“The Doctrine of *Amalavijñāna in Paramārtha (499-569), and Later Authors to Approximately 800 C.E.”

“The Doctrine of *Amalavijñāna in Paramārtha (499-569), and Later Authors to Approximately 800 C.E.”

Michael RADICH

1. Introduction1

Like other ideas in his corpus, Paramārtha’s2 (Ch. Zhendi 真諦, 499-569) notion of amoluoshi 阿摩羅識, or *amalavijñāna, “taintless consciousness”, occupies an

1 Acknowledgements : I have been most fortunate in that I was able to use unpublished research materials circulated in the research seminar “Shintai sanzō to sono jidai 真諦三蔵とその時代” [“Paramārtha and His Times”] (April 2005-March 2010), coordinated by Prof. FUNAYAMA Tōru 船山徹, Institute for Research in Humanities, Kyoto University. Especially helpful were collected fragments of the Jiushi zhang cited in extant texts collected in ŌTAKE 2007(b), and studied by YOSHIMURA Makoto. I am also grateful for the invaluable opportunity to participate in this seminar. Its members have helped me tremendously. Above all I must thank Prof. FUNAYAMA for kindnesses too numerous to list. I also owe special thanks to Dr. ŌTAKE Susumu 大竹晋 for much valuable help. Prof. MUROJI Yoshihito 室寺義仁, Prof. YOSHIMURA Makoto吉村誠, Dr. IKEDA Masanori 池田将則 and Ching KENG 耿晴 saved me from a number of errors, and pointed me in helpful directions. I am grateful for funding from the Kyōto University Institute for Research in Humanities, which made it possible for me to travel to Japan and present a lecture summarising the present research, “Shintai ni okeru amarashiki to sono igi 真諦における阿摩羅識とその意義,” on October 18 2008. I thank Richard STANLEY for permission to cite his unpublished ANU doctoral dissertation. Finally, I thank Eunsu CHO for helping me to fi nd a useful article at the eleventh hour. Naturally, responsibility for any remaining errors is entirely my own.

Conventions: In citing the Taishō (T) and Xuzangjing (X) Chinese canons, I give the number of the text, followed by the volume, page, register and line nos., thus: T1616:31.863b05. Throughout, I regularly repunctuate citations from Chinese canonical texts without notice. I have aimed for this repunctuation to show clearly my interpretation of each passage.

2 For purposes of argument, I will throughout this paper use “Paramārtha” (abbreviation: P) to refer indifferently to the historical person and also to the corpus attributed to him. In so doing, I am eliding important problems in determining authorship of these texts. I have in preparation a study in which I attempt to use methods of computer-assisted statistical analysis to examine this problem of authorship. important place in the common understanding of the development of East Asian Buddhist thought. In particular, it is frequently linked to claims about the “sinifi cation”, or “making Chinese”, of Buddhist ideas. It has also often been interpreted as an attempt to forge links between Yogācāra and Tathāgatagarbha thought, that is, to bring about a synthesis between two major strands of Mahāyāna Buddhist doctrine. For these reasons, an accurate understanding of *amalavijñāna is important to our understanding of Buddhist doctrinal history. Towards this end, this paper studies primary sources for the doctrine of *amalavijñāna in detail, fi rst in Paramārtha’s extant corpus, and then in other sources to the close of the eighth century.

In Sections 2 and 3, I present a full analysis and translation of all passages in Paramārtha’s extant corpus mentioning *amalavijñāna, containing in total approximately twenty instances of the concept. I interpret each passage in relation to its context, and with full reference to available Chinese, Sanskrit and Tibetan parallels. I also present an analysis of the generally neglected Abhidharmakośa passage in which the word amalavijñāna occurs for the only known time in Sanskrit, and its relation to Paramārtha’s concept.

In Section 4, I present an analysis of what is said by more than twenty-fi ve authors, in several dozen texts, in about the first two and a half centuries after Paramārtha, comparing this evidence carefully with what is learnt about *amalavijñāna from examination of Paramārtha’s extant works. I proceed in three main steps, examining in turn: (1) claims about *amalavijñāna that are found only in later authors, and are not matched in Paramārtha’s works; (2) areas of overlap between later authors and Paramārtha; and (3) aspects of Paramārtha’s doctrine that are never repeated in later works.

Finally, in Section 5, I summarise the conclusions of this study. My main contentions will be as follows. (1) The neglected Abhidharmakośabhāṣya passage surrounding the mention of the word amalavijñāna does have some clear relations with Paramārtha’s idea of *amalavijñāna, especially as found in the Jueding zang lun. (2) In Paramārtha’s own works, we find not one but two largely distinct doctrines of

• amalavijñāna; one featuring in Jueding zang lun alone, and the other in the remaining works where the concept occurs. Of these two versions of the doctrine, it is likely that the Jueding zang lun is closer to the original doctrine. (3) There is relatively little overlap between Paramārtha’s own attested doctrine(s) of *amalavijñāna, and characterisations of the doctrine in later authors ― even when those authors purport to be describing the doctrine as Paramārtha taught it. (4) This suggests that Paramārtha’s ideas may sometimes have been misunderstood and misrepresented by the tradition. Later authors may thus not be very reliable sources of information about Paramārtha’s thought, and due to over-reliance upon them, our understanding of Paramārtha’s though may not be entirely accurate.

2. Paramārtha’s concept of *amalavijñāna in the primary sources

To my knowledge, there exists no complete study of Paramārtha’s concept of *amalavijñāna.5 In this section of this study, I will analyse all the instances of the term in Paramārtha’s corpus. We will fi rst examine the passages in which the concept occurs one by one. I will then attempt to summarise the doctrine of *amalavijñāna as it appears in these sources into a synthetic, general picture.

5 There is certainly no such study in any Western language. By “complete study”, I mean a study that takes full account of all the instances of this term in P’s corpus. Some Japanese and Chinese studies do survey nearly all of the sources I will treat below, but typically do not suffi ciently put passages in context; and I believe every scholar except YOSHIMURA misses at least some instances of the term. Few scholars have studied P’s texts against their parallels in reference to this question. Perhaps the closest study of primary passages is IWATA (1972[a]), which presents nearly all of the primary passages in which the term amalavijñāna appears in P’s extant corpus, and identifi es parallel terms in Indic texts. However, IWATA does not mention the four instances of the term in SWXL. Further, his work is largely restricted to translating single sentences and identifying Sanskrit parallels, and he pays far too little attention to the larger conceptual contexts. IWATA has further studied *amalavijñāna extensively, and often redundantly, in a long series of other articles, listed in my bibliography. These articles typically do not add anything not already said in this 1972 article. YOSHIMURA (2007a) surveys all pertinent passages, but only briefly, in preparation for consideration of later Shelun doctrines. Other important studies include relevant portions of KATSUMATA; Yinshun; LÜ Cheng’s essay on *amalavijñāna collected in Lü Cheng foxue lunzhu xuanji; and comments by Shengkai in his Shelun xuepai yanjiu. See also UI 6, 486-488, 539, 753-754; MOU 350-351, 355; FUKAURA 1, 338, 341-344; YE 15, 247, 253-255, 474; HAKAMAYA 10-13, 17. The most important Western language studies of *amalavijñāna to date are undoubtedly FRAUWALLNER; and GIMELLO, 277 ff. (“The Early Chinese Appropriation of Yogācāra and Tathāgatagarbha Buddhism”). See also LA VALLÉE POUSSIN (1928-1929), 109-113; DEMIÉVILLE (1952), 56 ff.; LIEBENTHAL 369 ff.; RUEGG 439-444 (“L’Amalavijñāna”); BUSWELL (1995), 77, 92-104; LUSTHAUS 369 ff; p. 379-380 n. 46; LAI, 76; PAUL 108, 145, 149; several pieces by YOSHIMURA (2002, 2007a, forthcoming).

The term 阿摩羅識 *amalavijñāna is not found in any parallels to the Paramārtha texts in which it appears. However, there are extant Sanskrit, Tibetan and Chinese parallels to passages in which the term appears, which sometimes even make it possible to identify terms to which it corresponds. These provide valuable clues to the meaning of the term and its context, and I have attempted to make full use of them.6 2.1 Amalavijñāna in AKBh

It has seldom been observed in secondary literature on *amalavijñāna that the term amalavijñāna is in fact attested in Sanskrit ― though not in the sense Paramārtha famously gives to it; and not in parallels to any of the passages where he expounds his 6 In utilising parallels to interpret P’s Chinese, I have tried to bear in mind two principles, which pull in opposite directions. (1) P most likely usually had before him a Sanskrit text that said something very similar to other versions of the text, and was trying to convey at least one plausible meaning of that text. Where possible, then, his Chinese must therefore be interpreted in a sense reconcilable with parallels. (2) As we shall see, it is also clear that P did at times depart from his source text in various ways, and this is corroborated by a general examination of his entire corpus. Therefore, where P’s text cannot plausibly be reconciled with parallels without doing violence to it, we must translate P, and not the parallels. amoluoshi 阿摩羅識.

The term occurs in Abhidharmakośa (AKBh) 5.28 ff. and the accompanying Bhāṣya (AKBh). The context is a discussion of when “latent tendencies” (anuśaya) attach to dharmas, how many attach to each dharma, of what kind they are, etc. AK here distinguishes between sixteen kinds of consciousness. The fi rst fi fteen kinds are associated with the three “realms” (dhātu: kāma, rūpa, ārūpya); the sixteenth, however, is “pure”, or “free of outfl ows” (anāsrava). Paramārtha uses several translations for this last “pure consciousness”.

The centre of gravity in this passage as a whole is this concept of “consciousness without outfl ows” (i.e. *anāsravavijñāna). Only in verses does Vasubandhu refer to this concept by the term amalavijñāna, and also by the epithet amala (twice, with vijñāna elided). In the prose Bh, by contrast, he consistently uses anāsrava. It is thus likely that he uses amala etc. metri causa. For Vasubandhu, the rare term amalavijñāna was thus most likely a nonce coinage, a mere poetic equivalent to *anāsravavijñāna.

Thus, the term amalavijñāna itself occurs only once in Vasubandhu’s Sanskrit: duḥkhahetudṛgabhyāsapraheyāḥ kāmadhātujāḥ | svakatrayaikarūpāptāmalavijñānagocarāḥ. (5.29) “The [[[dharmas]]] produced in the kāmadhātu that are to be abandoned by insight into [the Truths of] Suffering and the Origin [of Suffering]/ are the objects (gocara) of three [kinds of consciousness proper to] their own realm; of one [kind of consciousness] obtained in the rūpadhātu; and of pure consciousness (amalavijñāna).”

Paramārtha translates: 見苦集修滅 / 是欲相應法 / 自界三一色 / 無垢識境界.12 Strikingly, Paramārtha does not translate amalavijñāna 阿摩羅識, although it is the presumed underlying term for his own notion of *amalavijñāna. The term he does use, wugoushi 無垢識, only ever appears this once in Paramārtha’s corpus.

In 30ab, this “pure consciousness” is identifi ed merely by the epithet amala (with vijñāna elided); Paramārtha translates jingshi 淨識. Paramārtha uses this term twice in SWXL<1> and SBKL<2> to discuss *amalavijñāna. He also uses a closely related locution in JDZL<4>.

In a summary verse that concludes the discussion, the “pure consciousness” is again called amala; here, Paramārtha translates wuliu 無流, more usually his translation for anāsrava. This translation is far from arbitrary, but rather refl ects the real centre of gravity of the concept at issue. Aside from the exceptional (probably metri causa) instances of amala etc., Skt. too usually calls the same “pure consciousness” anāsrava (with “consciousness” elided but comprehensible from context). In one instance, Paramārtha translates wuliushi 無流識. Elsewhere, Paramārtha translates several times 無流心, for which we might expect Skt. *anāsravacitta.  

In this AKBh passage, then, the single instance of amalavijñāna clearly does not mean precisely what *amalavijñāna = amoluoshi comes to mean in other passages in Paramārtha. Neither is the concept of “consciousness without outfl ows”, whose discussion forms the larger context, identical to Paramārtha’s *amalavijñāna. This may be why scholars have tended to disregard this passage when studying Paramārtha’s concept of *amalavijñāna. Yet this passage may provide us with clues to part of the background of Paramārtha’s concept. Certainly, we can at least be sure from this passage that Paramārtha knew Vasubandhu’s use of amalavijñāna. It is therefore likely that when he elaborated his own notion, he was picking up on Vasubandhu’s term and infusing it with new content.19

Neither is the meaning of amalavijñāna in AKBh entirely unrelated to Paramārtha’s *amalavijñāna. As we will see below, two key parts of Paramārtha’s doctrine of *amalavijñāna are: (1) that it is free of defi lements (kleśa), a claim which is related to freedom from “outflows” (āsrava);20 and (2) that the realisation of *amalavijñāna brings freedom from the attachments that condition future rebirth, which also seems to be a consideration at play in the present AKBh passage’s consideration of anuśaya. These factors also make it likely this AKBh passage formed part of the background to Paramārtha’s coinage of his own *amalavijñāna.

To conclude, it is likely that: (1) Vasubandhu coined amalavijñāna as a nonce equivalent, metri causa, for *anāsravavijñāna; and (2) Paramārtha picked this rare term up as a label for his own ideas, and bent it fl exibly to that use.

We now turn to the consideration of “*amalavijñāna proper”, that is, the term *amalavijñāna as it is used to articulate the distinctively Paramārthian notion of a “pure consciousness” in the context of the Yogācāra system, usually represented by the transcription amoluoshi 阿摩羅識. The term appears in four texts: Jueding zang lun, Shiba

PRADHAN 302 samāsata ime ṣoḍasa dharmāḥ kāmarūparūpyāvacarāḥ pañcaprakārāḥ anāsravāś ca, T1559:29.260a09; but I think, as VP’s translation suggests (“dharmas pur” VP 4, 67), that this is in error for anāsravāḥ dharmāḥ. 19 This seems to be a time-honoured technique ― a thinker picks up a rare or unusual (and therefore suitably ill-defi ned) term from some nook or cranny, and reshapes it to fi t the new concept. Examples might include the standard concept(s) of dharmakāya; Zhiyi’s 一念三千; Yogācāra and Tathāgatagarbha uses of prakṛtiprabhāsvaracitta; and even Mahāyāna śūnyatā. 20 I n fact, the last paragraph of the passage I call JDZL<4> below, following the fi nal mention of *amalavijñāna in the text, explicitly mentions anāsravavijñāna. kong lun, Zhuanshi lun, and San wuxing lun. 2.2 Jueding zang lun

The term *amalavijñāna occurs most frequently in four passages in the Jueding zang lun 決定藏論 (JDZL), T1584.21 2.2.1 JDZL <1>

• Amalavijñāna occurs, most importantly, eight times in one short passage in JDZL. This passage is thus the most important source for information about what Paramārtha meant by the term. Signifi cantly, the context is a discussion of the “revolutionary transformation of the basis” (āśrayaparāvṛtti) of Yogācāra theory, and the relationship between it and the “storehouse” or “base” consciousness (ālayavijñāna).

“All that is included under [the head of] defi lements (煩惱, *kleśa) in the category (? 種) of karmic conditioning [i.e. ‘volition’, saṃskāra (-skandha), 行] is gathered together in the fundamental consciousness (ālayavijñāna); [thus, when,] because of intensive and repeated cultiva- 21 JDZL is a freestanding translation of a portion of YBh corresponding to the fi rst portion of the Viniścayasaṃgrahaṇī (XZ She jueze fen 攝決擇分). For Tib, I have referred to the Tokyo University reprint of the Derge version (Tōkyō daigaku . . . ed). Important secondary studies, which were helpful to me in preparing the present article, are HAKAMAYA; UI 6, 541-790 (“Kettei zō ron no kenkyū”). MATSUDA Kazunobu has identifi ed parallels to nearly half this text in a set of Sanskrit fragments in Nepal (see MATSUDA). According to MATSUDA’s account of the manuscript, it corresponds to JDZL 1025c26-1035a18 (extrapolating from equivalences MATSUDA gives to portions of the XZ YBh). This should mean that it contains a parallel to one instance of the term amalavijñāna, that at 1031a02-04, though MATSUDA also notes that the manuscript is damaged and it will not be possible to reconstruct a complete text on its basis, 18. However, it seems unfortunately that these Sanskrit fragments have not yet been published. tion (āsevita-bhāvita)26 on the part of the wisdom that takes Thusness as its objective support (tathatâlambanajñāna),27 the fundamental consciousness is abandoned, which is to say that a revolutionary transformation (*parāvṛtti) is brought about in the nature of the ordinary worldling (凡夫性, *pṛthagjanatva),28 such that all the qualities of the ordinary worldling (凡夫法, *pṛthagjanadharmāḥ) are discarded, then the fundamental consciousness is extinguished. Because this consciousness is extinguished, all defilements are extinguished, and by means of [the] counteragent[s] (對治, pratipakṣa) to the fundamental consciousness, the *amalavijñāna is realised (cf. Tib/XZ: “the basis undergoes a revolutionary transformation”, āśrayaḥ pravar- tate).

“The fundamental consciousness is impermanent (anitya), and is something attended by ‘outflows’ (有漏法, *sâsravadharma); [whereas] *amalavijñāna is permanent (nitya), and is something devoid of ‘outfl ows’ (無漏法, *anāsravadharma). [This is because] *amalavijñāna is 26 Tib. kun tu brten cing goms par byas, Skt. following HAKAMAYA 66. 27 真如境智, Tib. de bzhin la dmigs pa’i shes pas. 28 For *pṛthagjanatva for fanfuxing 凡夫性 in P, see e.g. HIRAKAWA 1, 241, which shows it is the only Skt. term so translated by P in the context of AKBh. Within JDZL itself, fanfuxing (XZ yishengxing 異生性) corresponds to Tib. so so’i skye bo nyid = *pṛthagjanatva at T1584:30.1024c1115 = T1579:30.587b25-29, D 23b4-5. (I am grateful to Ching KENG for pointing out the latter parallel to me.) It is thus implausible to reconstruct *pṛthagjanagotra with GIMELLO 326. realised 證 by means of the attainment of the path that takes Thusness as its objective support (真如境道, de bzhin nyid la dmigs pa’i lam, *tathatâlambanamārga).34 “The fundamental consciousness is accompanied by ‘badness’ (dauṣṭhulyasamanvāgama), [that is,] suffering as an effect [of karma] 苦果; *amalavijñāna is free of all ‘badness’ [i.e.] suffering as an effect [of karma]. “The fundamental consciousness is the basis 根本 for all the defi lements 煩惱, but does not act 作 as a basis for 根本 the noble path (聖道, *āryamārga). *Amalavijñāna, on the other hand, is

not the basis for the defi lements, but only 但 acts as a basis for the noble path and the attainment of the path. The *amalavijñāna acts as a ‘cause for the perdurance’ (*pratiṣṭhā-hetu) of the noble path, but does not act as a ‘cause for the generation’ (janma-hetu) [of it]. “The fundamental consciousness does not exert controlling power (自在, vibhutva) over good and neutral [[[dharmas]]; 善無記, kuśala, (kuśalâkuśala-) avyākṛta]. 34 Tib. reads, ” . . . because it transforms by [means of] the path that takes Thusness as its objective support”, de bzhin nyid la dmigs pa’i lam gyis bsgyur ba’i phyir ro. XZ: ” . . . because it is only possible for the basis to undergo revolutionary transformation on the basis of the noble path that takes Thusness as its objective support”, 緣真如境聖道、方能轉依故. “When the fundamental consciousness is extinguished, then things appear different.40 That is to say, the causes of bad [[[dharmas]]] and defi lements in future existences have been extinguished, and because these causes have been extinguished, the suffering of the fi ve ‘rampant skandhas’ 五盛陰苦41 will therefore not arise again in future existences. In this present existence, the bad causes of all defi lements are extinguished, and thus the aggregates of the ordinary worldling (凡夫陰, *pṛthagjanaskandha)42 are extinguished. [The practitioner obtains] controlling power (*vibhutva) [with regard to] the body of this [[[Wikipedia:present|present]] existence] 此身自在, [it] being therefore

whereas āśrayaparāvṛtti does exert controlling power over all good dharmas and dharmas unspecifi ed,” kun gzhi rnam par shes pa ni dge ba dang lung du ma bstan pa’i chos rnams la dbang mi byed la/ gnas gyur pa ni dge ba dang lung du ma bstan pa’i chos thams cad la dbang byed pa’o. XZ agrees. 40 阿羅耶識滅時、有異相貌. There is nothing corresponding to this sentence in Tib. or XZ. 相貌 also strikes me as somewhat colloquial. I therefore suspect that this might be a record of an explanatory lecture comment (see n. 98). 41 The term wu shengyin ku 五盛陰苦 is relatively rare. This is the only attestation of the term in the extant P corpus. The term first ever appears in the North in Narendrayaśas (under the Sui), T397:13.262b25; prior to that is confi ned to the South. 42 T he term fanfuyin is very rare in the canon (only fi ve instances), and other than here, never occurs in a translation text. Other instances are in Huisi 慧思 (515-577), Jingying Huiyuan, and Zhanran. Tib. and XZ parallels (see below n. 47) do not discuss the skandhas in this context. However, the skandhas are linked to āśrayaparāvṛtti (for which P is consistently substituting *amalavijñāna throughout the present passage) in MSg, which was arguably the most important of all Mahāyāna treatises for P. See RADICH §5.2.10, pp. 1159-1162. It is further notable that in MSg, the parāvṛtti (“revolutionary transformation”) of each of the skandhas in turn is described as leading to various special kinds of “controlling power” (vibhutva) proper to the Buddhas. Voluntary control over lifespan is not included among these powers, but there may be a connection between P’s incorporation of the skandhas here, MSg’s similar connection of skandhas to āśrayaparāvṛtti, and the fact that P, uniquely among our versions of the text, also talks here about controlling power (vibhutva) over lifespan (see n. 43 following).

like a magical creation 如化 (nirmāṇa). This is because [the practitioner] has abandoned all bad karmic consequences, and attained the cause and conditions of the *amalavijñāna; [he] thereby attains controlling power over the life force (jīvita) of the present body, [whereby he] can extinguish the causes and conditions of the life force (jīvitapratyaya ) in the body, and can also sever [his] lifespan [so that it is] completely extinguished with no remainder (*nirupadhiśeṣa). [By this same process,] all sensations (vedanā) are rendered pure, and so forth, as a sūtra explains in detail. tirely under the voluntary power of its possessor. In this connection, it is most likely signifi cant that elsewhere in YBh itself (in the fi nal chapter of BBh), voluntary power over lifespan is connected precisely with āśrayaparāvṛtti, for which P is here substituting *amalavijñāna; see SA- KUMA (1990) 2, 150-151; WOGIHARA 384; RADICH §5.2.6, pp. 1134-1135.

