uncertain how to understand this term from its several possible meanings, but later commentators (like authors of YBh and AKBh) determined that “exertion” is the correct meaning and used the form abhisamskdra to clarify their position. Or else, perhaps, they simply wanted to reject the specifically Sarvastivada- informed interpretation in terms of samskrta/asamskrta.

Regardless of which interpretation the commentators chose - either “with/without exertion” or “with conditioned/unconditioned object” - there is still significant uncertainty remaining in the un­derstanding of the list of the five anagamins. In the spark simile, it is obvious that the criterion used to classify the five kinds of sparks is the time a spark takes to be extinguished. So it is reasonable for us to expect this criterion also be applied to the anagamins. And indeed, the first two anagamins - antara-parinirvayin and upapadya/upahacca-parinirvayin — are differentiated in both the Theravada and the Sarvastivada interpretations by the time they take to reach parinirvdna, regardless of whether it occurs in the state of antarabhava or already reborn in the rupa-dhatu. But with asamskara- and sasamskara-parinirvayin, both the Theravada and the Sarvastivada fail to apply the time criterion. Instead, they are differentiated according to whether they require effort, or whether the objects of their meditation are conditioned or unconditioned. Moreover, the differing order of items 3 and 4 across various texts further reveals the uncertainty in interpretation. We will return to this issue after examining further the order of items 3 and 4.

5.4. The order of asamskara- and sasamskara-parinirvayin

The texts related to the Sarvastivada school (MA, SA, SP, JP, AVBS, MVS, SAH, AKBh) as well as the *SammatTyanikayasastra (SM) list the third and the fourth items of the list in a peculiar order: samskara-parinirvayin precedes asamskara-parinirvayin. The texts affiliated with other schools like the Dharmaguptaka and Theravada reverse the order of these two items. Interestingly, the SA contains both versions of the list. This difference in the order of items 3 and 4 suggest two lineages in the interpretation of the list. (To my knowledge, this pair of anagamins appears only in the con­text of the list of five anagamins, so we have no separate case for comparison.) Almost all Sarvastivada texts share one version, with the exception of the SA, also believed to be a Sarvastivada text, which contains both versions of the list. This suggests that the SA preserves some content that predates the Sarvastivada school and would appear to confirm that there is an earlier textual heritage shared by the Sarvastivada as well as other schools.

The Yogdcdrabhumi adds further evidence to support the ex­istence of an earlier, common textual lineage. The YBh contains the non-Sarvastivada version of the list, which lists asamskdra be­fore sasamskara. However, as the Chinese scholar Lii Cheng has convincingly shown, the VastusamgrahanT in the YBh is organized according to the matrka of the SA. The similarities between the SA and the YBh strongly suggest that the Yogacara and the Sarvastivada may have a common lineage with the SA that differs from that of the Theravada and the Dharmaguptaka.

Vasubandhu comments explicitly on the different orders of asamskdra and sasamskara. In his AKBh, he usually follows the interpretation of the anagamin list based mainly on the MVS, or the Sarvastivada tradition, and explains the items in the list following the order in the MVS. However, after explaining items 3 and 4 ac­cording to the Sarvastivada order, he notes that asamskara appears first in the sutra, and that is the more reasonable order because the asamskara-parinirvayin needs less effort to reach nirvana. No other text before the AKBh mentions this ordering issue. It is especially interesting that even the MVS, which is famous for its characteristic way of exhaustively listing positions from different Buddhist groups, does not mention this issue.

These textual data are not sufficient to indicate the exact history of these texts, but it is likely that the proto-SA, which was shared by different traditions, already had both versions of the list. Later, as the Sarvastivada and Yogacara traditions separated, they treated the list in different ways; the Sarvastivada adopts one version and Yogacara, the other. Vasubandhu notices the difference and choos­es to stand with the Yogacara.

The ordering issue with respect to items 3 and 4 in the anagamin list is inevitably related to the issue discussed previously concern­ing the wording of the terms with or without the prefix abhi, and together illustrate the efforts of ancient Buddhist commentators to formulate an acceptable interpretation of the terms asamskara- and samskara-parinirvayin in the list. We can imagine that in order to make sense of the anagamin list and the spark simile, they engaged in various commentarial and doctrinal maneuvers, and perhaps even altered the text. Moreover, as I have mentioned previously, the inconsistency in the application of time as the criterion for classi­fication in the explanation of items 3 and 4 further reveals the gap between the sutra texts and these later commentaries.