Parallel texts say nothing about voluntary power over lifespan here. Tib. has only, “The body that is like a magical creation continues to exist,” sprul ba lta bu’i lus kun tu gnas pa; and XZ similarly has “Although the body remains, it is like a magical creation” 其身雖住猶如變化. This would seem to be an interpretation of the passage closer to the notion of “Nirvāṇa with remainder” (sôpadhiśeṣanirvāṇadhātu) than P’s. It is thus remotely possible that P’s text is intended to convey a similar meaning, in which case we could also read zizai, unusually, as meaning something like “the body [of the present existence] continues to exist under its own [momentum]”. “Thus, we should know that it is (1) by means of thorough comprehension and analysis; (2) by means of the cultivation of wholesome thoughts; and (3) by means of the realisation of the *amalavijñāna48 that the fundamental consciousness and defilements together are extinguished.”49

We may summarise the doctrine of *amalavijñāna in this rich passage as follows.

cause [these conditions still] exist, [the person] continues to experience sensation (vedanā) up to the limits of his body and lifespan (kāyaparyanta, jīvitaparyanta). This is why it is says in words such as these in the sūtras, ‘These various [ongoing] sensations in this existence are only experienced for so long as [this existence lasts],’ and so forth;” kun gzhi rnam par shes pa de’i spangs pa’i mtshan nyid ni de spangs ma thag tu len pa rnam pa gnyis spong ba dang/ sprul ba lta bu’i lus kun tu gnas pa ste/ phyi ma la sdug bsngal yang ’byung bar byed pa’i rgyu spangs pa’i phyir/ phyi ma la yang ’byung bar byed pa’i len pa song ba dang/ tshe ’di la kun nas nyon mongs pa’i rgyu thams cad spangs pa’i phyir/ tshe ’di kun nas nyon mongs pa’i gnas ngan len thams cad spong ba dang/ gnas ngan len thams cad dang bral shing srog gi rkyen du gyur pa tsam kun tu gnas so// de yod na lus kyi mtha’ pa dang/ srog gi mtha’ pa’i tshor ba myong bar byed de/ de’i phyir mdo sde las kyang ’di na de’i tshor ba thams cad de tsam gyis na yongs su gtugs par ‘gyur ro shes rgya cher ji skad du gsungs pa lta bu’o// XZ agrees almost perfectly.

I t looks as though two new factors have been introduced in P: (1) P has introduced the old doctrine of voluntary control over lifespan for the liberated person in the present existence (cf. n. 43); (2) P is paraphrasing the ideas of the paragraph, rather than translating closely; in this process, as we have seen (see e.g. n. 42, 43, 46), part of his concern may be to fi ll in his audience on doctrines (especially those pertaining to āśrayaparāvṛtti) which he knew to be contained elsewhere in YBh, but to which they had no access.

48 T his sentence is a summary of a large section of the preceding exposition. The fi rst two categories hark back to parts of the text we have not examined. Relevant for us is the fact that “the realisation of the *amalavijñāna” is the category under which the text summarises the entire section quoted. 49 「一切行種煩惱攝」者、聚在阿羅耶識中。得真如境智、增上行故、修習行故、斷阿羅耶識、即轉凡夫性捨凡夫法、 阿羅耶識滅。 此識滅故、 一切煩惱滅、 阿羅耶識對治故、 證阿摩羅識。阿羅耶識是無常、是有漏法；阿摩羅識是常、是無漏法。得真如境道故、證阿摩羅識。阿羅耶識為麁惡苦果之所追逐；阿摩羅識無有一切麁惡苦果。阿羅耶識而是一切煩惱根本、不為聖道而作根本；阿摩羅識亦復不為煩惱根本、但為聖道得道得作根本。阿摩羅識作聖道依因、不作生因。阿羅耶識於善無記不得自在。阿羅耶識滅時、有異相貌、謂來世煩惱不善因滅。以因滅故、則於來世五盛陰苦不復得生、現在世中一切煩惱惡因滅故。則凡夫陰滅、此身自在、即便如化、捨離一切麁惡果報、得阿摩羅識之因緣故。此身壽命便得自在、壽命因緣能滅於身、亦能斷命盡滅無餘。一切諸受皆得清淨、乃至如經廣說。一切煩惱相故、入通達分故、修善思惟故、證阿摩羅識。故知阿羅耶識與煩惱俱滅, T1584:30.1020b08-28. For Tib. and XZ parallels, as quoted or referred to in fns immediately above, see D zhi 8a2-8b4, T1579:30.581c08-24. See also HAKAMAYA 40-42, 65-67; SAKUMA (1990) 2, 155-161; part of the passage is also translated in GIMELLO 326.

• Amalavijñāna is realised when ālayavijñāna is abandoned through the operation of its counteragents. The process whereby these counteragents operate is equivalent to intensive and repeated cultivation on the part of the wisdom that takes Thusness (tathatā) as its objective support (ālambana); it is also spoken of as the attainment of the path that takes Thusness as its objective support. The abandonment and extinction of the ālayavijñāna amounts to a radical transformation in the being of the practitioner, which is identifi ed as “revolutionary transformation [of the basis]” ([[[āśraya]]-] parāvṛtti). The resultant state, in which *amalavijñāna is realised, is diametrically opposed to the state of the ordinary worldling: it is free of defilements; it is also free of all the qualities (dharmāḥ), the skandhas, and the very nature (xing 性) of the worldling (pṛthagjana). In this state, all causes of future suffering have been brought to an end. By contrast to the ordinary states of consciousness grounded in ālayavijñāna, this state is permanent; free of “outflows” (anāsrava); free of karmic conditioning (saṃskāras); free of “badness” (dauṣṭhulya); acts as a basis for the noble path; and exerts control over good and neutral dharmas. The state brought about by the realisation of *amalavijñāna is also characterised by power of control over the body and over lifespan.

Comparison with the parallels in Chinese and Tibetan allow us to add, more certainly than on the basis of Paramārtha’s text alone, that *amalavijñāna is clearly a functional equivalent of āśrayaparāvṛtti. This is confirmed by the fact that even in Paramārtha’s text, which does not specifi cally mention āśrayaparāvṛtti, *amalavijñāna is characterised as “free from ‘badness’ (dauṣṭhulya)” ― an old characterisation of āśrayaparāvṛtti itself. It is also confirmed by the fact that quasi-commentarial paraphrases in Paramārtha’s text highlight other ideas known to be connected to āśrayaparāvṛtti doctrine in Yogācāra lore, despite the fact that parallel texts do not mention those ideas ― the transformed relationship to the skandhas, power over the body and lifespan, and the transformation of sensation (vedanā).56 2.2.2 JDZL <2>

JDZL next mentions *amalavijñāna at the end of a discussion about the different kinds of seeds (bīja) possessed by three different classes of persons ― those still in bondage (bandhana), those on the path but requiring further training (śaikṣa), and those beyond further training (aśaikṣa). The end of this discussion notes that it has been based upon the fundamental consciousness that is “not established” (rnam par ma gzhag pa, *avyavasthita). The text then considers the alternative perspective, in which seeds are considered on the basis of the “established” (vyavasthita) fundamental consciousness. The discussion in this latter connection is very brief, but Paramārtha differs signifi cantly from parallel texts. Tib. and XZ merely say: “On the basis of the ‘established’ fundamental consciousness, in brief, it should be known that the seeds of all dharmas exist (yod) in/upon the basis of that [fundamental consciousness ], and they are to be known respectively as either seeds that have not yet been abandoned and seeds that ought not be abandoned [at all].

Thus, the discussion here is only phrased in terms of distinguishing between bad seeds and good seeds. One set needs to be abandoned but has not been yet, whereas the other set must be retained in order to attain to the liberated state.

Paramārtha puts the same point this way: “[In the perspective of] the established exposition [of the fundamental consciousness, we would rather say]: All worldly dharmas take the fundamental consciousness as their basis, whereas all transcendent dharmas (一切諸法出世間者, *sarvadharmāḥ lokôttarāḥ) and dharmas of the path not to be abandoned (? 無斷道法 *aprahāṇīyamārgadharmāḥ?) have *amalavijñāna as their ba- sis.”

In other words, Tib. and XZ’s YBh merely distinguish between different types of seeds, and specifi cally locate them in one and the same ālayavijñāna. Paramārtha, by contrast, speaks rather of two different types of vijñāna, one the basis for worldly and defiled dharmas, and the other the basis of transcendent (lokôttara) dharmas. Paramārtha substitutes this point for the original text’s assertion that it is important to distinguish between the two types of seeds; and he inserts this distinction in a place that seems to be speaking of ālayavijñāna from a kind of “ultimate” or “definitive” (vyavasthāna) perspective.

This passage thus adds to our picture of *amalavijñāna the detail that *amalavijñāna is the separate basis for transcendent dharmas (lokottaradharmāḥ), whereas worldly dharmas (laukikadharmāḥ) are based on ālayavijñāna. Further, it is perhaps implied that *amalavijñāna so defined is consciousness as it appears in the perspective of defi nitive truth. 2.2.3 JDZL <3>

JDZL next mentions *amalavijñāna in another discussion of “seeds” (bīja). The basic question at issue is a possible contradiction between the claim that all seeds are universally pervaded by “badness” (dauṣṭhulya), and the claim that there is a class of “transcendent” qualities (lokôttaradharmas) which lead to liberation. What seeds give rise to these lokôttaradharmas? The basic answer is that lokôttaradharmas are produced from a different class of seeds, which are based upon (ālambana) Thusness itself as their necessary condition, and thus circumvent entirely the order of “impregnated” (vāsanā) seeds and their attendant “badness”. The text then explains the difference between four classes of beings (those who have not attained nirvāṇa, Śrāvakas, Pratyekabuddhas and Buddhas/Tathāgatas) on the basis of the relationship in each between these seeds of lokôttaradharmas based upon Thusness and the two “obstructions” (āvaraṇa). The text summarises the difference between the two orders of seeds thus (Tib. and XZ): “It should be understood that the continuance (rjes su ’jug pa, *anuvṛtti) of lokôttaradharmas [once they have] arisen is due to the increased strength of the revolutionary transformation of the basis (gnas gyur pa’i stobs bskyed pa las, *āśrayaparāvṛttibalâdhānāt ). This [fundamental transformation of the basis] is the counteragent (gnyen po, pratipakṣa,‘antidote’) to the fundamental consciousness (ālayavijñāna), is [itself] without fundament (kun gzhi ma yin pa, *anālaya), is a realm/element without ‘outfl ows’ (zag pa med pa’i dbyings, anāsravadhātu), and is free of conceptual elaboration (spros pa med pa, *niḥprapañca).”

By comparison, Paramārtha reads: “The continuum (相續, *saṃtāna) produced by the lokôttaradharmas can only be established 住 on the basis of the *amalavijñāna, since 以 this continuum acts as the counteragent to the fundamental consciousness (ālayavijñāna); [this continuum is otherwise] itself without fundament (無住處, more literally ‘without a place wherein it is established’), a realm/element without ‘outfl ows’ (anāsravadhātu), with no deleterious function 無惡作務, and free of all defi lements (煩惱, kleśas).”

Once more, the contrast between ālayavijñāna and āśrayaparāvṛtti is at stake, and as in JDZL<1> above, *amalavijñāna is substituted for āśrayaparāvṛtti. New is the idea that the counteragent of ālayavijñāna is a continuum produced by transcendent dharmas (lokottaradharmāḥ), and that this continuum is based upon *amalavijñāna. 2.2.4 JDZL <4>

Finally, JDZL mentions *amalavijñāna once more within the context of another discussion of the groundless counteragent to the ālayavijñāna.68 Significantly, it is clear from all three versions that the original text is indeed discussing a category of pure consciousness. Although the term *amalavijñāna only occurs once, it is necessary to take into account the fairly long passage that comprises the context in which the term is used to fully appreciate what is at stake.

The passage fi rst asks how “the wise” (mkhas pa, *paṇḍita) [i.e. the Buddhas] can free themselves not only from the immediate passions of attachment and aversion, but even from the underlying latent bondage to the basic conditions of existence, which gives rise to karma and therefore to future existences. In answer, the text explains how 68 T he passage in full runs from T1584:30.1030c21-1031a15, corresponding to XZ T1579:30.595b06-c06, D 43a6-44a4. See also on this passage UI 6, 785; and HAKAMAYA 10-17. full liberation of the wise differs from the state of the householder (gṛhin) or the monastic (pravrajita) who is not liberated. Householders are bound to future re-existences by the twofold fetter of desire and aversion (malice, ill-will), in addition to acts by which they harm sentient beings.71 Unliberated monastics are bound to future re-existence by excessive attachment to precepts and rules, which they mistake for the truth.72

These two conditions comprise the foil against which the attainment of full liberation by “the wise” is explained. This is where *amalavijñāna comes into play. XZ and Tib. explain as follows: “By means of plentiful cultivation of the [‘supreme’ XZ only] counteragent, these twofold desirous latent tendencies (anuśaya) are abandoned, and because of this abandonment, one is freed from taking the four skandhas [such as] objective form, sensation etc. as the objective support (ālambana) because of defi lement; the continuum is thereby eternally (gtan du, *nityam) severed. The cessation of consciousness attended by latent tendencies (anuśaya) [brought about] by this severing of the continuum is not grounded (based, gnas, *āśritya) upon the bases of consciousnesses associated with form, sensation etc., because of the completely purifi ed conscious-

skan dhas are presumably the same set referred to later in the passage as “material form, sensation etc.” (P: 色等; XZ: 色受等; Tib. gzugs dang tshor ba la sogs pa). (2) Tib. speaks clearly throughout of bondage to these skandhas in terms of anuśaya (bag la nyal); XZ agrees in places (二種隨眠, 595b15; 有隨眠識, 595b17). (3) It is clear later in the passage that these four skandhas, as a set, are understood as a “basis” (P: [四]住處; XZ: 住, 安住; Tib. gnas pa = *āśraya, *sthāna etc.). (XZ also speaks unusually of the fourfold set as a fourfold embodiment 四身, 595b07, b13; for skandhas as “bodies” in earlier Chinese tradition, see RADICH n. 1617, and more generally §4.3.6 and p. 556 ff.) 71 di ltar khyim pa’i phyogs la brten pa dag ni brnab sems dang/ gnod sems kyi mdud pa dag gis yul la rjes su ’dzin pa’i rgyu las byung ba dang/ sems can la gnod pa byed pa’i rgyu las byung ba’i las kun nas slong bar byed do// 72 rab tu byung ba’i phyogs la brten pa dag ni tshul khrims dang/ brtul zhugs mchog tu ’dzin pa dang/ ’di bden no snyam du mngon par zhen pa mchog tu ’dzin pa’i mdud pa dag gis las kun nas slong bar byed do// The text goes on to specify the ways that such attachment to rules and precepts functions as an analogy to the cruder twofold defi lements of the householder, inasmuch as excessive valuation of the rules is functionally analogous to desire, and the corresponding disregard (apavāda, “degradation, deprecation”) of nirvāṇa is functionally analogous to aversion. Both of these errors are said (in Tib.) to be mere mental constructs (yid kyis rnam par rtogs pa) and therefore to function, as their householder analogues do, to bring about further karmically conditioned existence. (XZ here says somewhat cryptically that only the attachment to the fourfold “embodiment” comprising the fi rst four skandhas remains, because it is a product of imaginative construction belonging to the manobhūmi 當知四身繫唯在意地分別所生.) ness (rnam par shes pa rnam par dag pa) that comprises the counteragent (pratipakṣa) of that [basis]. Thus, this [[[transformation]]] is known as ‘the complete pacification (upaśānti) of the cause of the basis’. Because of the cessation of the cause, there will in future never occur any [initial] apprehension or ‘[coming to] fulfi lment’ (yongs su rdzogs pa, *paripūrṇa) of a body, nor any activation (’jug pa, *vṛtti) of a continuum; thus is it called ‘the complete pacifi cation of the basis’.

“That purifi ed consciousness (rnam par shes pa rnam par dag pa) which is attained as ‘the counteragent of that [basis]’ is [itself] said not to be a basis (gnas pa ma yin pa, *anāśraya?). From it as cause, therefore, it is not possible [for further re-existence] to develop (mngon par rgyas pa). Because of the intensive cultivation (bhāvanā) of the ‘Liberation Gate (vimokṣamukha) of Emptiness (śūnyatā)’, [this state] is deemed to be ‘unconditioned’ (*asaṃskāra, *anabhisaṃskāra). Because of the intensive cultivation of the ‘Liberation Gate of No Desire (smon pa med pa, *apraṇihita)’, it is deemed to be ‘completely satisfi ed’ (saṃtuṣṭa). Because of the intensive cultivation of the ‘Liberation Gate of No Marks (mtshan ma med pa, *animitta)’, it is a ‘basis’ (gnas pa). Thus, for the reasons given, from ‘there being no further development [of future re-existence]’ to ‘its being a basis’, it is liberation (rnam par grol ba, vimokṣa).”  

Paramārtha reads (in a somewhat more abbreviated presentation):

“Due to the cessation that comprises the counteragent to these defi lements, the desire to take material form etc. [i.e. the fi rst four skandhas] as an objective support (境, ālambana) ceases; and due to this cessation, the defi led consciousnesses that take the [fi rst] four [[[skandhas]]] as a basis no longer [fi nd] a basis (諸識有惑於四住處則不復住). Because these consciousnesses that comprise the counteragents [plural in original, 諸對治識] are truly purifi ed, we know that the basis is pacifi ed (*upaśānta); [and] because the cause (緣, *rgyu) ceases, there will in future never be any re-arising of apprehension or bringing to completion (具足 paripūrṇa) of a continuum. Thus is it called the ‘pacifi cation (upaśānti) of the basis and reason for [re-]existence’ 有緣住靜. “The *amalavijñāna, which is the counteragent (pratipakṣa) of temporal consciousness 世識, ltar na de ni gnas pa’i rgyu nye bar zhi ba yin par rig par bya’o// rgyu ’gags pa’i phyir phyi ma la lus yongs su ’dzin pa dang/ yongs su rdzogs par byed pa dang/ rgyun ‘jug pa rnams ‘byung bar mi ’gyur te/ de ltar na de ni gnas pa nye bar zhi ba yin no// de’i gnyen por gtogs pa rnam par shes pa rnam par dag pa gang yin pa de ni gnas pa ma yin pa zhes bya’o// de yang rgyu las mngon par rgyas par mi ’gyur ba yin no// rnam par thar pa’i sgo stong pa nyid yongs su bsgoms pa’i phyir mngon par ’dus ma byas pa yin no// rnam par thar pa’i sgo smon pa med pa yongs su bsgoms pa’i phyir chog shes pa yin no// rnam par thar pa’i sgo mtshan ma med pa yongs su bsgoms pa’i phyir gnas pa yin no// de ltar mngon par rgyas pa med pa nas gnas pa’i bar gyi phyir shin tu rnam par grol ba yin no// 從此以後、由多修習勝對治故、復能永斷貪愛身繫二種隨眠。由此斷故、煩惱所緣色受等境亦不相續、以究竟離繫故、由此所緣不相續故。有隨眠識究竟寂滅、於色受等諸識住中不復安住、由對治識永清淨故。是名「識住因緣寂止」。又由當來、因緣滅故、於內身分不取不滿、決定無有流轉相續。是名「識住寂止」。又復對治所攝淨識名「無所住」。由彼因緣故、名「不生長」。由善修習空解脫門故、名「無所為」。由善修習無願解脫門故、名「為知足」。由善修習無相解脫門故、名「為安住」。如是不生長故、乃至安住故、名「極解脫」。 is utterly pure, and is said not to be a basis (不住, anāśraya), and thus this consciousness cannot

T1593:31.118a24-27. In this context, shishi refers to the necessary condition of the unbroken continuity of the “continuum” saṃtāna/saṃtati of the sentient being through all its lifetimes or incarnations; 世識謂生死相續不斷識, T1595:31.181c12; 為明䱾生果報無始以來三世生死相續不斷故、須立世識, T1595:31.184a17-19.

S imilarly, the term occurs in this sense in the Xianshi lun (顯識論, *Khyātivijñāna-śāstra, XSL). XSL is a puzzling text for various reasons, but seems to expound a category of consciousness (“manifesting consciousness”, khyātivijñāna = 顯識) otherwise primarily known from LAS. Here, the term also appears as part of the same list of vijñapti. Here, however, we have the added twist that nine vijñapti are identified as types of *khyātivijñānāni, as in the text’s title (nos. 1 and 4-11). Against this, the remaining two types of vijñapti are identifi ed as two kinds of vastupratikalpavijñāna (“consciousness imagining differentiation between phenomenal things” = 分別識, a category also deriving from the same LAS contexts), i.e. (2) the embodied subject; and (3) sense [[[Wikipedia:organ|organ]]]; 「顯識」者有九種：一身識、二塵識、三用識、四世識、五器識、六數識、七四種言說識、八自他異識、九善惡生死識。其次「分別識」有 二種：一有身者識、二受者識, T1618:31.878c27-879a02. In other words, it seems that for XSL, *khyātivijñāna are the categories (識, vijñapti) in which consciousness 識 manifests itself (顯, /khyā) as apparent objects of experience, whereas vastupratikalpavijñāna are the categories in which consciousness appears as a pseudo-subject, which therefore is conceived of as the agent of false imagining (parikalpa, /kalp, cf. vastupratikalpa) of phenomena as existent “things” (vastu).

F inally, the term also appears in P’s translation of Ratnâvalī 1.97: “Such temporal [Tib. ’gro < loka, thus *laukikadharmāḥ?]dharmas/ Are the fuel for the burning of consciousness/ With its due portion of the light of discrimination/ This fuel of temporal consciousness blazes, and then fades away;” 如是等世法 / 是然識火薪 / 由實量火光 / 世識薪燒盡, T1656:32.495b21-22, corre- sponding to Tib, rnam shes me yi bud shing ni/ ’gro chos ’di kun yin par ’dod/ de dag ji bzhin rab ’byed pa’i/ ’od dang ldan pas bsregs nas zhi, HAHN 39. This is the only other instance in which the word is used in anything like the sense here in JDZL, i.e. as referring simply to consciousness rather than vijñapti.

Returning to JDZL, shishi seems is intended here to convey two implications about the ordinary defi led consciousness to which *amalavijñāna is the counteragent: (1) it points to consciousness as it is related to the continuum (rgyun, *saṃtāna, *saṃtati: see also the immediate sequence of the present JDZL locus) of existence bound to the ordinary “world” saṃsāra; (2) it connects to shishi in the vijñapti list, where shi refers specifi cally to time (the “three times” 三世 of past, present and future), indicating that there is not only something “worldly” about this consciousness, but that its worldliness is intimately related to the very stuff of time. My translation as “temporal consciousness” (more literally “world consciousness”) is intended to convey some of this ambiguity: “worldly” and “bound to time”. (There may even be an implication that the “counteragent”, i.e. *amalavijñāna or the “purifi ed consciousness” is atemporal in the sense that it is timeless, i.e. eternal.) This concept of “temporal consciousness” may be an echo of the old notion that consciousness is the origin (samudaya) of the suffering world (loka) (see below n. 175), connected to the Yogācāra idea of the bhājanaloka 器世 (“container world”) etc. function as the cause of further existence. Because of the thorough cultivation of the ‘Liberation Gate (vimokṣamukha) of Emptiness’, it cannot give rise to karma 不能生業. Because of the thorough cultivation of the ‘Liberation Gate of No Desire’, it ‘knows contentment’. Because of the thorough cultivation of the ‘Liberation Gate of No Marks’, it is based in the immovable 住於不動. For these foregoing four reasons, [it is equivalent to] the attainment of full liberation.”80

“[A consciousness that is pure in this manner, even when it] sees the metamorphoses in phenomenal things 觀行於塵, does not have any attachments to [notions of] ‘I’ and ‘mine’ (aham iti, mamêti, ātmâtmīya, etc.), and thus, even when visible form (rūpa), etc. [i.e. phenomenal dharmas] are destroyed and pass out of existence 滅壞, the mind does not [feel] any hunger 渴愛 [for them; for more of the same]. In these respects 此諸相, the mind is utterly pure 心極清淨. Because consciousness is pure 識清淨, [it] does not pass out of existence of its own accord 不自滅壞, nor is it destroyed by other conditions [external to it] 亦復不為他緣所滅. Because there is [thus] no [longer any] continuum (相續, saṃtāna), it is not reborn again into the places of the ten directions, and it does not hanker after life and death; thus it is called ‘desireless’ (P. nicchāta, “without cravings”). [If] we liken the mind to a tree, and sensation (受, 80 此諸煩惱對治滅故、 欲取色等以為境者即得永滅。 以此滅故、 諸識有惑於四住處則不復住。諸對治識實清淨故、如是得知、住處寂靜。以緣滅故、於未來世當生具足應得相續、不復更生。是名「有緣住靜」。阿摩羅識對治世識甚深清淨、說名「不住」。復次此識不為緣生。空解脫門善修習故、不能生業。無願解脫門善修習故、則能知足。無相解脫門善修習故、住於不動。如前四義、得正解脫, T1584:30.1030c27-1031a07.

vedanā)83 to its shadow (chāyā), then at this time, neither exists [any longer]; [for] where the tree no longer exists, so, too, its shadow no longer exists 是故無樹是故無影. Because the temporal mind (世心, *laukikacitta) has been extinguished 滅, [this state] is called ‘complete cessation’ 盡滅 in the here and now [現; dṛṣṭe dharme/ihaiva]. With reference to 故 the gradual (次第, krameṇa) liberation of undefiled mind (mind ‘without outflows’, 無漏心, anāsravacitta) in which residual practice is necessary (學 . . . 解脫, *śaikṣavinirmukti), [this state] is said to have been ‘made peaceful’ (得寂靜, *śānta). With reference to liberation in which no further practice is necessary (無學解脫, *aśaikṣavinirmukti) it is said to have been ‘purifi ed’ 得清淨. 83 The mention of the cessation of vedanā in close proximity to talk of liberation “in the here-andnow”, “becoming cool”, etc., also recalls to mind the classic formula “all that is sensed right here, being not rejoiced in, will become cold”, idheva . . . sabbavedayitāni anabhinanditāni sītībhavissanti, as e.g. at Itivuttaka 44, WINDISCH 38; MASEFIELD 35. Thus, we see, running in order through the passage: “does not rejoice in” (Tib. only, mngon par dga’ ba med pa, cf. P. anabhinanditāni) [life and death]; “in the here and now” (cf. P. idheva); the cessation of the sensations (vedanā, cf. P. sabbavedayitāni); and the notion of coolness (śīta etc., cf. P. sītī bhavissanti). In its locus classicus in It 44, this formula is associated with the distinction between nirupadhiśeṣanirvāṇadhātu and sôpadhiśeṣanirvāṇadhātu, “Nirvāṇa with/without a remainder of attachment”, and it likely that this distinction is in the back of the YBh author’s (“Asaṇga’s”) mind here, too, given that (1) the text lays out a gradated schema of several types of liberation; and (2) it is discussing liberation as a process whereby consciousness frees itself from grasping at (other) skandhas, where nir/sa-upadhi was often interpreted precise as having/not having (a remnant of) grasping at skandhas = upādānaskandha.

Because the four remaining [skandhas] have been extinguished, [it is a state that has] attained the power of Brahman (得梵自在, for *brahmabhūta? 91.”92

better sense in light of the allusion to the Saṇgīti sutta running through the passage (see n. 82). It is difficult, then, to see what might have led P to translate “purified” here, since it dampens the resonance of the allusions the passage is based upon. See discussion below.

For brahmabhūta here, see parallels, and the Nikāya/Āgama passage cited n. 82.

觀行於塵、於我我所無所取著、是故、色等諸塵滅壞、心無渴愛。如此諸相、心極清淨。識清淨故、不自滅壞、亦復不為他緣所滅。無相續故、於十方處不更入生、於命於死無貪欲。故說「無求欲」。心譬如樹、受喻如影、於時二無、是故無樹、是故無影。世心滅故、說「現盡滅」。是無漏心學解脫故前次第、說「得寂靜」。無學解脫故、得清淨。四餘滅故、得梵自 在, T1584:30.1031a07-15. Comparison shows that the P translation is slightly scrambled and terse (whether this is because he was working from a different version of the text, or due to the translation process), especially towards its end; without the parallels, it would not be entirely possible to determine accurately the intent of some wording. XZ and Tib. here differ from one another in minor details, but the gist of both is the same. The following is a translation of XZ, noting key differences in Tib: “[Such a consciousness] does not grasp at ‘I’ and ‘mine’ (aham iti, mamêti; ātmâtmīya, etc.) with regard to any [of the things subject to] metamorphosis (行, rgyu bar gyur pa); thus, it does not feel fear (恐怖, Tib. ‘distress’, yi chad pa), even when visible form, etc. [i.e. the dharmas of the phenomenal world] undergo destruction (壞, Tib. ‘change, transformation’, gyur). In virtue of this feature (相貌, rtags), it is manifest that that [consciousness] has been purifi ed in its very essence (自體已得清淨, Tib. ‘is an essence that is pure[-ifi ed]’, rnam par dag pa’i bdag nyid du snang ngo). Moreover, because this consciousness has been permanently purifi ed (又由彼識永清淨故, Tib. because of the purity of this consciousness rnam par shes pa de ni rnam par dag pa nyid kyi phyir), it enters spontaneously into [the state of] tranquility (任運自然入於寂滅, Tib, ‘into cessation’ rang gi ngang kho nas ’gag par ’gyur), without any dependence upon other causes. Because the continuum of consciousness (識相續, rnam par shes pa de’i rgyun, tasya vijñānasya saṃtānaṃ) is thus cut off once and for all, it never again will tumble through (流轉, Tib. ‘enter’, ‘fall into’, ’jug pa) the worlds of the ten directions, and does not hanker after (希求, Tib. ‘fi nds no delight in’ . . . la mngon par dga’ ba med pa) life or death; it is therefore said to have left behind desire. Further, because all sensation (vedanā) is like a shadow to the tree of consciousness, and because that [consciousness], from that time on, will never again exist, it is said to have left behind its ‘shadow’. The extinction of defi led consciousnesses (諸無漏識, zag pa dang bcas pa gang yin pa, yat sâsravaṃ [vijñānaṃ, XZ] once and for all in the here and now (於現法中, tshe ’di la, *dṛṣṭe dharme/ihaiva), it is called ‘cessation’ (寂滅, mya ngan las ’das pa, nirodha). The gradual, stepwise (隨其次第, rim gyis, krameṇa) liberation of undefi led consciousnesses (consciousnesses without ‘outfl ow’, 諸無漏識, zag pa med pa gang yin pa, yad anāsravaṃ [vijñānaṃ]) in which residual practice is still necessary (隨其次第有學解脫 *śaikṣavinirmukti) is called ‘peace’ (寂靜, zhi ba, śānta). The liberation in which practice is no longer necessary (*aśaikṣavinirmukti) is called ‘cooled’ (清涼, bsil ba, śīta); because the basis [for it, viz. consciousness] in the other (餘依) [skandhas; Tib. phung po] has been permanently extinguished, [this consciousness] is called ‘purifi ed’ (清淨, tshangs pa gyur pa, brahmabhūta).” 又於行等都不執著我及我所、由此

This passage is very instructive. Looking just at XZ and Tib, YBh clearly does understand that: (1) ordinary consciousness, when associated with the other skandhas, is the base for ordinary defi led existence that is entangled in saṃsāra; (2) the counteragent to this base is a kind of purifi ed consciousness (*viśuddhaṃ vijñānam); (3) the liberation brought about by the operation of this counteragent is indeed a cessation of defi led consciousness; (4) this realisation equals the severing of the continuum (saṃtāna); (5) the realisation of this state guarantees that the realised being will never take incarnation, i.e. will never again in future apprehend a body, and thus will not suffer; (6) the purifi ed consciousness is itself not a basis (for future existence in suffering); (7) this purified consciousness is identical with liberation. In these features, the doctrine of YBh echoes the following features of Paramārtha’s *amalavijñāna doctrine: it posits a kind of pure consciousness which comprises the counteragent to a kind of defi led consciousness, which is the basis for ordinary, defi led existence; liberation entails the cessation of the defi led consciousness and the severing of the continuum;93 it involves the cessation of future suffering; and the cessation of future suffering is connected to the complete ending of embodiment.94 Paramārtha’s translation of this passage seems quite faithful, and adds little except that it names the purifi ed consciousness *amalavijñāna.

The passage adds the following to our picture of *amalavijñāna. (1) *Amalavijñāna

因緣、色等壞時亦不恐怖。由此相貌、顯彼自體已得清淨。又由彼識永清淨故、不待餘因、任運自然入於寂滅。此識相續究竟斷故、於十方界不復流轉。於命及死不希求故、名「永離欲」。 又所有受是識樹影、 彼於爾時不復有故、 名「永離影」。 諸有漏識於現法中畢竟滅盡故、名「寂滅」。諸無漏識隨其次第有學解脫、名「為寂靜」。無學解脫、名曰「清涼」。餘依 永滅故、說「清淨」 T1579:30.595b26-c06; rgyu bar gyur pa na yang bdag dang bdag gir cung zad kyang mi ’dzin to// des na de gzugs la sogs par gyur kyang yongs su yi chad par yang mi ’gyur ro// rtags des na de rnam par dag pa’i bdag nyid du snang ngo// rnam par shes pa de ni rnam par dag pa nyid kyi phyir bdag nyid rgyu gzhan la mi ltos par rang gi ngang kho nas ’gag par ’gyur ro// rnam par shes pa de’i rgyun rgyun chad pa’i phyir phyogs bcur ’jug pa yang med do// ’tsho ba dang ‘chi ba la mngon par dga’ ba med pa ni grib ma med pa zhes bya’o// yang na tshor ba rnams ni rnam par shes pa shing ljon pa lta bu de’i grib ma lta bu yin te/ de dag de’i tshe na mi ’byung bas/ de’i phyir yang de ni grib ma med pa zhes bya’o// zag pa dang bcas pa gang yin pa de ni tshe ’di la mya ngan las ’das pa yin no// zag pa med pa gang yin pa de ni rim gyis slob pa’i rnam par grol bas zhi ba yin no// mi slob pa’i rnam par grol bas ni bsil bar gyur pa yin no// phung po ’gag pa’i phyir tshangs par gyur pa zhes bya’o. In Tib, the presentation of the conceit: “consciousness = ‘tree’, sensation = ‘shadow’,” and the following epithets of the liberated state differ, but not in any way that affects the gist for our purposes here. I am grateful to ŌTAKE Susumu for saving me from some errors in the attempted reconstruction of Skt. equivalents here. Cf. JDZL<3> above. Cf. JDZL<1>, p. 55 above.

is explicitly said to be the counteragent of defi led consciousness. (2) *Amalavijñāna qua the counteragent is overtly said itself not to have (or be) a basis. (3) The passage makes explicit an association between the “consciousness” in question and the vijñānaskandha, and thus seems to clarify somewhat the connections between *amalavijñāna and the cessation of the skandhas already touched on in earlier passages.

The fi nal paragraph is especially important, even though it does not overtly mention *amalavijñāna.

Most strikingly, Paramārtha says the polar opposite to parallel texts. In XZ and Tib, the pure consciousness does not have to depend upon any other causes in order to cease, but rather enters into cessation of its own accord. At the moment of liberation, consciousness ceases to “exist”, and liberation consists in this cessation. In Paramārtha, pure consciousness “does not pass out of existence of its own accord, nor is it destroyed by other conditions [external to it].” In other words, the liberated, purifi ed consciousness ― *amalavijñāna ― is permanent, as we already saw at JDZL<1>. However, it is difficult to be sure what to make of this reversal in meaning. Given the sometimes slightly garbled state of the remainder of the text, which apparently betrays problems in translation, the reversal may result from simple error. This impression might be reinforced by the fact that even Paramārtha goes on immediately to say, ” . . . there is no longer any continuum (saṃtāna/saṃtati).” (At least so long as “continuum” refers to consciousness, this would appear to be in direct contradiction to the assertion that purified consciousness “does not pass out of existence”.) However, Paramārtha’s text nowhere else really departs from the gist of the original as seen through parallel texts. Why, then, only at such a crucial juncture? This divergence from the underlying text thus may be deliberate, and for this reason, a significant component of Paramārtha’s doctrine of *amalavijñāna ― one that sounds rather close to “eternalism” (śāśvatadṛṣṭi, nityadṛṣṭi ).

In this fi nal paragraph (in parallels as well as Paramārtha), the passage also explicitly connects the liberating realisation of pure consciousness to anāsravavijñāna/ anāsravacitta, which was a dominant theme in the AKBh amalavijñāna passage. This may show that there is indeed a conceptual link between the AKBh passage and Paramārtha’s use of *amalavijñāna.

Finally, this last paragraph reinforces the idea that the purifi ed consciousness is the vijñānaskandha. Liberation is the process whereby that consciousness is freed from attachment to the other four skandhas (aggregates). Here we hear echoes of a very old model, in which consciousness is the apparent subject of both transmigration and liberation. I take this model to be extremely pertinent to the doctrine of *amalavijñāna. We will return to this point below. 2.3 Shiba kong lun

The Shiba kong lun 十八空論 (T1616, hereafter SKBL), or “Treatise on Eighteen [Modes of] Emptiness”, is a text of a type that has been referred to as a “lecture text”. SBKL is clearly based upon two sections of the Madhyântavibhāga (MAV), from Chapters 1 and 3 respectively. It intersperses apparent citations from or paraphrases of that text with comment and expansion. Beyond its discussion of *amalavijñāna, the text is of great interest because it makes use of a number of apparently Chinese categories. Since Paramārtha separately translated the “root text”, MAV (中邊分別論, T1599), for at least one of the instances of *amalavijñāna, we are in the unusual position of being able to see how Paramārtha himself alternately interpreted the passage into which he inserts the term. We are also fortunate to have a sub-commentary (ṭīkā) on Vasubandhu’s MAV Bhāṣya by Paramārtha’s close contemporary Sthiramati. The term *amalavijñāna appears twice in SBKL. 2.3.1 SBKL <1>

• Amalavijñāna fi rst appears in SBKL corresponding to comment upon MAV 1.21-22. The text is discussing the category of prabheda-śūnyatā, “emptiness differentiated” [i.e. into different aspects or characteristics], before moving on to piṇḍârtha-śūnyatā, “emptiness in general”, i.e. a general summation of things that can be said of all the modes of emptiness collectively. In MAV, this corresponds to a section in which emptiness is “proven” as a doctrinal tenet. SBKL reads:

“Here begins part four, proofs (道理, *sādhana?) that emptiness is differentiated. There are three. “(1) [The proof that emptiness can be differentiated according to its] purity and impurity. (i) If we were to say that emptiness (śūnyatā) were absolutely 定 impure, then it would be impossible for all sentient beings to attain liberation, because, [[[emptiness]]] being absolutely impure, it could not be made pure. (ii) [On the other hand,] if we were to say that [[[emptiness]]] is absolutely pure, then there would be no point in cultivating the path (*mārgabhāvanā), because even before one had attained liberation at [the stage of] the path without taints, emptiness would already be aboriginally 本 and innately 自然 pure of its very substance 體. There would therefore be no defi lements to obstruct wisdom, nor anything that could extirpate [them], and all sentient beings would automatically attain liberation without relying on effort; [however,] it is evident that sentient beings do not [in fact] attain liberation without effort, and thus, emptiness is not absolutely pure. On the other hand, it is also the case that liberation is [in fact] attained by dint of effort, and thus we know that emptiness is not absolutely impure. This is the proof of puritycum-impurity and impurity-cum-purity 淨不淨不淨淨. “Additional comment: If we say that the principle (li 理) of emptiness is absolutely impure, then all efforts would be ineffi cacious 無果報, because the essential nature (自性, *svabhāva) of the element of emptiness (空界, *śūnyatādhātu) itself would be impure; and therefore, even when the path had arisen 生道, one would remain incorrigibly worldly/profane 俗不可除, so that the path would be useless. Because it is not thus 無此義故, we know that emptiness is not impure by nature.

“Question: If this is the case, then given that there is no impurity by essential nature 自性不淨, there should also be no purity by essential nature 自性淨. How can it be ascertained 分判 that the dharma-realm (法界, *dharmadhātu) is neither pure nor impure? “Answer: *Amalavijñāna is the aboriginally pure [Skt. ‘luminous’] mind (自性清淨心, prakṛtiprabhāsvaracitta).109 It is only because it is tainted 污 by adventitious dirt 客塵 that we speak of it as ‘impure’; because of adventitious dirt, [that is,] we establish that it is [also] impure.”

This passage adds to our growing picture of *amalavijñāna as follows.

• Amalavijñāna is now identifi ed with prakṛtiprabhāsvaracitta, i.e. the aboriginal “innate purity of mind”. It accounts for the pure aspect of emptiness, whereas the impure aspect is accounted for by adventitious defilements. Moreover, both the association with prakṛtiprabhāsvaracitta, and the association between *amalavijñāna and the dichotomy of intrinsic purity and innate defi lements, link *amalavijñāna to the pure Thusness of the Ratnagotravibhāga (RGV), and via RGV to tathāgatagarbha doctrine (and its Chinese offshoot, the doctrine of “Buddha Nature”, foxing 佛性). Further, the pure aspect of emptiness with which *amalavijñāna is identifi ed here is also identifi ed in turn with the dharmadhātu; this is the beginning of a process that links *amalavijñāna into a chain of identifi cations for (aspects of) the Mahāyāna “absolute”.

109 I WATA (1972[a]) claims that prabhāsvara . . . cittasya, i.e. something like prakṛtiprabhāsvaracitta, is the Skt. “original” for *amalavijñāna here, but it is clear from the fact that SBKL also gives a term clearly corresponding to prabhāsvaracitta that *amalavijñāna does not translate that term, but that rather, a correspondence is being explicitly asserted between the two terms. Further evidence that P does know the difference is that in the corresponding locus in his translation of MAV itself, he simply gives 心本自性清淨故 for this line (prabhāsvaratvāc cittasya), T1599:31.453b01.

The passage translated here comprises only part of a longer section. In the sequence, it becomes clear from the wording of the argument that, as above, emptiness is being identified with the dharmadhātu (cf. the unusual use of *śūnyatādhātu above, probably as a kind of intermediate term to this identifi cation),114 and with Thusness.115 All three are therefore implicitly also identifi ed with *amalavijñāna qua the “innate purity/luminosity of mind” (prakṛtiprabhāsvaracitta). The identifi cation of more than one of these terms as the topic or locus of absolute purity thus implies an identifi cation of the truest pure substance of mind with the truest substance of all things. 2.3.2 SBKL <2>

Strictly speaking, the second passage in SBKL may not in fact mention *amalavijñāna, but rather a “pure/luminous consciousness” (清淨心, prabhāsvaracitta), which is then qualifi ed as *amala. The passage in question corresponds to MAV 3.14 and Bh. MAV Ch. 3 treats various kinds of reality (tattva, also “truth”, “real” etc.). The ninth category under which it does so is prabheda-tattva, or “reality as it is differentiated”.116 This category in turn is divided into seven different aspects under which reality may be known, approached or apprehended.117 The third of these aspects is “the reality of representation only” (vijñaptitattva),118 which refers to the fact that in reality, all dhar-

China, especially in Huayan contexts. 114 863b24 , b27 etc. 115 These sections are not matched by anything in MAV. As it has for emptiness, the text ex- pounds purity and impurity for the dharmadhātu, 863b22-c05; and Thusness (tathatā), 863c05-c24 (note in this connection that Sth also brings in Thusness in his exposition of the passage in MAVṬ: tathatāyām āgantukair malaiḥ saṃkliṣṭatābhyupagantavyeti etc., YAMAGUCHI 1, 59.22-60.1). In each case, the basic proposition that the topic is “pure-impure” is expounded in different ways: in the case of the dharmadhātu, that it is “pure in some respects, and pure in others” 或淨或不淨; in the case of Thusness, that it is “both pure and impure” 淨不淨. The implicit identifi cation effected by this argument between emptiness, the dharmadhātu, and Thusness is important; that *amalavijñāna is also identifi ed with the pure substance or essence of emptiness implies a further identification between *amalavijñāna and all three of these instances. The use of 自性淨 to describe the dharmadhātu, e.g. 863b28-29, echoes the predication of 自性清淨 of mind/*amalavijñāna. 116 Note the parallel with the treatment of “emptiness differentiated” in SBKL<1> above. 117 These seven types of tattva are also found in other early Yogācāra texts in addition to MAV, including SdhN, as Sth’s MAVṬ notes; see SdhN 8.20.2, LAMOTTE (1935) 99, 219; and MSA 19.44 and Bh, LÉVI (1907, 1911) 1, 167-168; 2, 275; JAMPSAL et al., 304. 118 I here provide the equivalent found in MAV (where the orthography is in fact －tatva), which is certainly the more relevant here, in that MAV is the text upon which SBKL is based. Note, however, that in other sources (see n. 117), the category may also be called vijñaptitathatā (as mas are merely representations (vijñaptimātra). We will see below that Paramārtha also connected *amalavijñāna to this category of the third tattva in SWXL<1>.119 “Third, we explain the reality of representation only 唯識真實 (*vijñapti[-mātra]-tattva). [This consists in] discerning that all dharmas are only pure consciousness 淨識, such that there is no subject of ignorance, and also no object of ignorance 無有能疑亦無所疑,120 as is explained in detail in the Treatise on Representation Only 唯識論. There are two senses in which it is proposed that there is only representation. “(i) First, [in the perspective/stage of] practice (prayoga, 方便), [the doctrine of representa- the other items in this sevenfold rubric are －tathatā); see e.g. LÉVI (1907, 1911) 1, 168; and cf. the term in SWXL, for which see below n. 144. I am grateful to ŌTAKE Susumu for pointing out this variant. 119 Further on P’s interpretation of vijñapti-[mātra-]tattva, see SBKL T1616:31.864c29-865a03, and GIMELLO 322. 120 The distinction between nengyi 能疑 and suoyi 所疑 is unique to the present passage and a citation of it in the Zong jing lu (which erroneously attributes it to Kumārajīva’s 十二門論), T2016:48.609c03-09. (Chengguan, T1736:36.212b03-05, seemingly uses the two terms in a different sense.) tion only] refers to the perception that there is only ālayavijñāna, and no other objects beyond it 無餘境界. This results in the realisation of the dual emptiness of object and mind 境智兩空, and the complete extirpation of deluded consciousness. This is what is termed ‘representation only [in the perspective/stage] of practice’. “(ii) Second, representation only [in the perspective/stage] of perfect insight (正觀, *abhisamaya). [In this perspective,] we dispose of 遣蕩 both the deluded consciousness and mind of saṃsāra, and of its object, [such that] both are completely purifi ed, and there is only the taintless (阿摩羅, *amala) pure/luminous mind (清淨心, *prabhāsvaracitta).” however, that the translation of prayoga and even prayogamārga is, once more typically for P, variable even within this chapter; the more regular translation of prayoga throughout the chapter is jiaxing 加行.) This may be compared to P’s Bh to MSg 3.3: 一切法、謂有為無為、有流無流、及四界、三乘道果等。如此等法、實唯有識。何以故？一切法以識為相、真如為體故。

若方便道、以識為相。若入見道、以真如為體, T1595:31.200a19-22. Here, too, the contrast is between (the path of) practice (方便道, prayogamārga) and (the path of) insight (見道, darśanamārga); further, (the path of) practice is related to the realisation that the characteristic of all dharmas is consciousness/representation (shi), while (the path of) insight is related to the realisation that their substance (ti) is Thusness. Thus, it is likely that in correlation with zhengguan “perfect insight”, “true understanding”, fangbian is intended to refer to prayoga, “practice”. I am grateful to ŌTAKE Susumu for pointing me to this background.