However, Masefield suggests another possible interpretation for items 3 and 4 in the anagamin list. He observes that grass, branches, sticks, and other kindling as fuel for burning are consist­ently used in the sutras as a simile for the skandhas (Pali khandha). Therefore, Masefield suggests that the asamskdra-parinirvayin can be understood as the kind of anagamin who is reborn in the rupa- dhatu and gets involved in new skandhas, but to a minimal extent; the skandhas burn out quickly like a small amount of fuel. And the sasamskdra-parinirvdyin is the kind of anagamin who gets in­volved with more skandhas, which take a longer time to burn out. I agree with Masefield that this interpretation seems more reason­able and more consistent because it uses the time that is required before nirvana occurs as the main criterion that distinguishes the five a nd gamins. It also conforms better with the spark simile in the sutra texts than the explanations presented in either the Southern or Northern Abhidharma texts.

6. Conclusion: The variations of the anagamin list and the dispute over antarabhva The textual data investigated here regarding the antarabhava dis­pute and the anagamin lists in different texts cannot reveal what the Buddha’s “original” teaching was, which was the goal of an­cient commentators and historians. But they do provide very valuable historical information that helps to clarify the history of the texts, of the development of doctrines, and of Buddhist tradi­tions. As shown in previous sections, texts that support the notion of the antarabhava tend to enumerate upapadya-parinirvayin as the second item of the anagamin list, while texts that reject the antarabhava are more likely to have upahacca-parinirvayin. Also, the Yogacara texts have abhisamskdra instead of samskdra in items 3 and 4 of their lists. These examples show that doctrines did indeed influence the form that texts took. Also, by grouping the texts according to the lists they contain, I have been able to show in section 5.1 that the text groupings re­inforce the Mahasanghika chronicles’ description of relationships among Buddhist schools. This indicates that the Mahasanghika ac­count might be more credible than other accounts regarding these relationships among the Sthavira Buddhist schools.

Finally, the variation in the order of items 3 and 4 in the list, the difference in wording of samskdra vs. abhisamskdra, and their spellings in relation to the interpretations of these two items, sug­gest a gap between sutra texts and later commentarial texts, and reveal the creative struggle of later commentators when attempting to understand ambiguous passages from the sutras and bring them into conformity with their own views.

Texts and abbreviations

All Pali texts are quoted front the Chattha Satigayana CD published by the Vipassana Research Institute and proofread with the Pali Text Society editions. Abbreviations of Pali text names follow the Critical Pali Dictionary.


CPD Critical Pali Dictionary
PED Pali Text Society Pali-English Dictionary
AKBh Abhidharmakosa(bhasya)
Abhidharmakosabhasyam of Vasubandhu ed. P. Pradhan, Patna: K. P. Jayaswal Research Institute 1975
Xuanzang’s translation: Taisho no. 1558
Paramartha’s translation: Taisho no. 1559
AS Mahayanabhidharmasamuccaya Taisho no. 1605
•fe- iA.
AVBS *AryavasumitrabodhisattvasamgrhTta Taisho no. 1549
SAS Dlrghagama Taisho no. 1

Mahavibhasasastra Taisho no. 1545
Samyuktagama Taisho no. 99 *Samyuktabhidharmahrdaya Taisho no. 1552 ® Gandhari Sahgltisutra *Sariputrabhidharmasastra Taisho no. 1548
YBh *SammatTyanikayasastra Taisho no. 1649 HKeEhEw
Sahgitiparyayapada Taisho no. 1536

Yogacarabhumi Taisho no. 1579 WWffiifEaw; Sanskrit Sravaka- bhumi ed. by Karunesha Shukla, P. Jayaswal Research Institute, Patna 1973.