This passage is the fi rst time we have seen *amalavijñāna associated with the doctrine that there is representation only. In Sanskrit, there is a distinction, however tricky, between “representation(s) (vijñapti) and “consciousness” (vijñāna). However, in this Chinese there is none ― both are shi 識. It seems clear that the present passage plays on this polysemy, and is predicated on thinking in Chinese: there is only shi, and that is of two kinds, ālayavijñāna (at the intermediate stage of practice) and *amalavijñāna (at the stage of insight). This exposition situates *amalavijñāna in relation to an interpretation of vijñaptimātra/weishi typical of the Paramārtha corpus (see e.g. ZSL below), here articulated in terms of the distinction between fangbian weishi and zhengguan weishi. As opposed to a halfway-house understanding of vijñaptimātra that hypostasises ālayavijñāna and imagines it to be all there is,126 *amalavijñāna is associated with an ultimate understanding in which ālayavijñāna is gotten rid of altogether (as above, e.g. in JDZL<1> and <3>, where *amalavijñāna is associated in various ways with its counteragents). This fi nal perspective is also associated with a non-dualism that transcends the subject-object dichotomy (here, of subject and object of ignorance). Finally, it is also worth noting that the overall context of this passage associates *amalavijñāna with reality or Thusness (tattva, tathatā), thus rendering direct a link that was only indirect in SKBL<1> above, where it was made via emptiness as middle term. 2.4 Zhuanshi lun (ZSL)

The Zhuanshi lun (轉識論, *Pravṛttivijñāna śāstra?127 T1587, hereafter ZSL) com-

is “perfected in [the sense of] being free of erroneous inversion” (aviparyāsapariniṣpattyā); YAMAGUCHI 1 135.9, STANLEY 179. 126 An interpretation which incidentally corresponds quite accurately to characterisations of Vijñaptimātra thought as “idealist”, frequently found in the secondary literature; a Vijñaptimātra response to charges of “idealism” might thus be that the term only characterises an imperfect or incomplete Vijñaptimātra. 127 The title of this text poses interesting problems. The term zhuanshi 轉識 does not actually appear anywhere in ZSL except the title; the vijñānapariṇāma that is the topic of the text is rather called (more logically) shizhuan 識轉. Aside from this, zhuanshi only appears in two places in P’s extant corpus: once in the possibly problematic Yijiao jing lun, T1529:26.285c13-14; and once in a passage in MSg and Bh, where it is part of a verb-object construction meaning “to transform the aggregate of consciousness”, 轉識陰依故, Tib. rnam par shes pa’i phung po [[[gnas]] su] gyur pa’i phyir (*vijñānaskandhâśraye parāvṛtti[-tvāt]?); T1593:31.130a22, T1595:31.253b27-28, LAMOTTE (1973) 1, 86.

On the other hand, although it is elsewhere also quite rare, the term zhuanshi does appear in the following series of texts. Beginning with Guṇabhadra, all three Ch. versions of LAS use the term for pravṛttivijñāna, pravṛtti or vṛtti, Tib. ’jug pa’i rnam par shes pa etc. (443 C.E.): in Guṇabhadra, T670:16.483a29-b3, 484a13-14, corresponding respectively to NANJIO 38.13-15, prises a translation of Vasubandhu’s Triṃśikā, with commentary seamlessly interwoven with the root text.128 In this sense, ZSL too may answer to the description of “lecture text”. As for SBKL, we have the possibility here of some direct comparison with Sthiramati’s interpretation, as his commentary on the Triṃśikā is also extant.129 Sthiramati also comments in detail upon the verse, sometimes to similar effect,130 but

44.8; in Bodhiruci (513 C.E.), in addition to passages corresponding to these Guṇabhadra passages at T671:16.522a16-20 and 523a19-23, also at 515a06-08=NANJIO 2.13, 523.c10-16=NANJIO 47.3-8, 538c02-04=NANJIO 126.18, 540b25-27=NANJIO 136.12, 559c01-04=NANJIO 235.17, 571c1213=NANJIO 300.11; and in corresponding loci in Śikṣānanda’s translation (700-704); see also SUZUKI (2000) 120, 412. A large number of these loci feature the conceit of the arising of the pravṛttivijñāna as “waves” upon the “ocean” of the ālayavijñāna: e.g. ebhir mahāmate caturbhiḥ kāraṇair oghântarajalasthānīyād ālayavijñānāt pravṛttivijñānataraṇga utpadyate, NANJIO 44.8 etc. The only other place it appears before P is in Bodhiruci’s Daśabhūmika sūtra śāstra T1522:26.172b17, where it corresponds to Tib. ’jug pa’i rnam par shes pa (= *pravṛttivijñāna), ŌTAKE (2005) 2, 488, 489 n. 10. Thereafter, the most important place where the term also appears is AF, where it is clearly derived from LAS in at least some instances (being associated, for instance, with 業識 karma[-lakṣaṇa-]vijñāna, 分別事識 vastupratikalpavijñāna, 現識 khyātivijñāna etc.; see T1666:32.577b06-12, HAKEDA (“evolving consciousness”) 48; 579b20-23, HAKEDA 69; 581a26-29, HAKEDA 87.

As ŌTAKE Susumu has pointed out (personal communication), in P’s own texts, the more usual term for pravṛttivijñāna is by contrast shengqishi 生起識: see MSg T1593:31.115c17, 116a03, (119c22, not in Tib.), 121b29-c03 = ’jug pa’i rnam par shes pa, LAMOTTE (1973) 1, 12-13, 42; MAVBh to 3.22, T1599:31.457c16-17, = pravṛttivijñāna, NAGAO 48. (The term shengqishi, further, is not found before P, and only a few times in XZ after him in translation works, and so is a strong marker of his genuine style.)

Thus, while the evidence is only circumstantial, it seems a term derived from LAS, but never certainly found in P’s own corpus, has been applied as the title of this P text. The same term is further associated with AF, i.e. the most famous instance of the interpolation of non-P ideas into the P corpus, which is itself associated with LAS-derived ideas. Rather than reading the title in terms of the content of ZSL itself, then, and reconstructing *Vijñānapariṇāma śāstra, it seems more consistent with this scenario to read the title in terms of the LAS provenance of the term zhuanshi, and to reconstruct *Pravṛttivijñāna śāstra. This refl ects a presumed intent of whoever applied the title to align the text with ideas derived from LAS and possibly AF.

Note, finally, that Daoji, who quotes ZSL by name, must have seen the text in whatever modifi ed form it assumed when this title was applied to it; conversely, we can conclude that the text must have acquired the title at least before his citations (in 633-637); see below p. 131. 128 A full translation of ZSL is the centre of the only book-length study of P’s works in English, PAUL; see 153-167. For apt criticisms of PAUL’s translation, see reviews by DE JONG and COX. 129 See L ÉVI (1925) and LÉVI (1932), 61-123. 130 L ÉVI (1925), 36. Sth agrees with P that the fi rst vijñānam refers specifi cally to ālayavijñāna; and that various particular sense-consciousnesses are produced in the process of pariṇāma the two commentaries differ in important respects.131 There is nothing in Sthiramati that sounds like the *amalavijñāna passage translated here.132

• Amalavijñāna occurs twice in one passage, at the end of a long comment on Triṃśikā 18. That verse proclaims:

“[This] consciousness (i.e. ālayavijñāna) is [possessed of/identical to] all seeds; due to the mutual infl uence [of consciousnesses one upon the other], [its] phenomenal transformation (pariṇāma) goes from one form to another, and thereby each fi gment of the imagination comes into being [in turn].” (though P specifi es a wider range of evolutes). 131 In particular, P’s translation and interpretation of tathā tathā/ yāty anonyavaśād is strange and diffi cult to account for. At this point, P’s text also adds comments on the distinction between subject and object of discrimination, and equivalences between these categories and parikalpitasvabhāva and paratantrasvabhāva, none of which are matched in Sth’s text. P’s emphasis on the disproof of external objects alone is also not paralleled in Sth. 132 That is to say, no caveat that the kind of vijñaptimātra here expounded is limited or provisional, no concern with a “pure” aspect of vijñaptimātra system or practice, and no mention of any concept that might answer to *amalavijñāna itself. For a French translation of Sth’s comments here, see LÉVI (1932), 107-108.

This verse explains how apparent phenomenal experience originates from the mind as a product of the interactions of karmic forces (seeds, bīja). The apparent objects of phenomenal experience are mere fi gments of false imagination (vikalpa), and the endlessly transforming stream of such experience is a series of transformations (pariṇāma) of consciousness itself.

Paramārtha’s ZSL then comments in detail on this verse, foot by foot. These comments basically treat the verse as showing the falsehood, i.e. the non-existence in the ultimate perspective of truth, of external phenomenal objects only. On this reading, the verse disproves the independent status of external things, but not of consciousness itself. In a move clearly related, in content if not in name, to the doctrine of two perspectives on vijñaptimātra we saw above in SBKL <2>, the text characterises this aspect of vijñaptimātra as the “impure aspect” (不淨品) ― a clear parallel to “representation only in the perspective of practice” (方便唯識). This leads the text directly to the notion of *amalavijñāna. It seems clear (though implicit) that this exposition is intended to present *amalavijñāna as the “pure aspect” of vijñaptimātra (both as practice and object of realisation), by contrast to the “impure aspect” just examined: “Question: If we do away with the phenomenal object, but allow consciousness (vijñāna) to remain 遣境在識, then we can call this principle ‘consciousness only’ (唯識, vijñaptimātra, ‘representation only’). But once both object and consciousness have been done away with, what ‘consciousness’ is there to be demonstrated/realised (成, /siddh) [in order that we can call the resulting state ‘consciousness only’]?” “Answer: In establishing ‘consciousness only’, it is in the fi nal analysis only for the purposes of argument (卒終為論) that one [proposes] merely doing away with the object and retaining the mind. In fact, the true purport [of ‘consciousness only’] is that one does away with the object because one wants to render mind empty (遣境為欲空心), and for this reason, the principle [of ‘consciousness only’] is only truly realised (是其義成) when both object and consciousness disappear at once. This simultaneous disappearance of both object and consciousness is precisely the perfected nature (pariniṣpannasvabhāva); and the perfected nature is precisely the *amalavijñāna.”140

As in SKBL<2> above, we here see *amalavijñāna related to a more perfect aspect or version of vijñaptimātra/weishi doctrine. Again, this “higher weishi” is superior, and ultimately true, because it not only disposes of external objects, but also of ordinary consciousness. The most signifi cant new aspect of the doctrine here is the identifi cation between the perfected nature (pariniṣpannasvabhāva) and *amalavijñāna; once more, this is associated with a non-dualism that obviates the subject-object dichotomy. 2.5 San wuxing lun

The San wuxing lun (三無性論 T1617, *Niḥsvabhāvatātraya-śāstra, “Treatise on the Threefold Absence of Essential Nature”, hereafter SWXL) corresponds to the Cheng wuxing Chapter (成無性品, “Proof of the Absence[s] of Essential Nature”) of the Xianyang shengjiao lun (顯揚聖教論, T1602, hereafter XYSJL). The term *amalavijñāna 140 問： 遣境在識、 乃可稱「唯識」 義、 既境識俱遣、 何識可成？答： 立唯識、 乃一往遣境留心、卒終為論。遣境為欲空心、是其正意。是故境識俱泯、是其義成。此境識俱泯、即是實性、實性即是阿摩羅識, T1587:31.62c15-20. My translation differs significantly from PAUL’s. The second instance of the term *amalavijñāna occurs in the phrase immediately following, 亦可卒終為論是摩羅識也. There are reasons to believe there is a textual problem here. (1) *Amalavijñāna is uniquely here represented by 摩羅識 alone, whereas in all other instances of the term it always has a preceding syllable for Skt. a- (阿摩羅識, 庵摩羅識, etc.); note, however, that Song, Yuan, Ming and Palace editions of the canon have a here. (2) As we noted in the preceding fn. 139, the odd phrase 卒終為論 is unique, in the entire canon, to the present passage. Its repetition at such close proximity may be a sign of a scribal error. (3) It is very diffi cult to extract any sense from this sentence. (PAUL’s translation here, “Additionally, we can say in the final analysis that this is Pure Consciousness,” is a guess at best, and does not acknowledge the strangeness of the Chinese syntax.) In support of this conjecture, we might note that the text seems also to be corrupt in other places; at the very beginning of the passage discussed here, for example, we see an apparently meaningless repetition of the phrase 及所餘七識種子 thus: 為諸法種子及所餘七識種子 及所餘七識種子, T1587:31.62b29-c02, where it seems clear the copyist’s eye has been drawn back to the fi rst instance of zhongzi and he has redundantly reduplicated its sequence. These possibly corrupt passages are the same in all editions of the text available to me (including the southern Qisha and two derivatives of the Kaibao canon, namely the Koryǒ and Jin versions). UI silently corrects the latter passage, UI 6, 416. In any case, whether or not the sentence is indeed corrupt, it is diffi cult enough to make sense of it that it adds nothing to our analysis of the overall meaning of the term *amalavijñāna. appears in two passages in this text.142 2.5.1 SWXL <1>

• Amalavijñāna occurs three times in a passage expounding the third of the seven tattvas, i.e. “the reality of representation only” (vijñaptitattva, vijñaptitathatā). As we saw above, *amalavijñāna is expounded in the same connection in SBKL <2>.143

“(3) The reality (Thusness) of representation only (識如如, *vijñapti [-mātra-]tattva).144 By this is meant that all conditioned phenomena (一切諸行, *sarvasaṃskāra) are nothing more than representation/consciousness (識, *vijñapti, *vijñāna). This ‘consciousness’ is called ‘reality’ (如如, tattva) in two senses: (i) it is a comprehension free from error (攝無倒,

• aviparyāsasaṃgraha?145); (ii) it is immutable (無變異, *avikāra).146

“(i) In saying it is ‘a comprehension free from error’, we mean that all dharmas, i.e. the twelve sense bases (入, āyatana) etc., are nothing more than representation/consciousness, and that beyond [this] deluded consciousness 亂識,147 there are no other dharmas. Thus, all dharmas are comprised by consciousness and nothing more.148 The discernment of this principle is called

• Āryadeśanāvikhyāpana, Āryaśāsanaprakaraṇa, Śāsanôdbhāvana etc. The chapter in question begins at 557b04. Parts of XYSJL and therefore of SWXL also correspond in some measure to portions of YBh, and I will refer to relevant parallels below. On the Xianyang shengjiao lun, see

SCHMITHAUSEN (1987) 2, 261-262 n. 99. 142 P AUL incorrectly asserts that it only appears in SWXL <2>; PAUL 142. 143 This passage is translated in its entirety in G IMELLO 317-319, and I have benefi ted greatly from consulting this translation and annotations in the process of making my own. A partial translation is also given in DEMIÉVILLE (1929), 41-42. GIMELLO also comments extensively on this passage, 320 ff. 144 The reader is referred to comments introducing SBKL<2> above p. 76, for the meaning of this concept and its place within larger rubrics. Note that the translation terminology varies: for 識如如 here SBKL has 唯識真實; see n. 118. 145 Following G IMELLO’s suggested Skt. and explanation, 318 n. 206. 146 See G IMELLO 318 n. 207, esp. the illuminating reference to MAV 3.11, NAGAO 41. 147 P uses 亂識 to translate both bhrānti “error, delusion” and abhūtaparikalpa “imagination of that which does not really exist” in his MAV. See GIMELLO 318 n. 208. This term is a fi ngerprint of P’s style; see n. 337. 148 Note that here, in being associated with luanshi, weishi 唯識 takes on a stronger sense of “consciousness only” than would be suggested by Skt. vijñaptimātra “representation only”. This same dynamic is noticeable at other places in P’s expositions of weishi also. ‘comprehension free from error’;149 and it is because it is free from error that it is ‘reality’150 (如如, ‘Thusness’ ). However, this ‘reality free from error’ is not yet the [higher] Thusness free of characteristics (無相如如, *alakṣaṇatattva). “(ii) In saying ‘it is immutable’, we show that this deluded consciousness 亂識 is in fact 即 manifested in consciousness of the pseudo-objects 似塵 識 proper to the imagined and interdependent [[[essential]] natures] (分別依他, *parikalpita[-svabhāva], *paratantra[-svabhāva]). Be- 149 I have differed from G IMELLO (who translates “comprises”, “comprisal” etc.) in translating saṃgraha as “comprehension”. I am attempting to convey what I take to be a play on words. On the one hand, vijñaptimātra is said to “comprehend” what is free from error in the sense that it includes it or encompasses it. On the other hand, the discernment with which vijñaptimātra is realised is itself a “comprehension” in the sense that it consists in arriving at an understanding of or insight into this basic fact of the reality of existence. It is this “comprehension” (understanding), I believe, that is “freed from error” (aviparyāsa). At least in Skt., the verbal root / grah admits of a similar polysemy to English “comprehend” in this regard, though it is a stretch to naturally interpret Chinese she in the cognitive sense. 150 G IMELLO 318 n. 209 refers us to the third of the ten tattvas in MAV Ch. 3. This third category is aviparyāsatattva, “reality free of error”, which is the reality that is discerned when we understand that existence is characterised by impermanence, suffering, absence of self, and impurity. (It may help to recall that prabhedatattva, or “reality as it is differentiated”, is the ninth of this tenfold list; and is the master rubric under whose sevenfold head the present vijñaptitattva is discussed as the third item). cause the imagined essential nature never exists, the interdependent essential nature also does not exist; and the inexistence of these two [[[essential]] natures] is *amalavijñāna (阿摩羅識). Uniquely, this consciousness alone is ‘immutable’, and for this reason it is [fi t to be] called ‘reality’ (‘Thusness’). “[Now,] the former ‘reality’ [i.e. (i) ‘Thusness/reality free from error’] only dispenses with the twelve sense bases (āyatana) [and] all [such] dharmas distinguished in the Lesser Vehicle. [This discernment or comprehension] merely frees the [[[understanding]] of the] twelve sense bases [etc.] from [[[predicative]]?] error 唯十二入非是顛倒. Here, however, the tenets of the Greater Vehicle demolish 破 the twelve sense bases [etc.], such that they are [seen to be] completely non-existent 並皆是無, and only the fi gments of deluded consciousness 亂識. Thus, [in this perspective,] the twelve sense bases [etc.] themselves are in fact errors 則為顛倒, but there is still no [[[predicative]]?] error 非顛倒 with regard to the deluded consciousness that is in fact all that exists. Thus it is called ‘reality’ (‘Thusness’).

“However, the substance 體 of this [deluded] consciousness is still mutable 猶變異. Next, [therefore,] in consideration of [its inexistence in terms of] imagined and interdependent natures, we do away with 遣 this consciousness. [In this perspective,] only the *amalavijñāna is free of [the] error [even of positing bare existence per se, 唯阿摩羅識是無顛倒]; [only *amalavijñāna] is immutable, and is therefore the true ‘reality’ (Thusness) 真如如. “Even in [interpreting] the former theory of consciousness only, we should adopt this interpretation of ‘consciousness’. Thus, we fi rst do away with external objects by means of [the posit that] there is only deluded consciousness; next, through disposing of the deluded consciousness by means of *amalavijñāna 阿摩羅識, [we realise that] there is ultimately only pure consciousness 究竟唯一淨識.”

As in ZSL above, *amalavijñāna is here identified with pariniṣpannasvabhāva (“perfected nature”). As in both ZSL and SBKL<2>, it is also associated with the superior interpretation of vijñaptimātra/weishi doctrine, which does away with ordinary consciousness as well as its objects, and so obviates the subject/object dualism. Once more, *amalavijñāna is identified with a kind of reality or Thusness (tattva, tathatā), as in SBKL above (especially SBKL<2>, where the link was direct). The text adds to our understanding of *amalavijñāna by further specifying a further sense in which the “reality” with which it is identifi ed is said to be real: it is free from change (avikāra; echoing JDZL<1>, “*amalavijñāna is permanent”). The text also aligns the “lesser” understanding of vijñaptimātra/weishi doctrine with “hīnayāna”, and asserts that only the superior understanding that perceives *amalavijñāna is worthy of the name “Mahāyāna”. 2.5.2 SWXL <2>

The last passage in which *amalavijñāna is mentioned is also found in SWXL. The root text is discussing the relationship between two groups of “characteristics” or “marks” (lakṣaṇa): (1) one group of fi ve characteristics, namely (i) signifi er (名言, *abhilāpa); (ii) signifi ed (所言, *abhilāpya); (iii) meaning (義, *artha?); (iv) attachment/hypostasisation (執著 *abhiniveśa); (v) non-attachment (非執著, *anabhiniveśa); (2) “three characteristics” corresponding to the three essential natures of Yogācāra, namely: (i) the characteristic of [all things as] imaginar y constructs (parikalpitalakṣaṇa); (ii) the characteristic of [all things as products of] mutual interdependence (paratantralakṣaṇa); (iii) the characteristic of [all things as they appear when brought to] perfection (pariniṣpannalakṣaṇa). 