Aung, Shwe Hsan and Caroline A. F. Rhys Davids (1969). Points of contro­versy, or, Subjects of discourse. London, Pub. for the Pali text society by Luzac & Co.
Bareau, Andre (1955). Les sectes bouddhiques du petit vehicule. Saigon, Ecole francaise d’Extreme-Orient.
Bodhi, Bhikkhu (2000). The Connected Discourses of the Buddha. Oxford, PTS.
Cai Yaoming 8)10 (1998). 'WE (HWE) = Pan
ding a han jing bu pai gui shu qian she de nan ti." Fa Guang (111) 24.
Chung, Jin-il (2008). A survey of the Sanskrit fragments corresponding to the Chinese Samyuktagama = KEJtFjE!—ja Za a han
jing xiang dang fan wen duan pian yi lan. Tokyo, Sankibo Busshorin.
Cox, Collett (1998). Kasmira Vaibhasika Orthodoxy. In Sarvastivada Buddhist scholasticism. C. Willemen, B. Dessein and C. Cox. Leiden; New York, Brill: 138-254.
Cuevas, Bryan Jare (1996). “Predecessors and Prototypes: Towards A Con­ceptual History of the Buddhist Antarabhava.” Numen 43, 263-302.
(2003). The hidden history of The Tibetan book of the dead. New York: Oxford University Press.
Dhammajoti (2007). Abhidharma doctrines and controversy on perception. Hong Kong: Centre of Buddhist Studies, The University of Hong Kong.
Enomoto, Fumio (2000). ‘Mulasarvastivadin’ and ‘Sarvastivadin.’ In Vividha- ratnakaranddaka'. Festgabe fiir Adelheid Mette. C. Chojnacki, J.-U. Hartmann, V. M. Tschannerl. Swisttal-Odendorf: Indica et Tibetica Verlag, 239-250.
Frauwallner, Erich (1956). The earliest Vinaya and the beginnings of Buddhist literature. Roma: Is. M. E. O.
(1995). Studies in Abhidharma literature and the origins of Buddhist philosophical systems. Albany , N.Y, State University of New York.Gethin, Rupert (1994). “Bhavaiiga and rebirth according to the Abhidhamma.” The Buddhist Forum III.
(1998). The Foundations of Buddhism. Oxford, New York: Oxford University Press.
Glass, Andrew and Mark Allon (2007). Four Gandhari Samyuktdgama sutras: Senior Kharosthifragment 5. Seattle: University of Washington Press.
Harvey, Peter (1995). The Selfless Mind. Surrey: Curzon Press.
Hiraoka, Satoshi (2000). “The Sectarian Affiliation of Two Chinese Sam- yuktagamas." Journal of Indian and Buddhist Studies 49(1), 500-506.
Katsura, Shoryu (1974). A Study ofHarivarman's Tattvasiddhi. University of Toronto. PhD.
Kritzer, Robert (1997). “Antarabhava in the Vibhasa.” Maranatha'. Bulletin of the Christian Culture Research Institute fl— |'7l flJflfl:\'flfl'-fl 0 fl 5, 91(11)—69(33).
(1998a). "Semen, Blood, and the Intermediate Existence." Indogaku Bukkydgaku kenkyu 46/2, 1031-1025.
(2000a). “The Four Ways of Entering the Womb (garbhavakranti)." Bukkyb bunka 10, 1-41.
(2000b). “Rupa and the Antarabhava.” Journal of Indian Philosophy 28, 235-272.
(2005). Vasubandliu and the Yogacarabhumi: Yogacara elements in the Abhidharmakosabhasya. Tokyo: International Institute for Buddhist Studies of the International College for Postgraduate Buddhist Studies.
La Vallee Poussin, Louis de (1988-1990). Abhidharmakosabhasyam (English translation by Leo, M. Pruden). Berkeley: Asian Humanities Press.
Lamotte, Etienne (1987). Karmasiddhiprakarana: the treatise on action by Vasubandhu (English translation by Leo M. Pruden). Berkeley: Asian Humanities Press.
(1988). History of Indian Buddhism', from the origins to the saka era (English translation by Sara Webb-Boin). Louvain-la-Neuve: Universite catholique de Louvain, Institut orientaliste.
Langer, Rita (2007). Buddhist rituals of death and rebirth', contemporary Sri Lankan practice and its origins. Abingdon, New York: Routledge.
Masefield, Peter (1987). Divine revelation in Pali Buddhism. Colombo: Sri Lanka Institute of Traditional Studies.
Salomon, Richard (1999). Ancient Buddhist Scrolls from Gandhara'. the British Library Kharosthifragments. London: British Library.
Schmithausen, Lambert (1985). Beitrdge zur Schulzugehdrigkeit und Text- geschichte kanonischer undpostkanonischer Materialien. Zur Schul- zugehdrigkeit von Werken der Hinayana-Literatur. Gottingen: Van- denhoeck & Ruprecht.
Wayman, Alex (1974). The Intermediate-State Dispute in Buddhism. In Buddhist Studies in Honour of I. B. Horner. L. Cousins, A. Kunst and K. R. Norman. Dordrecht/Boston: D. Reidel Publishing Company, 227-239.
Wijesekera, OH de A. (1945). “Vedic Gandharva and Pali Gandhabba." Uni­versity of Ceylon Review 3, 73-107.
Yinshun ERffli (1981). = shuo yi qie you
bu wei zhu de lun shu yu lun shi zhi van jiu. Taipei: Zhengwen.
(1983). = Zu a han jing lun hui bian. Taipei: Zheng­wen.
(1988a). E[= Yin du zhifo jiao. Taipei: Zhengwen.
(1988b). = Yuan shi fo jiao sheng dian zhi ji
cheng. Taipei: Zhengwen.