The text then considers whether the fi ve encompass or comprise (攝, *saṃ/grah) the three, or vice versa. It answers that of the fi vefold rubric, categories (1) signifi er and (2) signifi ed are comprised in all three members of the threefold rubric; (3) “meaning” is comprised of the imagined characteristic (parikalpitalakṣaṇa) alone; (4) “attachment/hypostasisation” is comprised of the interdependent characteristic (paratantralakṣaṇa) alone; and (5) freedom from attachment is comprised of the perfected characteristic (pariniṣpannalakṣaṇa) alone.

In SWXL alone, a comment follows, bringing *amalavijñāna to bear: “The reason the fi rst two characteristics [of the fi vefold rubric] are comprised of all three characteristics [in the threefold rubric] is as follows. “(1) The characteristic [called] ‘the signifi er’ is the names of all things (dharma) and language. This signifi er is a product of consciousness. (i) Consciousness arises in the apparent guise of the signifi er, and for this reason it is of the nature of what is imagined (parikalpitasvabhāva). (ii) The subject [or ‘agent’] of [this] imagination, viz. consciousness 能分別識, is of the nature of what is produced by interdependence (paratantrasvabhāva). (iii) Since the signifi er constituting the imagined object 所分別名言 does not exist, the consciousness constituting the agent of imagination also does not exist, and [the discernment of] this [very fact] is the nature of [things as they are when brought to] perfection (pariniṣpannasvabhāva). For this reason, this fi rst [category] is comprised of all three natures 三性攝.

“(2) The second characteristic is also comprised of all three natures. The characteristic called ‘the signifi ed’ is the meaning/referent (*artha) intended by the signifi er 所目義, that is to say, all things 一切諸物, which are also products of consciousness. (i) Where only consciousness exists, it is of the nature of what is imagined (parikalpitasvabhāva) that there arise the apparent characteristics of things 似物相起. (ii) The agent of imagination, viz. consciousness, is of the nature of what is produced by interdependence (paratantrasvabhāva). (iii) That these two, as above, do not exist, is of the nature of [things as they appear when they have been brought to] perfection (pariniṣpannasvabhāva). “(3) The third characteristic [‘meaning’, i.e. reference] is only comprised of the nature of what is imagined (parikalpitasvabhāva) for the following reason. This ‘characteristic [of things] whereby word and meaning correspond to one another’ 名義相應相 refers to [the fact that] a word is designated for a thing. We make [the word] correspond to the thing, [so that,] by means of the word, it is possible to represent the thing 因名得顯物. However, this meaning of the word in fact does not exist, and because of the fact that it is so characterised by inexistence 無相義故, [we know that] it is only of the nature of what is imagined (parikalpitasvabhāva).

“(4) The fourth characteristic [‘attachment’, i.e. hypostasisation] is only comprised of the nature of [what is produced by] interdependence (paratantrasvabhāva) for the following reason. In this ‘characteristic of attachment to both word and meaning’, we distinguish the agent of this attachment, and thus [determine that] it is only of the nature of interdependence. Because [this category] does not explicitly [address] the object of attachment, it is not comprised of the nature of what is imagined. The preceding [category], however, only brings out the object of imagination, and not the agent of imagination; and thus it is not of the nature of interdependence. “(5) The reason the fi fth characteristic is comprised only of the nature of [things as they appear when they have been brought to] perfection is as follows. This state ‘characterised by freedom of attachment to both word and meaning’ is the *amalavijñāna, in which there is no distinction between object and wisdom/intuition 境智無差別阿摩羅識. The third and fourth [[[characteristics]], i.e. the apparent object and subject of deluded knowing comprised by imagined and interdependent natures] are in fact no different from this perfected nature; [the] only [difference is that] each of them establishes [a category that] manifests precisely one partial aspect [of the truth].”

In this passage, as previously in ZSL and SWXL<1>, *amalavijñāna is once more identified with pariniṣpannasvabhāva. This passage adds to our picture of *amalavijñāna a new dimension of non-dualism. Where previously the non-dualism was defi ned in negative terms, as the obviation of a false dualism between subject and object of delusory thinking, here, its content is defi ned positively as a relation of indistinction between perfected gnosis (wisdom, jñāna) and its object. In the emphasis in its exposition of the earlier members of the fi vefold rubric, the passage also exposes a certain dimension of the relationship between *amalavijñāna and language. Previously, in SWXL<1>, we were already told that *amalavijñāna is identical with “reality” in part because it is beyond error (aviparyāsa); given the frequent connection between viparyāsa and language, we could perhaps have inferred that this meant it was free of ordinary language also. Here, however, this aspect of *amalavijñāna is made explicit. All aspects of the operation of language and its referential function are associated with the imperfect parikalpita- and paratantrasvabhāva. *Amalavijñāna, by contrast, is associated exclusively with pariniṣpannasvabhāva, in large part because it is “free of attachment to word and meaning”. 3. Summary and analysis: Two doctrines of *amalavijñāna

Let us now review the picture of *amalavijñāna that emerges from the primary texts taken in the aggregate. For the purposes of this summary analysis, it will be convenient to divide the primary texts (excluding Abhidharmakośabhāṣya) into two groups: (1) Jueding zang lun (JDZL); (2) other texts. There are several reasons for this division.

First, the JDZL passages in which *amalavijñāna appears have counterparts in parallel versions of the text in Sanskrit, Tibetan and Xuanzang’s Chinese. By contrast, in SBKL, ZSL and SWXL, there is uniformly little or nothing in parallel texts that corresponds to passages expounding *amalavijñāna. This suggests strongly that in SBKL, ZSL and SWXL, we are dealing with sub-commentarial or “lecture” passages (whether by Paramārtha and his team at the point of translation, or by some later hand).

Second, JDZL is traditionally supposed to have been translated (or composed) earli-

名得顯物。此名義實無所有、無相義故、但是分別性。第四相但為依他性攝者、此執著名義二相、辨其能執故、但是依他性、不明所執、故非分別。前但出所分別、不出能分別、故非依他。第五相唯為真性所攝者、此不執著名義二相、即是境智無差別阿摩羅識故。第四第三 亦不離真實性、但其所立、正為偏顯一義耳, T1617:31.873c09-26. My translation of the last sentence is tentative. est of all the texts under consideration here, during the Liang (before 557). The facts here are complex and tenuous;168 but in the absence of fi rm evidence to the contrary, we

168 JDZL is fi rst mentioned in the catalogues in the Da Zhou kanding zhongjing mulu 大周刊定䱾 經目錄, composed by Mingquan 明佺 under Empress Wu (r. 690-705). Here, it is ascribed to P and the Liang in the Song, Ming and Yuan versions of the text, 梁天竺三藏真諦譯, T2153:55.407c15; in the Korean version, however, this line is missing, so that JDZL ends up appearing to be ascribed to Gautama Prajñāruci (active c. 516-543?), alongside his 迴諍論 Vigrahavyāvartiṇī T1631. JDZL is not more precisely dated until the next catalogue of the canon, Zhisheng’s (智昇, 669-740) Kaiyuan lu 開元釋教錄 (which dates to 730). Even then, Zhis heng only dates the text on the basis of the fact that it contains an interlinear note glossing a transliteration term by saying “in Liang, this is said . . .” 梁言; T2154:55.538b05; relevant JDZL passage at T1584:30.1018c09. So it is not until 120-160 years after P’s death that JDZL turns up, under that title, in the catalogues, and then the dating is based upon one very slender piece of evidence. Note that this means that the fi rst recorded references to the title Jueding zang lun are actually found outside catalogues; e.g. the reference in Wŏnch’uk (see below, p. 107-108, 145) predates Mingquan.

The situation is further complicated, however, by the fact that older catalogues, beginning with Fei Changfang’s (費長房, d.u.) somewhat unreliable Lidai sanbao ji 歷代三寶紀 in 597 (T2034:49.99a04), list a lost Shiqi di lun 十七地論 among P’s works, which seems to have been a partial translation of YBh (see further also n. 283). It is possible that our present JDZL is a surviving remnant of that text; for example, we will see below that there is at least one clear instance in which T2807 refers to contents included in JDZL by the title Shiqi di lun (see n. 283). If JDZL is a remnant of the Shiqi di lun, it may be relevant that that text too is supposed to date to early in P’s translation career, dating from the fourth year of the Taiqing era 太清四年 (approx. 550).

Judging from its title, we might well expect that Shiqi di lun (“Treatise on the Seventeen Stages”) should have been a translation of the Maulībhūmi portions alone of YBh. However, this seems not to have been the case. T2807 cites a portion of the text corresponding to the present JDZL, i.e. the Viniścayasaṃgrahaṇī (see once more n. 283). Bhikkhu Huimin has further pointed out that one of three Shiqi di lun passages referred to in P’s MSgBh is also from the Viniścayasaṃgrahaṇī: 十種法正行、如『十七地論』說, T1595:31.224b18-19 (Huimin has a19, in error), corresponding to T1579:30.706c22 (in juan 74 of XZ); Huimin [1994], 6 (I am grateful to ŌTAKE Susumu for pointing out this reference to me). In unpublished work, ŌTAKE has also found passages quoted from the Shiqi di lun that correspond to sections of the Maulībhūmi ranging from XZ juan 4 to juan 48. It therefore seems that whether or not it was a full translation or, as seems more likely, merely a set of excerpts, Shiqi di lun covered passages corresponding to a very wide range in the present YBh, including parts outside the Maulībhūmi. Thus, while it is possible that JDZL is a remnant of Shiqi di lun, Shiqi di lun cannot have been identical to our present extant JDZL. We should also note that when Huijun summarises the seventeen stages as laid out in the text (which he calls Shiqi jing 十七經), stages nine to fourteen comprise five sets of ten stages, totalling fifty stages, known to be unique to Chinese Buddhism, i.e. (1) ten “faiths” 十信, (2) ten “abodes” 十住, (3) ten should still follow the traditional bibliographers. If JDZL was thus the fi rst text in which *amalavijñāna was mentioned, and followed only at a distance of perhaps a decade by the other texts under consideration, the concept may have undergone considerable development between JDZL and the other texts. For this reason, we should be alert to possible differences in *amalavijñāna doctrine as it is expounded in JDZL, against the doctrine of the other texts.

Third, close study of the style of these texts suggests that SWXL, SBKL and ZSL may be part of a reasonably close sub-group in Paramārtha’s corpus, but differ in important respects from JDZL. These stylistic considerations reinforce the hints from the bibliographic tradition that the circumstances, and indeed collective authorship, of our texts may have differed in important respects.169

“practices” 十行, (4) ten “dedications of merit” (here broken up into two sets, for the hīnayāna and Mahāyāna) 捨[十]小乘廻向大乘, 大乘十迴向地, and (5) ten “grounds” 十地 (X784:46.569c11-13; ŌTAKE, personal communication). This would seem to make it unlikely that the text was simply a straight translation of any Indic version of YBh.

SWXL, by contrast, is attested as early as the Zhong jing mulu 䱾經目錄 by Fajing (法經, d.u.) of 594 (T2146:55.141b10) (much more reliable than the Lidai fabao ji); here it is already assigned to the Chen. It is also found in the Gu jin yijing tu ji 古今譯經圖紀 by Jingmai (靖邁, d.u.), dating to 664-665, where it is assigned even more fi rmly to a period in which P was supposed to have been in residence at Zhizhi si 制旨寺, which various sources place either “for some time” after 562 or between 563 and 567; T2151:55.365a01, 364c20. Given that ZSL is supposed originally to have been part of the same larger text as SWXL, viz. the Wuxiang lun 無相論, this information may apply to it also. Mention of an independent ZSL, however, is fi rst seen in the Da Zhou lu, where it is ascribed to P and assigned to the Chen dynasty, T2153:55.408a01; Zhisheng concurs, adding the detail that assignment to the Chen was confirmed by another catalogue, the Chulun ti danben 出論題單本, T2154:55.609a20-21. Finally, SBKL, like SWXL, is attested quite early, in the Zhong jing mulu 䱾經目錄 by Yancong (顏琮) of 602, where it is ascribed to the Chen, T2147:55.153c16; as with SXWL, Jingmai ascribes it to the Zhizhi si period. In the cases of what I am here calling the “other” texts (SWXL, ZSL and SBKL), then, our most concrete information is also late, dating to a century after P’s time.

Uncertain though it may be, this information suggests that JDZL may be ten years or more older than the other three texts, and thus date to a signifi cantly different period of P’s career, when his team was of very different composition, etc. 169 I will address the complex issue of authorship and style in a study currently in preparation. To give only one example, JDZL calls ālayavijñāna 阿羅耶識 ― 71 times! ― but the term is previously entirely unknown (and is only attested three times in one text even thereafter). JDZL’s frequent use of this term is thus unique in the P corpus in this regard.

Fourth, we fi nd that there are important differences between the *amalavijñāna doctrine of JDZL and that of the other three texts, but the other three texts present a relatively uniform version of the doctrine. 3.1 *Amalavijñāna in JDZL

We can summarise the most important points of the rich doctrine of *amalavijñāna in JDZL as follows. *Amalavijñāna corresponds closely to āśrayaparāvṛtti. The term corresponds to āśrayaparāvṛtti in textual parallels; it is also spoken of as parāvṛtti in JDZL itself, where context shows clearly that āśrayaparāvṛtti is meant; and it is free of “badness” (dauṣṭhulya). The text (especially in Paramārtha) equivocates paradoxically over the exact status of this āśrayaparāvṛtti ― in places it is said not itself to have, or be, a basis; but elsewhere it is said to be the basis for this or that (e.g. the path, lokôttaradharmas). The identity with āśrayaparāvṛtti is doubtless also connected to the interesting doctrine, propounded with special emphasis in JDZL<4>, that *amalavijñāna, identifi ed with a purifi ed vijñānaskandha, stands in a radically transformed relationship to the other four skandhas. The other four skandhas, when the object of grasping or attachment, are clearly spoken of as the “basis” (āśraya) for further rebirth, because clinging to them leads to appropriation of a new body (incarnation) after the end of one lifespan.

There seems to be a number of respects in which JDZL’s *amalavijñāna-cumāśrayaparāvṛtti warrants the epithet “pure”. It is the counteragent, or the result of the operation of counteragents, to ālayavijñāna and other features of the defi led state. It is associated with (a separate basis for) the “transcendent dharmas” (lokôttaradharmāḥ). It is also pure because it is diametrically opposed to the ordinary defi led state in many specific respects: most importantly, it is free of defilements (kleśa, 煩惱) and “outflows” (āsrava), and their causes; it is also free of all other qualities associated with the ordinary worldling (pṛthagjana). The purity of *amalavijñāna is also reflected in the fact that in it, the sensations (vedanā) are rendered pure. Its purity is also refl ected in its association with Thusness. *Amalavijñāna is realised by intensive cultivation of the wisdom that knows Thusness (tathatā), and the very notion of “taintlessness” (amala, nirmala) may well be intended to recall expositions of “purifi ed Thusness” in RGV, other parts of YBh, etc.

• Amalavijñāna is also permanent (nitya), and this is emphasised by Paramārtha’s assertion, against parallel texts, that amalavijñāna-cum-āśrayaparāvṛtti does not cease of its own accord.

Apart from the naming of the liberated state (āśrayaparāvṛtti etc.) as *amalavijñāna, Paramārtha adds little to ideas already found in YBh. In effect, he has just given a clear name to an already-present confi guration of ideas about the implications of liberation for consciousness. The name, moreover, is not inappropriate. The parallel passages also discuss liberation as a liberation of consciousness from defi lements, in a manner that amply justifi es the epithet “pure”; and the phrase “pure consciousness”, in some form, even appears in one place in parallel versions of the text.

It is impossible to miss, throughout these passages, the ring of the Yogācāra “Mahāyāna Abhidharma” ― much of the talk is of seeds (bīja), defi lements (kleśa), “outfl ows” (āsrava), aggregates (skandha), gradualist models of paths of practice and realisation (including śaikṣa and aśaikṣa), etc.

Some of this doctrine echoes the AKBh amalavijñāna passage, not only in the general “Abhidharma” language, but in specifi cs. In both contexts, at issue is a consciousness pure specifi cally in that it is free of “outfl ows” (āsrava), and of defi lements under other names. This purity matters so much because it enables us to attain freedom from eventual rebirth, or more precisely (especially in AKBh) freedom from the “latent tendencies” (anuśaya) that constitute the most subtle level of grasping after the bases of future rebirth. If Paramārtha picked up the term amalavijñāna from AKBh 5.29 ff. and used it to name the doctrine of pure consciousness elaborated in these YBh passages, it was thus an artful move.

We see echoes here of an old doctrine of consciousness as the subject of transmigration and liberation. These echoes have not been noted by previous *amalavijñāna scholarship. In brief, this old doctrine is as follows.

In some texts, such as the Mahānidāna sutta (DN 15), viññāṇa is presented as the sine qua non for embryonic development; as “descending” into the mother’s womb, “leaving” in cases of miscarriage, etc. Viññāṇa is also presented as a “surviving” factor, called in the Majjhima nikāya the “consciousness that evolves [into the next life]” (saṃvattanikaviññāṇa). In both these connections, the understanding of viññāṇa is closely related to the place it assumed in the standard twelvefold chain of dependent origination (paticcasamuppāda).173 Another related idea is that consciousness arising from moment-to-moment sense experience is the “origin of the world” (lokassa samudayaṃ) with all its suffering, and that the suffering world ceases when consciousness ceases. Wijesekera has further shown that viññāṇa so understood is connected to the notion of the gandharva/gandhabba, as the “being that enters into the womb” on conception, a “being of the intermediate state” (antarabhavasattvam); this notion in turn is linked with manas, mind or “soul”, and, ultimately, with an ancient and sprawling network of various mythemes reaching back to the Vedas and apparently beyond. The resonances between these ideas and JDZL are strengthened by other key details. Even in the earliest texts, as in JDZL, viññāṇa is continually drawn back into rebirth precisely because it is still associated with anusaya (anuśaya). Further, viññāṇa is spoken of in the Sampasādanīya sutta (DN 28) as a “stream” (viññāṇasotaṃ), in a manner that clearly recalls the “continuum” or “fl ow” (saṃtāna) at issue in JDZL.

In the early texts, this same consciousness is sometimes understood as continuing 173 In this formula, viññāṇa is pivotally placed as the key link in the process leading to rebirth. On the emergence of the twelvefold model, problems of internal consistency in it, and its possible basis in more than one earlier model, see the still seminal LA VALLÉE POUSSIN (1913). even into rarefi ed meditative states as far as nevasaññānâsaññâyatana, and there are indications that it, or perhaps “mind” construed more generally, was in some sources regarded as the quasi-“subject” of the process and state of liberation itself. Such a consciousness is said to be “unestablished” (apatiṭṭhita-viññāṇa): a representative statement says that when lust (rāga) has been abandoned with regard to each of the fi ve khandhas in turn, “there is no basis (ārammaṇam) for the establishing of consciousness (patiṭṭhā viññāṇassa na hoti);” such a consciousness is liberated (vimuttaṃ), steady (ṭhitaṃ) and satisfi ed (santusitaṃ), and has attained nirvāṇa (parinibbāyati). It seems clear that in such a state, consciousness is still undergoing experiences of some kind, which are free of suffering, and that it is such experience that comprises liberation. (However, we must also note other passages that depict liberation as a “cessation” of viññāṇa altogether. ) It is also noteworthy that this state is already spoken of in terms of there being no “basis” (ārammaṇam = Skt. ālambanam) for consciousness (although the term is not āśraya, JDZL clearly echoes these ideas). In “Volition”, SN 40.10.3, the attainment of this “unestablished” state without a basis is further clearly linked to the absence of “latent tendencies” (anusaya = anuśaya).

These and related ideas find resonances in the JDZL doctrine of the ordinary vijñānaskandha, attached to the other four skandhas as “base” and therefore undergoing repeated rebirth and suffering; and of *amalavijñāna as a metamorphosed or purifi ed transmutation of this vijñāna, which is freed from suffering, the subject of a pure kind of experience, and eternal.

Modern scholars have often taken *amalavijñāna as a kind of bridgehead, intended to enable annexation of tathāgatagarbha doctrine into Yogācāra. Perhaps the doctrine did lend itself to that use, and perhaps Paramārtha was at pains, elsewhere in his work, to bring about such a rapprochement. However, very little in JDZL calls to mind tathāgatagarbha doctrine ― only, faintly, that *amalavijñāna-cum-āśrayaparāvṛtti is associated with “pure Thusness”; that it is permanent (nitya); and that it somehow pre-exists the liberatory state (thus resembling tathāgatagarbha as a ground for potential realisation). Moreover, this is all also present in parallel texts, and therefore was presumably in Paramārtha’s source. The connection between tathāgatagarbha and *amalavijñāna in JDZL may thus be weaker than scholars have thought. 3.2 *Amalavijñāna in SWXL, ZSL and SBKL

In SXWL, ZSL and SBKL, *amalavijñāna is identifi ed with a “higher” or complete understanding of vijñaptimātra/weishi (“consciousness/representation only”), in which the practitioner realises the unreality not only of objects, but also of ordinary defiled consciousness itself (of ālayavijñāna). A state is thus attained that transcends the usual epistemological dualism of subject and object. Further, in explicit connection with this “higher weishi”, *amalavijñāna is identifi ed with the pariniṣpannasvabhāva (“perfected nature”) (ZSL, SWXL). This status of *amalavijñāna as the “perfected nature” hinges on the obviation of the delusory dualism of subject and object that pertains in ordinary consciousness. It is also said to be also expressed in the non-dualism of perfected gnosis and its object in the liberated state (SWXL). *Amalavijñāna is further associated with (the pure aspect of) Thusness or reality, particularly in the context of the exposition of the seven kinds of tattva (SBKL, SWXL).

The concept of *amalavijñāna in these texts is thus relatively consistent and uniform. However, some claims are unique to a single text. Most importantly, only in SBKL is *amalavijñāna identifi ed as “aboriginally luminous mind” (prakṛtiprabhāsvaracitta), which is tainted by adventitious defi lements (āgantukakleśa). This aboriginally luminous mind is further identifi ed with the fundamentally pure aspect of Thusness (tathatā), and also of the related domains of emptiness (śūnyatā) and the dharmadhātu. SWXL<1> alone claims that the aspect of Thusness constituted by *amalavijñāna is real (or “thus”) because it is immutable, i.e. not subject to metamorphosis (avikāra). SWXL<2> alone claims that it is real, also, because it is beyond ordinary language. On the whole, however, divergences between the texts are minor, and the coherence in their doctrine of *amalavijñāna is more striking.

There is still relatively little, in this version of *amalavijñāna doctrine, that recalls tathāgatagarbha doctrine. The strongest echoes of tathāgatagarbha ideas are found in SBKL, where *amalavijñāna is identifi ed with prakṛtiprabhāsvaracitta as obscured by “adventitious dust 塵”.

The relatively uniform *amalavijñāna doctrine of these three texts is strikingly different from that of JDZL. JDZL does not associate *amalavijñāna with vijñaptimātra/ weishi doctrine in any form. It unsurprisingly, therefore, never breathes a word of transcendence of the subject-object dualism. Neither does it identify *amalavijñāna with the “perfected nature” (pariniṣpannasvabhāva). Despite its clear concern with the purity of consciousness and mind, it also does not link *amalavijñāna and prakṛtiprabhāsvaracitta. Nor does it link *amalavijñāna to emptiness or the dharmadhātu; nor claim that it is beyond ordinary language.

By contrast, these three texts do not associate *amalavijñāna with āśrayaparāvṛtti as JDZL does. They do not treat *amalavijñāna as the vijñānaskandha, nor consider its relationship to other skandhas. They do not associate *amalavijñāna with the problematic of rebirth or its escape. There is, correspondingly, no discussion of defi lements or “outfl ows” (āsrava). *Amalavijñāna is never said to be the “counteragent” (pratipakṣa) to anything. These texts never broach the question of the relationship between *amalavijñāna and the path, or various kinds of dharmas (e.g. lokôttaradharmas, pṛthagjanadharmas), or the stages of śaikṣa or aśaikṣa. They never speak of the purifi cation of vedanā. Nothing in these texts echoes the AKBh passage in which the term amalavijñāna is attested in Sanskrit. There is also nothing, in these texts, of the echoes with the old doctrine of consciousness as the subject of transmigration and liberation.

More generally, in JDZL we observed a close entanglement of *amalavijñāna doctrine with the Yogācāra “Mahāyāna Abhidharma”. Here we see, rather, attempts to connect *amalavijñāna more directly with core elements of innovative doctrines more particular to Yogācāra itself ― that there is “only consciousness” (vijñaptimātra); three natures (trisvabhāva); theory of language; the non-dualist nature of true gnosis; and the relationship between that gnosis and the reality or Thusness it knows. In fact, SWXL<1> even seems to scorn an Abhidharmic understanding of reality (twelve āyatana etc.) as “hīnayāna”.

In contrast to these stark differences, there are only minimal areas of overlap between the presentation of *amalavijñāna in JDZL and the other three texts. In both sets of material, *amalavijñāna is associated in some regard with Thusness; and in both, the permanence of *amalavijñāna is stressed.

What, then, are we to make of these differences? We cannot be sure what they mean, but some possibilities suggest themselves. The traditional bibliographies assert that JDZL is much earlier than other texts featuring *amalavijñāna. There also seems to be a closer fi t between its *amalavijñāna and amalavijñāna in AKBh. These are also the only passages in which *amalavijñāna features in a “translation” rather than a “lecture” or “sub-commentarial” context. JDZL<1> is also the only passage that expounds *amalavijñāna at length, whereas it is elsewhere often merely mentioned. Finally, stylistic differences and other structural considerations make it possible that *amalavijñāna passages in the other texts may have been added by a later hand. JDZL is thus far more likely than the other texts to preserve for us the fi rst known exposition of *amalavijñāna.

It is possible that the differences between the “Yogācāra Abhidharma” of JDZL and the “core Yogācāra” of the other three texts is merely a function of the different topics at issue. The two versions of *amalavijñāna doctrine, while different, do not directly contradict one another, and so may be mutually consistent developments, in different directions, of the same doctrine by the same hand. It is also possible that both sets of materials are equally the work of Paramārtha(’s group), but that the doctrine was further developed between JDZL and the other texts, and that the composition of the authorial team shifted in the interim as well. SWXL, SBKL and ZSL may therefore show us a later version of *amalavijñāna. Alternatively, *amalavijñāna passages may have been interpolated later into SWXL, ZSL and SBKL, perhaps as subcommentary. Such a later hand might still be quite closely related to Paramārtha’s group.

Whatever the reasons, within Paramārtha’s corpus we thus fi nd not one but two relatively distinct doctrines of *amalavijñāna. We now turn to the question of how accurately either of these versions of the doctrine was communicated to the later tradition. 4. *Amalavijñāna in later sources

There are relatively few sources for Paramārtha’s doctrine of *amalavijñāna. It is natural, then, to turn for further information to the testimony of later scholastics. Such scholastics report aspects of *amalavijñāna doctrine never found in Paramārtha’s extant corpus, and this may be because they had access to additional facts. Such authors may have seen texts since lost; sometimes they refer to such texts (or at least claim to). They were also closer in time and space to Paramārtha, and may have learnt things by hearsay that were never written down.

However, we cannot always be sure that later reports are accurate. These authors may have quoted from memory, or at second-hand. They may not have had access to Paramārtha’s texts, but could be relying upon hearsay. Ideas might have been ascribed to Paramārtha to lend them authority, even if actually elaborated by someone else ― much as was the case with the entire Awakening of Faith. Some portrayals of Paramārtha’s ideas may have been coloured by sectarian polemics.

For these reasons, we must determine the extent to which later authors concur with or diverge from Paramārtha, as a way of judging their reliability. To this end, I will here summarise reports about *amalavijñāna to 800 C.E. under three heads: (1) what later authors report that we do not find in Paramārtha; (2) agreement between Paramārtha and later authorities; and (3) what we fi nd in Paramārtha that is not reported by later authorities. 4.1 What later sources say that Paramārtha does not

In later reports of *amalavijñāna, we find much that is never found in Paramārtha’s extant corpus. First, later authors use various terms never found in Paramārtha; second, the specifi c content they attribute to the doctrine of *amalavijñāna is extremely various. In what follows, I consider this material in approximate chronological order. 4.1.1 A proliferation of various terms

Turning fi rst to terms, we fi nd that later sources frequently refer to *amalavijñāna by various names that Paramārtha’s texts do not use.

I have already mentioned above the fact that later sources sometimes use the term wugoushi 無垢識 to refer to *amalavijñāna.191 Wugoushi is identified with “ninth consciousness”, and therefore with *amalavijñāna, already in Huijun (慧均, d.u., fl. 574-580s?) and in Daoji (道基, 577-637); by Kuiji (窺基, 632-682); by Wŏnch’uk (圓測, 613-696); by Wŏnhyo (元曉, 617-686); by Taehyŏn (大賢, 太賢, d.u., active c. 742-765); by Chengguan (澄觀, 738-839); and by Zongmi (宗密, 780-841). Tunnyun (遁倫, d.u., Silla monk of the eighth century) also knows this term.

We also fi nd *amalavijñāna called jingshi 淨識, e.g. by Zhiyi (智顗, 538-597), Huijun and Daoji.

We also fi nd that there are many variants of the transcription term that never appear in Paramārtha: (1) *Amalavijñāna is called anmoluoshi 菴摩羅[[[識]]] by Zhiyi; by 191 The term wugoushi is already mentioned in connection with Yogācāra doctrine by Jingying Huiyuan 淨影慧遠 (523-592), but he thinks that it is a term for ālayavijñāna (T1851:44.524c25-26). This is similar to the meaning the term has in the Cheng weishi lun, T1585:31.13c19-27); Huizhao (惠沼, d. 714) understands the term in a similar manner, as one of the eighteen names of ālayavijñāna mentioned in the Cheng weishi lun (成唯識論了義燈 T1832:43.729c03). Kuiji;205 by Zhanran (湛然, 711-782), perhaps following Kuiji;206 and once by Tunnyun, citing Kuiji.207 (2) In the Vajrasamādhi sūtra T273 alone (prior to the Song), *amalavijñāna is referred to by the variant anmoluoshi 庵摩羅[[[識]]].208 (3) It is called anmoluoshi 唵摩羅[[[識]]] by Wŏnhyo;209 in Amoghavajra’s (不空金剛, 705-774) T1177A; by the anonymous (perhaps late eighth century) Shi moheyan lun T1668; by Amoghavajra’s disciple Huilin (慧琳, d. 820); and in a sub-commentary on T1668. (4) The transcription amoluoshi 阿末羅[[[識]]] was used by Kuiji; by Wŏnch’uk (圓測, 613-696); by Tunnyun, citing Kuiji; by Taehyŏn citing Kuiji; and by Tankuang (曇曠, c. 700-788). (5) *Amalavijñāna is called anmoluoshi 菴末羅識 by Kuiji; and, citing him, by Tunnyun. (6) The transcription amoluoshi 阿磨羅識 is used by Li Shizheng (李師政, d.u., fl . 627-649);221 and by Tankuang. 205 T1829:43.179a01-05.

206 T1912:46.221c02-12; the contents of Zhanran’s comment also support the suspicion that he was getting his information from Kuiji. 207 T1828:42.605b19-23. 208 T273:9.368c28-29 (twice); see n. 209, 211. 209 T1730:34.978a01-28 (several times including quotes from the root text), 980b10, 980c07, c10-11, 981a24, 981a26; note that these instances include quotes from the root text, in which Wŏnhyo follows his own orthography rather than that of (our extant) VSS.

Of course, orthography was sometimes a matter of some indifference. For example, when Wŏnhyo quotes the Vajrasamādhi sūtra in his Vajrasamādhi sūtra lun, he uses a variant orthography from the one we fi nd in the extant text of the VSS itself.223 However, the majority of sources do still use the standard Paramārthian transcription amoluo 阿摩羅. Where transcriptions in later authors depart from this standard, variant transcriptions are also retained by later authors who quote those authors in turn (e.g. Tunnyun or Taehyŏn’s quoting Kuiji). Thus, it seems authors did stick with a given transcription, and departures from the usual transcription may indicate, for example, that an author had heard the term but not read it; or that his knowledge of it came from a source other than the texts of Paramārtha himself. 4.1.2 “Ninth consciousness/nine consciousnesses”

Perhaps the most signifi cant variant term we encounter for *amalavijñāna is “ninth consciousness” (jiushi 九識). Numerous later sources frequently say that Paramārtha expounded *amalavijñāna as such a “ninth consciousness” over and above the “standard” eightfold model of consciousness of normative Yogācāra. However, Paramārtha’s extant texts never say that *amalavijñāna is a “ninth consciousness”.224 4.1.2.1 Reference to *amalavijñāna as “ninth consciousness”

The claim that *amalavijñāna is a ninth consciousness is found as early as Zhiyi,225 Jingying Huiyuan (淨影慧遠, 523-592),226 Huijun,227 and Jizang (吉藏, 549-623);228 in the anonymous Dunhuang text She dasheng lun zhang 攝大乘論章 T2807;229 in the She

the āmalaka (myrobalan) tree, fruit etc. P himself, by contrast, uses amoluo 阿摩羅 even for āmalaka: e.g. T669:16.468b15, 469a23-24. 223 See n. 209. 224 We must be careful to distinguish cases where P uses the phrase jiushi, but to refer to the ninth of the eleven vijñapti, e.g. T1595:31.181c14-15. See p. 106, and n. 245, 79. 225 T1716:33.742b01-10, 744b22; T1783:39.4a12-13; T1778:38.686a06-11; X356:20.42b08-09. 226 T1843:44.176a08-13, 179a20-29, 179c13-17. 227 See below n. 272. 228 T1824:42.104c08-09. 229 T2807:85.1016c10-11. The date of the group of Dunhuang texts including T2807 is uncertain, but the best estimates of modern scholars, including ODA Akihiro, KATSUMATA Shunkyō, Shengkai and Ching KENG, tend to place each sometime between 590-640. For a recent summary of research on this subject see Shengkai, Shelun xuepai yanjiu 1, 47-59; see also, especially on T2805, the forthcoming Harvard PhD dissertation of Ching KENG. I am grateful to KENG for pointing me to this information (personal communication, September 2008). das heng yi zhang 攝大乘義章 T2809 (probably by Daoji);230 in other writings of Daoji;231 in Zhiyan (智儼, 602-668);232 in Li Tongxuan (李通玄, 635-730);233 in Kuiji;234 in the Vajrasamādhi sūtra; in Wŏnhyo (based on the VSS); in Zhizhou (智周, 668-723); in Tankuang; and in Zhanran. 4.1.2.2 A supposed special text on “ninth consciousness”

In addition to the reports of the authors just cited, there is a complicated body of further evidence that similar ideas may have been contained in a lost text attributed to Paramārtha, apparently primarily on this subject. The text in question is given a wide variety of titles containing the phrase “nine consciousnesses/ninth consciousness” (jiushi 九識). 230 T2809:85.1036b28. On Daoji’s probable authorship of this text, see n. 382. 231 See below §4.1.3.7. 232 T1869:45.522c24-25; T1870:45.543a20. 233 T1739:36.722b15; 722c22; 723a06-07; 723b09-11; 736a20-b03; 741b29-c01. Li Tongxuan is notable because, in discussing *amalavijñāna, he never uses any transcription term (he rather calls it ādānavijñāna!); his understanding of the doctrine is also highly peculiar; see n. 482. 234 T1829:43.179a05; T1830:43.239a11-19; T1861:45.282c24-25. 1) ZSL contains the phrase “as it says in chapter on the doctrine of ‘Nine Consciousnesses’ 如九識義品說”.242 However, this is not conclusive proof that Paramārtha himself knew of such a text, let alone that he wrote one. ZSL apparently contains a sub-commentarial layer, perhaps by a later hand,243 and this comment may belong to that layer.244 Further, YŪKI Reimon has argued the possibility that this comment refers to a text like the Xianshi lun (XSL), which expounds not a ninefold system of “consciousnesses” (識, vijñāna) but the first nine out of eleven “categories” or “ideas” (識, vijñapti) (the xianshi = khyātivijñāna of XSL’s title).245 Further, the comment falls in the course of a discussion which is in fact on ālayavijñāna, i.e. eighth consciousness.246 The ZSL comment may thus refer to a text entirely separate from the issue of *amalavijñāna as “ninth consciousness”. 2) The preface to Awakening of Faith (AF) claims that a Jiushi yi zhang 九識義章 was translated by Paramārtha, and gives quite specifi c circumstances and dates.247 However, this is a preface to a text whose own attribution to Paramārtha is generally regarded as spurious; and the preface is also considered inauthentic.248 3) Bibliographic sources report that Paramārtha was the author of a text called the Jiushi yiji 九識義記. However, the fi rst report of this text is in the often unreliable Lidai sanbao ji 歷代三寶紀, among a crop of texts dated to Paramārtha’s period of activity under the Liang dynasty, to which Fei Changfang (費長房, d.u., fl . under the Sui) attaches

passage; but I stress that all these translations are provisional. 242 T1587:31.62a04. 243 See F UKAURA 1, 315 ff. 244 ZSL is not mentioned in the catalogues until the Da Zhou kanding mulu, composed under Empress Wu (r. 690-705). However, we can be certain that the text already existed (though not precisely by this name), and that it already contained this reference to a “chapter on the doctrine of nine consciousnesses”, from the fact that it is quoted by Daoji, writing at the latest in 637; see n. 375 below. 245 Y ŪKI, 41-42. See also n. 224 above. On the set of categories at issue here, see n. 79. 246 Y ŪKI 42. 247 T1666:32.575a28-b01. 248 See D EMIÉVILLE (1929) 11-15. CHEN Yinque, however, has pointed out that some historical details in this preface could only have be known by someone very close to the original context in which P and his group worked, so that we cannot dismiss all of its contents out of hand. Prof. FUNAYAMA has rightly stressed that we must take this into consideration in weighing the testimony of the preface about the Jiushi yi zhang (personal communication, October 2008). a suspicious precision about dates and places of translation. The Da Tang neidian lu repeats this report verbatim; other Buddhist bibliographers through to Zhisheng (730) do not pick it up. 4) Gyōnen says that Daoji’s Shelun zhang is a commentary on the Jiushi zhang 九識 章. From context, it is clear that Gyōnen probably understands the Jiushi zhang to be the source of doctrines Daoji has just attributed to Paramārtha in a passage Gyōnen has quoted. Gyōnen seems to have had Daoji’s Shelun zhang before him as he wrote (he quotes it extensively). If Gyōnen has this information direct from Daoji, this is probably our strongest piece of evidence that a Jiushi zhang existed and was already attributed to Paramārtha in the Shelun school of Daoji’s time. However, Gyōnen does not say that the Jiushi zhang was in fact by Paramārtha, and it is possible that Daoji was reading a Jiushi zhang by another author (for instance, his teacher Jingsong; see below) which described Paramārtha’s doctrines at second hand. It is also possible that Gyōnen knows of a supposed erstwhile Jiushi zhang by Paramārtha, and has merely inferred that Daoji is commenting upon it. 5) Wŏnhyo also refers to what seems most likely to be a similar text (a zhang), but without specifying its title. 6) Wŏnch’uk cites a Jiushi zhang 九識章, and even says specifi cally, in his commentary on the Saṃdhinirmocana sūtra, that the text was quoting a “‘Ninth Consciousness’ Chapter” of JDZL. This is the fi rst time we see the claim that JDZL contains such a special chapter, here found in tandem with the notion of a freestanding text on the same topic. The most impressive thing about Wŏnch’uk’s evidence is that he cites the Jiushi zhang as a source for ideas that are not directly connected to the doctrine of *amalavijñāna or ninth consciousness. Here, it is diffi cult to imagine any motivation for Wŏnch’uk (or any intermediate source, if the citation is indirect ) to ascribe the passage to the Jiushi zhang, except that such a text indeed existed and contained it. This evidence cannot reassure us that the Jiushi zhang these scholars cite was correctly ascribed to Paramārtha, but it strongly suggests that some text of that name certainly did exist.

7) In a Vinaya text by Dajue 大覺 (712), we also see the claim that JDZL contains a “Ninth Consciousness Chapter” 九識品, referring to *amalavijñāna. 8) Tunnyun reports that according to Huijing 惠景, Paramārtha “established the doctrine of nine consciousnesses 九識義 on the basis of a citation from the ‘Ninth Consciousness Chapter’ 九識品 of JDZL.” Tunnyun is sceptical, and notes that there “never was” any such chapter in the corresponding part of YBh. Tunnyun notably tells us explicitly that he only has this information by hearsay. 9) Tankuang reports that the theory of ninth consciousness is found in “the ‘Treatise on Nine Consciousnesses’ (Jiushi lun 九識論), translated by Paramārtha”. 10) In a statement either paraphrasing Dajue or quoting a third common source, Chengguan 澄觀 (738-839) also refers to a “Ninth Consciousness Chapter” of JDZL 決定 藏論九識品. 11) A Jiushi lun 九識論 in two juan is attributed to Paramārtha in a catalogue of manuscripts from Japan dating to 753. However, the entry in question is included in a list of manuscripts “yet to be copied” (misha 未写), so there is no guarantee that the compiler of the catalogue actually sighted the text in question.

We should also note the ambiguous case of Huijun, who refers at one point to doctrines expounded in MSg and a “jiushi yi” 九□ 義并『攝論』. Jiushi yi here could simply mean “[in expounding] the tenet of ninth consciousness”, but given that later reports held that Paramārtha wrote a text entitled the Jiushi yiji 九識義章 etc. (AF preface, Fei Changfang), we must also recognise the possibility that this is an abbreviated reference to the title of a text.

Against these sources, we must weigh another set of references, which sometimes purports to trace the idea of “ninth consciousness” or “nine consciousnesses” to similar source texts, but without referring to a special text or chapter with “ninth consciousness/nine consciousnesses” in the title. Such authors are largely early.

1) Huijun (conceivably our earliest witness ) says: “What is the practice in seventeen stages? It is as laid out in the Sūtra of Seventeen [Stages] 十七經, which is cited by the Trepiṭaka Paramārtha to prove the tenet that there is a ninth consciousness 有九矧. His treatise 彼論 says, “nine kinds of mind 九品心”, and thus there exists a ninth consciousness 有第九矧. However, that sūtra [i.e. Shiqi jing] has not been translated here [in China 此間 ], and [[[Paramārtha’s]] claim] is thus diffi cult to believe.”

It is key that Huijun does not say the “Shiqi jing” contained a chapter 品 specially on “ninth consciousness” (i.e. a Jiushi pin 九識品). Rather, he says that the text speaks of jiu pin xin 九品心, most naturally read “nine kinds/categories of mind”. This claim fi ts better with our extant JDZL (if we grant the identifi cation between *amalavijñāna and “ninth consciousness”) than any claim that any partial YBh translation by Paramārtha contained an entire special section devoted to “ninth consciousness”. This is also the earliest reference in our extant record to the exposition of ninefold mind in JDZL. We might thus suspect that the idea of a “Ninth Consciousness Chapter” arose later by a reversal of Huijun’s word order (九品心 → 九心品), combined with the idea that Paramārtha wrote a special text on this topic.

2) The anonymous Shelun School Dunhuang text, She dasheng lun shu 攝大乘論疏 T2805, mentions a passage in which JDZL “expounds ninth consciousness” 『決定藏論』釋九識中. This passage is of particular interest because T2805 is here not itself expounding *amalavijñāna or “ninth consciousness”, so this characterisation of the text is clearly not motivated by T2805’s doctrinal agenda. The quote following this introduction is not a verbatim citation, but perhaps rather a paraphrase. However, the author clearly knew Paramārtha’s text. The cited passage falls just before JDZL<1>.

This passage, comments by Huijun, comments in T2807, and comments by Daoji (who cites parts of XWL)280 are the only instances where a later reference to “ninth consciousness” can clearly be shown to have in mind an extant Paramārtha text. T2805 is also the earliest text to say that “ninth consciousness” derives from JDZL, and the only text that actually recognisably cites JDZL in the process. T2805 also shows no special interest in “ninth consciousness” or *amalavijñāna, but rather merely mentions it here, in passing. It is thus early, accurate, and has no special interest in presenting any version of “ninth consciousness” doctrine. If T2805 also does not say that JDZL contains a special “chapter” on “ninth consciousness”, we may be glimpsing another intermediate stage in the elaboration of a legend of a special chapter on ninth consciousness in JDZL. 3) Another anonymous Shelun School Dunhuang text, the She dasheng lun zhang 攝大乘論章 T2807, says “it is said abroad” 外國傳云 that the “‘Bodhisattva Chapter’ of the ‘Treatise on Seventeen Stages’ [Shiqi di lun 十七地論菩薩品] goes into detail to distinguish an *amalavijñāna, and counts it as a ninth consciousness 廣辨阿摩羅識以為九 識.”282 Now, Shiqi di lun is the title of a lost partial translation of YBh by Paramārtha, and the present JDZL may be a remnant of it.283 The present JDZL does not contain a “Bodhisattva Chapter”;284 but the passage that concentrates most on *amalavijñāna, JDZL<1>, does open with an indication that it is discussing the entry of the bodhisattvas into the “stage of non-regression” (不退地, avaivartikabhūmi). Thus, the title “Bodhisattva Chapter” may be no less appropriate than “Ninth Consciousness Chapter”.285

T2807 thus traces *amalavijñāna to the exposition in JDZL, but refers to it by another alternate chapter name, also unattested in our extant text. This suggests that: (1) reference to texts by title was loose; (2) the tradition that JDZL contained a “Ninth Consciousness Chapter” was either not yet established, or at least not yet universal.

4) The wording of Huijun’s report ((1) above) is echoed by Daoji 道基. Daoji says that Paramārtha cited the “defi nitive exposition (? 決定說) of nine kinds of mind 九品心 in the Shiqi di lun” to prove a theory of ninth consciousness/nine consciousnesses.286 However, Daoji does not seem to have very concrete information about his sources. He claims, for example, that Paramārtha quotes LAS in expounding his theory, which is not supported by any extant evidence.287 Further, Daoji shows himself wary of the Shiqi di

who said that Shiqi di lun corroborated the doctrine of *amalavijñāna found in JDZL? Or does it simply indicate that the author presumed it was a “foreign tradition 外國傳” because the text is presented as a translation? Does the use of the title Shiqi di lun mean that the author had actually seen a text circulating under that title that expounded *amalavijñāna ― and as a ninth consciousness, to boot ― or that he was just attempting to put two and two together from various pieces of hearsay? 282 T2807:85.1016c21-23. 283 See n. 168. 284 The *amalavijñāna passages appear in a chapter entitled “The Stage/Ground of the Mind” 心地品. 285 諸菩薩入不退地 T1584:30.1020b05 286 又引『十七地論』決定說九品心、以為證驗, DBZ 22, 370a, 370b, 388b. 287 如眞諦三藏、引『楞伽經』「八九種種心」. . . DBZ 22, 370a, 370b, 388b. ŌTAKE Susumu points out that the wording of this quotation matches no Chinese translation of LAS, and suggests that this is because the quote genuinely dates back to P, who translated himself and directly from his knowledge of Skt. LAS; he regards this as supporting evidence for the likelihood that a special text on ninth consciousness by P really did exist (ŌTAKE 2007d). However, given the confusion that Daoji seems to evince in this very same passage about his other source, the Shiqi di lun (in which case he is possibly referring to or “citing” a text he himself never even saw, and thus basing himself on hearsay), it does not seem that we have strong grounds for confi - lun. He rejects it as a proof-text on the grounds that “has never circulated in this country [i.e. China] 此國未行”.288 Daoji himself thus never saw the Shiqi di lun, and any information he gives us about it is based on hearsay.289

Once again, Daoji does not say that the Shiqi di lun contained a special section whose title had anything to do with “ninth consciousness”. Daoji was a Shelun scholar in a direct line from Paramārtha himself. His teacher, Jingsong (one remove from Paramārtha), may have written a text entitled Jiushi yi zhang (see below), which we would expect his disciple Daoji to have known. If even Daoji does not say that Paramārtha’s Shiqi di lun (or JDZL) contained a special chapter on “ninth consciousness”, and if he nonetheless refers to Shiqi di lun (and LAS!) for his textual support, rather than a dedicated text by Paramārtha on ninth consciousness, then it seems unlikely such a text existed by the 630s.

In addition, Daoxuan (道宣, 596-667) reports that two other texts with similar titles were written by important Shelun figures. (1) Tanqian 曇遷 (542-607) is supposed to have written a Jiushi zhang 九識章.290 Tanqian was extremely influential in spreading Shelun School thought to the North, and a prominent fi gure under the Sui.291 (2) Jingsong (靖嵩, 537-614) is supposed to have written a Jiushi xuan yi 九識玄義.292 Jingsong was a disciple of Fatai 法泰 and so a “dharma grandson” of Paramārtha himself; he was also, like Tanqian, an important figure in the transfer of the Shelun school to the North.293 Jingsong was also pivotal because he was the master of Daoji, an important Shelun-school witness to later *amalavijñāna doctrine (see below). It is thus not impos-

dence that Daoji certainly quoted word-for-word from texts on paper, rather than roughly, from memory or hearsay. The slight difference in wording here might therefore only be evidence that Daoji’s “quotation” is actually simply a “near-enough” paraphrase. 288 DBZ 12, 370b; in fact, as we will see below, Daoji goes to some lengths to fi nd alternative proof- texts for the notion of ninth consciousness, precisely because he is so suspicious of the Shiqi di lun. 289 This ignorance about the text is mirrored by the fact that details of Daoji’s reference to the text also seem confused; his mention of 決定說 is perhaps a vague echo of the title of Jueding zang lun, and it is possible that he knows that these two titles have something to do with one another, but is not sure what. 290 Xu gaoseng zhuan 續高僧傳, T2060:50.574b04. It is important to remember, in assessing this report, that Daoxuan’s information is never repeated in any other source. 291 See C HEN Jinhua. 292 Xu gaoseng zhuan, T2060:50.502a02. 293 See n. 366. sible that one of these texts became associated in the tradition with Paramārtha, leading to the reports we have seen above of a similar text in Paramārtha’s name.

Finally, we should also consider that the earliest mention of a ninth consciousness, or nine consciousnesses, is found not in Paramārtha or his successors, but in the Laṃkâvatāra sūtra (LAS), as early as the translation of Bodhiruci. The verse in question reads: “The various consciousnesses, eight or nine in kind/ Are like waves on water.” This passage, or its doctrine, was frequently referred to as later scholastics discussed the concept of *amalavijñāna and ninth consciousness: for example, by Jingying Huiyuan, T2807, Li Tongxuan, Kuiji, Tunnyun, and Tankuang. This shows that later authors were interested in using LAS to furnish a scriptural warrant for *amalavijñāna; or using the notion of *amalavijñāna to interpret this cryptic passage in LAS; or using LAS to account for the errors perceived as inhering in the notion of *amalavijñāna; etc. Whatever tack the various later scholars took, these passages suggest that LAS is one possible alternative source of the enumeration of *amalavijñāna as a ninth consciousness.

Surveying this tangled body of evidence, we can discern several main points.

First, there seems to be considerable confusion about the title of the text(s) ascribed to Paramārtha (九識義品, 十七地論菩薩品, 九識義章, 九識品, 九識義記, 九識章, 九識品, 決定藏論有九識品, 九識論), which we find alongside a number of other locutions which may or may not even refer to a freestanding text (九品心, 九識義, 釋九識中). We also see much confusion about whether it this was a freestanding text or a chapter in a larger text (usually specifi ed as in JDZL), or (perhaps) both. This confusion does not inspire confi dence that many authors had actually seen such a text.

There is a clear association in some quarters between this supposed text and JDZL (or Shiqi di lun). If “ninth consciousness” refers to *amalavijñāna, this association is partly justifi ed; however, extant JDZL *amalavijñāna doctrine does not call it a “ninth consciousness”. Further, as we will see, the actual ideas about *amalavijñāna that are expounded in JDZL almost never appear again in the record.

We can further sort the above materials into three main groups: (1) the idea that JDZL or Shiqi di lun expounded the idea of ninth consciousness, without any claim about a section with a related title: Huijun, T2805 and T2807, and Daoji. (2) the idea that there existed a separate text on the same topic: beginning (possibly) with ZSL, or Fei Changfang and the AF preface; the root-text of Daoji’s Shelun zhang (only as reported by the much later Gyōnen); the Da Tang neidian lu; the Faxiang authors Wŏnch’uk and Tankuang; possibly Wŏnhyo; and the Japanese catalogue of 735. In Wŏnch’uk we find the unusual hybrid assertion that such a freestanding text (Jiushi zhang) cited a special chapter of JDZL. (3) The idea that JDZL contained a text with a title to do with ninth consciousness, expounding the same. This idea is fi rst seen in Wŏnch’uk, then in Dajue, Tunnyun and Chengguan. In Wŏnch’uk, moreover, we fi nd the unusual hybrid idea that this chapter was the source for a separate text called the Jiushi zhang.

Thus, the broad pattern seems to be as follows. Our earliest evidence contains two confl icting accounts. In the fi rst century after Paramārtha, one line of evidence holds that the doctrine of ninth consciousness is grounded in JDZL (among other texts; WXL, MSgBh etc. are also cited). This line of evidence is found in our best informed, most scrupulous early sources: Huijun and texts closely associated with the Shelun school, which go into detail about doctrines and accurately cite Paramārtha’s real texts. Alongside this, we have texts which, especially in the early period, only touch very fl eetingly on the whole problem, and never show themselves to be well informed about the concrete contents of Paramārtha’s texts. These texts hold that the doctrine of ninth consciousness is found in a separate text by Paramārtha containing the term “ninth consciousness/nine consciousnesses” in the title, but cannot agree on what that title is. At the same time, we also have accounts that inform us that two important Shelun School authors in the late sixth century, Tanqian and Jingsong, wrote texts with “ninth consciousness/nine consciousnesses” in the title. There is thus a possibility that the freestanding text of that title ascribed to Paramārtha might, like AF, have been incorrectly attributed to him, and actually authored by someone else.

When we arrive at the Faxiang authors, the characterisation of the textual basis for the theory of ninth consciousness changes. Wŏnch’uk seems to be pivotal. In him, we see an assertion, never repeated, that there exists a freestanding text, and that it cites a section of JDZL also named for the doctrine of ninth consciousness. Perhaps Wŏnch’uk was attempting to reconcile the two contradictory traditions that had preceded him, i.e. that there was a text with jiushi in the title, and that the doctrine was expounded primarily in JDZL.305 After Wŏnch’uk, all authors but one take to saying that ninth consciousness doctrine is expounded in a chapter of JDZL named for that doctrine. The idea that there existed a separate text only recurs in catalogues and in Tankuang.

While we cannot be sure, it thus seems likely that neither JDZL nor Shiqi di lun ever contained a section with the term “ninth consciousness” in the title. The idea that such a text existed seems rather to emerge over a century after Paramārtha’s death, as an attempt to reconcile conflicting traditions, and then to be repeated in a way that shows the extent of the authority of Wŏnch’uk (which he shared with Kuiji; see below).

It is even more diffi cult to know whether or not Paramārtha did indeed compose a freestanding text with jiushi in the title. On the one hand, the texts that do say Paramārtha wrote such a text are less clearly reliable. On the other hand, the portion of Fei Changfang’s Lidai sanbao ji about Paramārtha’s translations may have been based on a list drawn up by Cao Bi; CHEN Yinque has shown that the apocryphal AF preface still contains considerable accurate historical information; and Wŏnch’uk and Chinkai cite the Jiushi zhang on topics other than ninth consciousness. These facts and others discussed above seem to me to suggest that at least one text with a title like Jiushi zhang certainly existed, and that it was almost certainly ascribed to Paramārtha as early as 590. However, perhaps that attribution, and indeed the text itself, was not widely known for quite some time beyond that (otherwise why would Daoji not cite it?). Beyond this, however, it seems to me impossible to exclude either of two mutually contradictory possibilities: (a) the text was by Paramārtha, or (b) that it was by another author, and the attribution was apocryphal. 4.1.2.3 Did Paramārtha propound a ninth consciousness?

The idea that there are nine consciousnesses, and the identification of *amalavijñāna with the ninth consciousness, was certainly current and well-known in the scholastic Buddhism of North China from the early Sui (by the late 580s or 590). All our most reliable witnesses for this early period relay to us this idea. We have no evidence that is closer to Paramārtha and his circle than these witnesses, that might give us grounds to doubt this testimony. We must recognise the possibility that these ideas were genuinely propounded by members of the group (including perhaps Paramārtha himself), but were lost from the extant record of their texts.

On the other hand, these ideas are absent from the extant Paramārtha corpus. Further, as we will see, even our earliest witnesses were not very well-informed about Paramārtha’s actual doctrines of *amalavijñāna; many of the aspects of the doctrine we have seen above are entirely absent from their accounts, and each of them appears ignorant of the bare existence of at least some key texts, let alone their contents. We have also seen that the notion of nine kinds of consciousnesses could have been derived from LAS, and *amalavijñāna labelled a ninth consciousness in order to make sense of the LAS passage and furnish the Shelun school theory of mind with more textual support. It is also possible, then, that a nine-consciousness model grew up in the early Shelun school, to reconcile earlier convictions that the ground of mind was pure Thusness with the idea that ālayavijñāna was the repository of all defiled seeds. Given that our present evidence gives us no fi rm testimony of the existence of the idea of ninth consciousness until perhaps as late as 590, we cannot exclude the possibility that the doctrine of ninth consciousness/nine consciousnesses was an early post-Paramārtha development, which was then ascribed to him as founder of the Shelun school.

It therefore seems impossible to determine for sure whether or not Paramārtha or his group expounded ninth consciousness, or nine consciousnesses. 4.1.3 Later reports of concrete contents of the doctrine of *amalavijñāna In tracing the later development of *amalavijñāna doctrine, we must treat the reports of several individuals separately, since there is relatively little agreement between them. This lack of consensus alone suggests that there was a lot of creative interpretation mixed in with these reports. 4.1.3.1 Jingying Huiyuan

Jingying Huiyuan (淨影慧遠, 523-592) places *amalavijñāna under a broader rubric of “true” 真 consciousness, which is twofold, including also ālayavijñāna. He says that amala means “taintless” in Chinese 此云無垢, and also “originally pure” 本淨. He says further that it is referred to as “taintless” in the sense that the substance 體 of what is true (or Thusness) is permanent and pure 真體常淨故曰無垢. He equates it with the “Thusness aspect of mind” 心真如門, a term clearly derived from AF; he quotes AF as a proof-text in the next line, so connecting *amalavijñāna to tathāgatagarbha. Elsewhere, Huiyuan again associates “ninth consciousness” with the “Thusness aspect of mind” 心真如相, and ālayavijñāna with “the aspect of mind [that is subject to] arising, cessation and conditions” 心生滅因緣相. The ninth consciousness is the “substance of all dharmas” 諸法體. Both *amalavijñāna and ālayavijñāna are part of the same mind, but the difference is that ninth consciousness is the state in which language is cut off and conditionality is transcended, whereas eighth consciousness is the state in which [[[mind]]] conforms to the metamorphoses brought about by conditions.

Already, the concept of *amalavijñāna is clearly being interpreted in part through the lens of AF. As a result, the relationship posited here between *amalavijñāna and ālayavijñāna differs from that found in Paramārtha’s works. Paramārtha understood *amalavijñāna to be the counteragent to ālayavijñāna, and the two to be in a temporal relationship to one another, whereby ālayavijñāna existed only until liberation, and was then succeeded by fully realised *amalavijñāna. For Huiyuan’s post-AF analysis, by contrast, the two are different facets of the same “true” consciousness, *amalavijñāna in its pure, eternal, self-contained and transcendent purity, and ālayavijñāna as it is engaged with and even immanent in saṃsāra.

Huiyuan also links *amalavijñāna as “ninth consciousness” to the LAS passage mentioned above. Here again, he says that within the rubric of a ninefold analysis of consciousness, there are two possible analyses, depending upon whether one understands “true” and “false” 真妄 as (1) distinct 分別 or (2) as dialectically “analysed and then synthesised” 離合. (1) In the former perspective, the “true” aspect is twofold, and comprises *amalavijñāna on the one hand and ālayavijñāna on the other. (2) On the latter analysis, only “the fundamentally pure *amalavijñāna” 本淨阿摩羅識 is “true”, and ālayavijñāna is included among eight consciousnesses that are an “amalgam of true and false” 真妄和合. This approach is again redolent of AF. However, these are only two among a longer list of various modes of analysis consciousness admits of, which also include tenfold and elevenfold analyses.

It seems, then, that Huiyuan is wielding the concept of *amalavijñāna and related concepts in the pursuit of his own hermeneutic projects, and is not simply concerned with giving us accurate doxographic reports of Paramārtha’s own doctrine. Huiyuan also shows little sign of direct acquaintance with Paramārtha’s own pertinent texts. 4.1.3.2 Zhiyi

The bulk of Zhiyi’s (智顗, 538-597) discussion of *amalavijñāna is found in his Fahua xuanyi 法華玄義, most likely dating to around 593. Zhiyi maps the different levels of consciousness, in which he includes *amalavijñāna, onto his unique doctrine of “three dharmas” 三法 or “three rules” 三軌: *amalavijñāna is the “rule of Thusness [itself]” 真性軌; ālayavijñāna corresponds to the “rule of contemplation of Thusness” 觀照軌; and ādānavijñāna 阿陀那識 corresponds to “the rule of extending this understanding to the workings of Thusness” 資成軌. Zhiyi also says explicitly that the difference between *amalavijñāna and ālayavijñāna is just that *amalavijñāna is ālayavijñāna in which the seeds of wisdom exist and in which the “perfumation (vāsanā) of hearing” has grown, so that it undergoes a “revolutionary transformation of the basis” (āśrayaparāvṛtti) and is transformed into “Thusness after the path” (道後真如). He identifi es *amalavijñāna with something he rather idiosyncratically calls “the light of nirvikalpakajñāna” 無分別智光. Elsewhere, Zhiyi also says that *amalavijñāna is the “consciousness” of a Buddha, whereas ālayavijñāna is the consciousness of the bodhisattva, and ādānavijñāna, which he calls the “seventh” consciousness and identifi es with *prativikalpavijñāna 分別識, is proper to the two lesser vehicles.

As this brief overview shows, much of what Zhiyi has to say about *amalavijñāna is unique to him, and looks like the result of creative attempts to coordinate what he knows of *amalavijñāna from other sources with other doctrines and his own original system. 4.1.3.3 Huijun/Hyegyun

Important evidence about the image of Paramārtha’s doctrine in late sixth-century China is preserved in Huijun’s (慧均, d.u., fl . 574-590s?) Si lun xuan yi 四論玄義 X784. It is diffi cult to know exactly when Huijun was writing, but the Shelun school already canon, but it is diffi cult to be sure of their authenticity and date. Even if these two texts are not actually by Zhiyi, however, we should not automatically exclude their evidence. They may both nonetheless contain information as early as the period immediately after Zhiyi’s death, and thus still comprise some of our earliest evidence. (1) The 金光明經玄義拾遺記 X356 is a collection of fragments made in 1023 by Zhili 知禮 (960-1028) of the Shanjia 山家 faction, intended to disprove accusations from representatives of the Shanwai 山外 faction that parts of the Jin guangming xuan yi were apocryphal. It may contain sub-commentarial layers, and these are further of uncertain date; SATŌ 451. This text says *amalavijñāna is an “unmoving consciousness” (不動識), and says it is another name for prajñā and awakening 覺了;

X356:20.42b08-09. It further identifi es *amalavijñāna with the attainment of a sophisticated insight 觀 into the nature of mind, which has both empty and non-empty aspects; this insight does not hypostasise either the empty or the provisional, but understands their dialectical interrelationship; 60b10-13. It also maps this schema onto the analogy to gold, earth and impurities from MSg; *amalavijñāna equates with the gold, and is all that is left when full buddhahood is attained; 48c15-18. (2) The Chan men zhang 禪門章 X907 is probably not actually by Zhiyi, but is rather a commentary on Zhiyi’s 次第禪門. SATŌ thinks it probably dates after Zhiyi’s death (in 597), but otherwise is unable to speculate about its date; SATŌ 125, 276. This text includes *amalavijñāna in a string of different names which variously identify the absolute, all of which have in common that they strike the happy medium (madhyamapratipad) between the extremes of various false dualisms like conditioned/unconditioned, bondage/liberation, worldly/transcendent, defi led/pure etc. In this context, *amalavijñāna is identifi ed with ultimate truth (paramārthasatya), Buddha nature (foxing 佛性), Thusness, the “limit of what exists” (bhūtakoṭi), non-abiding, non-production etc; X907:55.645b17-22.

had some identity that it presented to outsiders, and thus Huijun’s knowledge of the concepts that concern us was fi ltered through a Shelun lens. The only *amalavijñāna text he mentions by name, he apparently misnames (Shiqi jing for Shiqi di lun), and he makes it clear that he has not seen the text himself. Huijun’s attitude to ideas he identifi es as belonging to Paramārtha is also probably coloured by his polemical hostility towards the Shelun school. that he most likely heard Falang lecture in person in 574, thus providing our most precise clue as to the dates of his activity. Other clues seem to indicate that he was a student of Falang from relatively early; that he was close to Falang over an extended period; and that he was slightly senior to Jizang; see MITSUGIRI 223-225, KANNO (2002) 87. It was long thought that the partial version of the Si lun xuan yi collected in the canon was his only surviving work, but modern scholars have discovered other parts of that text in Japan; see e.g. ŌCHŌ, and works listed in KANNO (2008) n. 1. Further, ITŌ has argued that the Mile jing you yi 彌勒經遊意 T1771, traditionally ascribed to Jizang, is also his (see ITŌ [1977] 847-848 for a summary of the reasons for this claim; also ITŌ 1973). It has also been proposed that the Dapin jing you yi 大品經遊意 T1696, also ascribed to Jizang, is by Huijun (CH’OE 26). Recently, CH’OE (infra) has proposed that Huijun may have been from Paekche 百濟, and also that the Si lun xuan yi may even have been composed in Paekche. (I will nonetheless refer to him as “Huijun”, not “Hyegyun”, because CH’OE’s theory is still new and speculative, and because Huijun was active in China and wrote in Chinese.) In the current state of our knowledge, it is not possible to know defi nitively the chronological relations between Huijun and Jizang (or their works), nor the exact date of the Si lun xuan yi. The full text of the Si lun xuan yi is thought to date at least to after Falang’s death in 581 (MITSUGIRI 225); it mentions events of the Sui, and figures like Huijue 慧覺 (554-606) and Huichong 慧衝, which would seem to indicate that the text was completed after Huijun was active in Chang’an 長安 under the Sui (Foguang dacidian 6029). As we saw, Huijun is thought to have been senior to Jizang; on the other hand, he also refers to Jizang (藏公, X784:46.599b02; KANNO [2008] 6). Despite this uncertainty in their chronological relations, I have placed Huijun before Jizang because he is thought to have been slightly senior; because the only fi rm date I know for his activities is his reference to hearing Falang lecture in 574; and because, to my knowledge, there is no fi rm evidence for his activity much beyond the time shortly after the death of Falang, whereas Jizang lived for several more decades. It is also possible that Huijun is the earliest among our sources after P. Both Jingying Huiyuan and Zhiyi were active into the 590s, and Zhiyi’s texts, further, were in many cases revised and expanded by his disciple Guanding, who lived until 632. Given that Huijun was active before the 580s, this may mean that some information in his text(s) about P predates these other sources.

Against this, however, we should also weigh the following factors. Huijun is potentially a particularly reliable witness not only because he is close in time to Paramārtha, but because he knows Paramārtha’s works better than many of our other witnesses. He quotes SBKL verbatim;334 he also accurately quotes MSg335 and unique parts of Paramārtha’s MSgBh.336 This puts Huijun among a very small number of later witnesses to *amalavijñāna/“ninth consciousness” doctrine who quote Paramārtha verbatim, or even cite texts and loci in which *amalavijñāna is in fact expounded. Huijun also knows at least one term that is found very rarely outside the writings of Paramārtha himself. In addition, Huijun is unlikely to himself be consciously applying AF concepts to the interpretation of Paramārtha’s ideas (though the ideas may have already passed through 334 非淨非不淨, X784:46.599c18, 645c06; quoting SBKL 非淨非不淨 T1616:31.863a27-28, 非淨非不淨 863b19; in the second passage, particularly, this phrase is associated immediately with *amalavijñāna and prakṛtiprabhāsvaracitta, for which see SBKL<1> above. See also FXL T1610:31.795c02. 335 此師九□義并『攝論』云：「此[[[界]]?]無始時、一切此[for 法?]依止」 X784:46.645b24, probably citing MSg: 此界無始時 / 一切法依止 T1593:31.114a01, cf. anādikāliko dhātuḥ sarvadharmasamāśrayaḥ | tasmin sati gatiḥ sarvā nirvāṇādhigamo ’pi ca, TrBh 37, 12-13. 336 彼『論』三種佛性中「自性住佛」 X784:46.599c20-21, and again in greater length at 602a02-06, referring to MSgBh T1595:31.200c23. some AF fi lter before they reached him). This is because Huijun knew AF, but was suspicious of it, believing it a Chinese apocryphon. On the whole, then, Huijun is a seemingly reliable witness - early, well-informed and scrupulous.

Huijun says, more than once, that *amalavijñāna/ninth consciousness is beyond language and conception or thought (想, possibly meaning *saṃjñā[-skandha]), and even, quite specifi cally, that it cannot be known by consciousness itself. He further associates “pure consciousness” 淨識 with prakṛtiprabhāsvaracitta 自性清淨心, citing in the process SBKL<1>, in which *amalavijñāna is indeed associated with that concept. He links these doctrines to tathāgatagarbha doctrine, but only indirectly, insofar as he ascribes to both the Dilun and Shelun schools the notion that an “aboriginal storehouse (or garbha) consciousness 本有藏識 is the substance of the essence of mind 心性之體”, which is prevented from being manifest by adventitious defilements; on this view, he says, the process of becoming a Buddha is identical with the removal of these defi lements. This view is associated clearly with the idea that “we do not speak of ‘Buddha nature’ only upon the attainment of buddhahood, but rather, it is precisely by means of the present manifestation of an original, hidden/latent mind that buddhahood is achieved; the original nature is neither changed nor lost, and thus we speak of ‘Buddha nature that always indwells’.”

In the context of describing a varied set of views about what comprises the substance 體 of Buddha Nature, Huijun also alludes to “ninth taintless (wugou) consciousness” 第九無垢識 (ascribed to the Shelun school, not to Paramārtha himself). The ninth and tenth positions he discusses are those of the Dilun and Shelun schools. He says that the Dilun masters hold ālayavijñāna, counted as an eighth consciousness 第八無沒識,344 to be the substance qua ontological cause (正因體, where 正因 is probably for kāraṇahetu345); the Shelun masters hold “ninth, taintless consciousness” to be the same. Huijun is critical of both views. In the course of this

344 The term 無沒識 is very unusual. So far as I can determine, it never appears in a translation text, and the earliest texts in which it appears seem to be the present text by Huijun, and texts by Zhiyi (T1716:33.699c15, 744b22; T1777:38.552a10; T1783:39.4a13-14), Jingying Huiyuan (X753:45.107c14-15) and Jizang (T1824:42.119a23-24); the term also appears early in the Dunhuang Shelun text 攝大乘論章卷第一 (T2807:85.1013a27). Jingying Huiyuan also says that wumo is the “proper translation” of ālaya, e.g. 「阿梨耶」者、此方正翻名為「無沒」, T1851:44.524c18, also 530b09-10. This leaves open the mysterious question of where this term for ālayavijñāna comes from. I have been unable to fi nd even the epithet wumo alone, clearly applied to any kind of consciousness, in any translation texts earlier than the texts cited here. 345 Shengyin 生因 (kāraṇahetu) means a cause due to which something comes into existence, i.e. an ontological cause; liaoyin 了因 (jñāpakahetu) means a cause due to which an act of perception or knowledge takes place, i.e. an epistemological cause; see OGIHARA s.v. shōin, ryōin. The distinction is explained in MPNS, T374:12.530a16-26 = T375:12.774c23-775a03, T374:12.593a1119 = T375.:12.841a01-10; and also in Kuiji’s commentary on the Nyāyapraveśa 因明入正理論疏 T1840:44.101b29-c28. Kāraṇahetu is variously compared to the seed from which a plant grows, the clay from which a pot is made, etc., while jñāpakahetu is commonly compared to lamplight that illuminates objects. The terms were known to P: see FXL T1610:31.798a07-10 (explaining different phases of the realisation of Buddha-nature); Rushi lun T1633:32.32c28-33a01 (the reason argued does not ontologically produce the ineternity of sound, but only brings about realisation of that ineternity, i.e. it is a jñāpakahetu for that ineternity, not a kāraṇahetu). I am grateful to Prof. FUNAYAMA Tōru for pointing out these references. For zhengyin opposed to liaoyin, in roughly the same sense as shengyin = kāraṇahetu, see MPNS: 有二種因：一者正因、二者了因。尼拘陀子以地水糞作了因、故令細得麁, T374:12..532b14-16.

exposition, Huijun mentions an understanding of “taintless consciousness” that seems accurate to what we have seen of it in Paramārtha, namely that it succeeds upon the cessation of ālayavijñāna 無沒盡顯無垢.348

To summarise, then: Huijun calls *amalavijñāna a “ninth consciousness”, and uses the term wugoushi for the same; he seems indifferently to identify the views of Paramārtha and those of the Shelun school; he understands that *amalavijñāna/ninth consciousness is beyond language and conception, and even unknowable; he connects the doctrine of this consciousness to prakṛtiprabhāsvaracitta; and he seems to report loose associations between this consciousness and Buddha nature/tathāgatagarbha (as part of a position indifferently ascribed to both Dilun and Shelun schools). 4.1.3.4 Jizang (writing c. 599-608)

Jizang (吉藏, 549-623) lists *amalavijñāna among a set of terms that are identifi ed by different schools or fi gures as what is “non-dual” (advaya, 不二). Jizang also does seem to correspond to the understanding that it is asaṃskāra, seen in Tib. and XZ parallels to JDZL<4> above. 348 X784:64.602a21; note, however, that this comment falls in the middle of a passage that is diffi - cult of interpretation. refers to the notion of *amalavijñāna in another doxographic passage discussing views of various schools on the “real/true” 實, saying that “scholars of the Mahāyānasaṃgraha” hold that the principle 理 of the twofold absence of self (nairātmyadvaya 二無我), the principle of the threefold absence of essence (*niḥsvabhāvatraya 三無性), and the *amalavijñāna can rightly be called “true/real”, but all else is false.352 In another passage, in his “Exposition of the Profundities of [the] Vimalakīrti [[[sūtra]]]” (淨名玄論, T1780), Jizang returns to this characterisation. According to him, “Adherents of the Mahāyānasaṃgraha and the *Vijñaptimātratā śāstra [唯識論 i.e. the Viṃśatikā]353 take non-attachment to the three natures as the principle 理 of the threefold absence of essence; this principle of the threefold absence of essence is *amalavijñāna, which is also the principle of the twofold absence of self. The ‘three natures’ are the interdependent nature (paratantrasvabhāva), the imaginary nature (parikalpitasvabhāva), and the perfected nature (pariniṣpannasvabhāva). . . The perfected nature is Nirvāṇa.”354

Elsewhere, Jizang also recalls a similar classifi cation to distinguish the ultimate, as it is characterised in all these schools, from the “Nirvāṇa of the true doctrine” (正法涅槃, *saddharmanirvāṇa) taught by his own position. In all these other cases, including that of *amalavijñāna as taught by the Shelun masters, he says that the instance in question is manifest upon the attainment of buddhahood, and this resultant state is called

Mahāprajñāpāramitôpedasa masters; and the Dilun school; where identifying the nondual as *amalavijñāna is the position of “the Shelun masters and Trepiṭaka Paramārtha” 『攝論』師、真諦三藏. See T1780:38.856c11-17, 912b09-18; T1853:45.66c02-06; 352 T1824:42.123c22-124a02. Jizang also mentions this same doxographic characterisation at 126c04-07. The other positions are: for adherents of the Abhidharma 阿毘曇人 (i.e. Satyasiddhi specialists) it is the principle of the Four Noble Truths, and most specifi cally of the third truth of cessation (nirodhasatya); for the [[[Wikipedia:adherents|adherents]] of the] “Mahāyāna of the south” (?南土大乘; this is the only time this phrase ever occurs in the canon) it is the “principle of the truth of refutation” (?破諦之理); for “those in the North” (i.e. Mahāprajñāpāramitôpadeśa exponents) it is prajñā that carries the mark of reality 實相波若. 353 Given that these are referred to as a separate group of scholars, Jizang may be referring to exponents of the text in Gautama Prajñāruci’s translation, T1588, translated around 540. For example, Tanqian’s biography reports that he studied this text before he went to the south, and so presumably before he had access to P’s texts (it is known that he only encountered the Mahāyānasaṃgraha, for instance, after he fled Zhou Wudi’s 577 persecution of Buddhism); CHEN Jinhua 14-15 n. 12. 354 T1780:38.897b06-16. dharmakāya because these schools maintain that certain dharmas defi nitively exist. Finally, Jizang also briefl y mentions a difference of opinion between the Dilun position, in which the “six consciousnesses” are eradicated to leave ālayavijñāna, and that of a certain Trepiṭaka and certain “masters” 三藏師 (presumably Paramārtha and the Shelun masters) who say that “the eighth consciousness is also eradicated, since it too is not pure; [only] the ninth, *amalavijñāna, is entirely pure”. Jizang again differs with these positions because they still posit a dualism of pure and impure.

Jizang’s explanation of the doctrine is rare in linking *amalavijñāna to three natures doctrine, and more specifi cally to the perfected nature. The general silence on this matter contrasts with the fact that it was so central to *amalavijñāna in Paramārtha. However, other members of the string of identifi cations Jizang ascribes to the doctrine are new: *amalavijñāna is also identifi ed with the nondual in a new sense, with the twofold absence of self, and with Nirvāṇa; the identifi cation with dharmakāya in T1853 is also new. Jizang also ascribes the positions he describes either to the Shelun school and Paramārtha indiscriminately, or else only to the Shelun masters, on occasion not even mentioning Paramārtha. Even if Jizang reports everything with fi delity, we apparently see here a version of the doctrine already fi ltered through the early Shelun school.

On the whole, Jizang’s presentation of *amalavijñāna, like those of Zhiyi and Huiyuan, is also clearly bound up with his own intellectual agendas, and his attempts to fi t the material into his own doctrinal system. 4.1.3.5 She dasheng lun zhang

The anonymous Dunhuang Shelun text She dasheng lun zhang 攝大乘論章 T2807 argues that the same consciousness can either be called eighth or ninth, and cites LAS in support. Like Huiyuan, it adduces the AF categories of a “Thusness aspect” 真如門 and “saṃsāric aspect” 生滅門 of mind to negotiate the apparent contradiction between eightfold and ninefold analyses, identifying “ninth consciousness” with the Thusness aspect. It then cites what it calls the “Chapter on Marklessness” (wuxiang pin 無相品, actually citing our extant SXWL) from the Wuxiang lun: “Because the imagined essential nature never exists, the interdependent essential nature also does not exist; and the inexistence of these two [[[essential]] natures] is *amalavijñāna (阿摩羅識).” It then says that this consciousness is “the ultimate, unique pure consciousness” 究竟唯一淨識. The text then says that a tradition from outside China reports that the Shiqi di lun contains a “Bodhisattva Chapter”, which gave an extensive exposition of *amalavijñāna as ninth consciousness.

This passage from T2807 probably conforms more closely to what we see in our extant Paramārtha corpus than any other later account of *amalavijñāna. Even here, however, we see a certain admixture of AF concepts. 4.1.3.6 Prabhākaramitra’s Mahāyānasūtrâlaṃkāra (tr. 630-633)

P r a b h ā k a r a m i t r a ’ s ( 波羅頗蜜多羅, 5 6 4 - 6 3 3 ) t r a n s l a t i o n o f t h e

Mahāyānasūtrâlaṃkāra T1604 famously mentions *amalavijñāna. The mention features as part of commentary on the verse corresponding to Skt. 13.19.362 This verse is part of a set dealing, in significant part, with “aboriginally luminous mind” (prakṛtipra bhāsvaracitta), which, we saw above, is connected with *amalavijñāna in Paramārtha’s SBKL. These verses assert that defi lement (niḥsaṃkleśa) and purifi cation (viśuddhi) do not really exist, but are illusory, like a magic trick or “space” (ākāśa) (3.16). This is likened to the way a flat picture, skilfully executed, appears to contain height and depth; similarly, there is in fact no dualism (dvaya) in the imagination of what is unreal (abhūtakalpa), but it appears as if dual (3.17). Water is intrinsically clear, even when tainted by mud, and when the mud is removed, the water is not changed, but rather, its original true nature simply becomes manifest (3.18). 3.19 spells out the parallel to this conceit in the case of the mind: “I have explained that the mind is pure in essence 心性淨 But is defi led by adventitious dirt 客塵; There is no essential purity of mind Apart from the Thusness of the mind 心真如.”

In other words, like water, the mind is pure all along and by its very nature; it is not the case that, when it is purifi ed, some new, pure mind is produced in the process.

In the Bhāṣya to this verse, Chinese features an extra sentence that does not correspond to anything found in the Sanskrit. “It is this mind[, equated with Thusness,] that is expounded as aboriginally pure 自性清淨. This mind is *amalavijñāna 此心即是阿摩羅識.” This reinforces the association of *amalavijñāna with prakṛtiprabhāsvaracitta. This association, made in SBKL, Huijun, and Daoji, appears again in Wŏnhyo and Chengguan. 4.1.3.7 Daoji (writing c. 633-637)

One of the most important moments in the history of *amalavijñāna doctrine after Paramārtha is found in Daoji’s (道基, 577-637) lost Shelun zhang 攝論章, which is   quoted by Gyōnen 凝然 (1240-1321) in the Kegon kōmokushō hatsugo ki 華嚴孔目章發悟記 15. Daoji’s most important comments about *amalavijñāna/ninth consciousness are contained in a single extended passage. “Question: Upon what sūtras and śāstras do Trepiṭaka Paramārtha and the Dharma Master Daoji base themselves, in proving the tenet of ninth consciousness? “Answer: In the fi rst [juan?] of [Daoji’s] She lun zhang [“Treatise on the Mahāyānasaṃgraha”], it says: ‘There are Dharma Masters who expound nine consciousnesses. For example, Trepiṭaka Paramārtha cites [the line], “Various kinds of mind, eight- or ninefold” from the Laṃkâvātāra sūtra; [he] also cites the defi nitive exposition 決定說 of nine kinds of mind 九品心 from the Shiqi di lun [“Treatise on the Seventeen Stages”] as proof. From then, right down to the present, the controversy has not ceased, so that later generations have no way of deciding [what is correct]. ‘Here, we will determine that it is correct to hold that: the sūtras expound six [kinds of consciousness]; some, however, say there are seven; the Laṃkâvatāra expounds eight; and the Wuxiang lun has nine. Among these various theories, the Laṃkâvatāra etc. expound only eight consciousnesses because they are expounding an abridged [version of the doctrine] (?據略但說八識).

‘On the other hand 或復說云, the doctrines of the Wuxiang [[[lun]]] lay out the nine in full (? 義具通陳其九). The “Chapter on *Pravṛttivijñāna 轉識品” in the Wuxiang lun says, “The [[[consciousness]] that is the] subject of perception is of three kinds: (1) Consciousness [[[arising]]] as a result [of karma] 果報識 (vipākavijñāna), that is, ālayavijñāna 䖘耶; (2) Consciousness that ‘grasps’ [or ‘is attached’] 執識, that is, ādānavijñāna 阿陀那; (3) Sense consciousness [literally ‘consciousness of sense objects’] 塵識, that is, the six consciousnesses [of eye, ear etc.].”373 When it has finished explaining thus, the text goes on to expound *amalavijñāna 阿摩羅識. Thus, the [same Wuxiang] lun says, “The simultaneous disappearance of both object and consciousness is precisely the perfected nature (pariniṣpannasvabhāva); and the perfected nature is precisely the *amalavijñāna.”374 The [[[Wuxiang]]]lun also says, “Ālayavijñāna 阿䖘耶識 is of eight different kinds, as is explained in the ‘Chapter on Nine Consciousnesses’.”375 This treatise, having expounded eight consciousnesses, in addition expounds separately a pure consciousness 淨識 called *amala 阿摩羅, and further says “as is explained in the ‘Chapter on Nine Consciousnesses’”. It is clear that the principle of the nine consciousnesses we are expounding here is paramount in it [? 勝焉: i.e. in that text?]. ‘In addition, the Mahāyānasūtrâlaṃkāra says, “By the transformation 轉 (*parāvṛtti) of eighth consciousness, ‘mirror-like wisdom’ 鏡智 (ādarśajñāna) is obtained; by the transformation of seventh consciousness, ‘the wisdom [that recognises the] equality [of all things]’ 平等智 (samatājñāna) is obtained; by the transformation of the fi ve [external sense] consciousnesses [sic], ‘wisdom of [[[Wikipedia:perfect|perfect]]] cognition’ 觀智 (pratyavekṣajñāna) is obtained; and by the transformation of the manovijñāna [sic], ‘wisdom that achieves its tasks’ 作事智 (anusthānajñāna) is obtained.”376 This sentence [refers to] eight consciousnesses. However, the same treatise also says, “It is this mind[, equated with Thusness,] that is expounded as aboriginally pure 自性清淨. This mind is amalavijñāna 阿摩羅識.”377 When we add this mind to the previous eight, are there not nine [altogether]? ‘Trepiṭaka Paramārtha may cite [the line], “Various kinds of mind, eight- or ninefold” from the Laṃkâvātāra sūtra; and also the defi nitive exposition 決定說 of nine kinds of mind 九品心 from the Shiqi di lun. However, even though the Laṃkâvatāra does indeed say “eight-

before P, and also almost never after him in translation texts). 373 ZSL: 能緣有三種：一、果報識、即是阿梨耶識。二、執識、即阿陀那識。三、塵識、即是六識, T1587:31.61c08-09. 374 ZSL: 境識俱泯、即是實性、實性即是阿摩羅識, T1587:31.62c18-19; see §2.4 above. 375 ZSL: 就此識中、具有八種異、謂依止處等。具如『九識義品』說, T1587:31.62a03-04. 376 M S A : 轉第八識得鏡智、 轉第七識得平等智、 轉第六識得觀智、 轉前五識得作事智, T1604:31.606c29-607a02. This text appears in Bh to the verse correspond to Skt. 9.67, LÉVI (1907, 1911) 1, 46; 2, 88; however, nothing in the Sanskrit corresponds to this Chinese. 377 MSA: 心真如、名之為「心」、即說此心為自性清淨。此心即是阿摩羅識, T1604:31.623a08-09; see n. 365 above and corresponding text. or ninefold”, it does not lay out the names [of the types of consciousnesses concerned]; and the Shiqi di lun was transmitted throughout (?攝傳?) India, but has never circulated in this country 此國未行. Thus, these two modes of exposition can hardly constitute proof [of the doctrine]. Here, we have taken up passages from the Wuxiang lun and the Mahāyānasūtrâlaṃkāra to explain that there are nine consciousnesses, taking them as reliable [proof-texts]. ‘Question: If nine consciousnesses are expounded, then why does the Laṃkâvatāra only expound eight consciousnesses? Answer: In the Laṃkâvatāra, only eight consciousnesses are expounded because [the text] only bases [itself on a view that] takes the objects (ālambana) as consciousness (?) 但據緣境為識. [However,] in the Wuxiang lun and the Mahāyānasūtrâlaṃkāra, both mind and object are taken as consciousness, and so it explains in full [all] nine(?) 以心境俱識、通說九乎. Or again, the Laṃkâvatāra only bases itself on [the point of view of] saṃsāra 生滅, [and so] expounds only eight consciousnesses; but in Wuxiang lun etc., the doctrine encompasses [both] true and deluded 真妄 [aspects].’”

Elsewhere, Daoji states that Paramārtha holds the substance of “pure consciousness” 淨識體 to be Thusness and the wisdom that takes Thusness as its object 如如及如如智380 (Daoji disagrees with this doctrine, as we will see below). The She dasheng yi zhang (攝大乘義章, T2809), which was preserved in an ancient manuscript in Japan381 and is probably by Daoji,382 also briefl y mentions the “ninth consciousness”.383

Daoji’s citations, from texts like ZSL, MSA, SWXL, AF and the Mahāyānâbhisamaya, are exceptionally accurate, and show that he is a scrupulous scholar. Daoji’s testimony is also particularly valuable because of his critical attitude towards Paramārtha (perhaps surprising, in a member of the Shelun school towards its ostensible founder). This shows him a cautious commentator, not ready to simply believe anything. For example, he says that Paramārtha lists six alternative names for the “pure consciousness” (阿摩羅, 無垢, 淨識). These names are found together in no 380 DZB 22, 371b. Cf. T226265:440c-441a. 381 It is sometimes mistakenly said to be a text from Dunhuang. I am grateful to I KEDA Masanori for allowing me to see unpublished work in which he traces some of the history of this manuscript. 382 On Daoji’s probable authorship of this text, see K ATSUMATA 795. I am grateful to both Ching KENG and IKEDA Masanori for pointing out KATSUMATA’s arguments.

383 T2809:85.1036b28-c05. This passage is difficult to interpret. It apparently identifies *amalavijñāna with an ultimate (paramārtha) pure consciousness identical with the truth 第一淨識, whose substance is Thusness 體是如如 (which, as we have seen, Daoji consistently holds elsewhere). The true essence of this consciousness is supposed to exist aboriginally 真性本有, and is identifi ed with a gnosis (jñāna) that has no inception and cannot be cultivated 非始修智. extant Paramārtha text. Daoji repeats twice that Paramārtha cites the authority of no sūtra or śāstra for these names, and concludes, “My suspicions have not been allayed, and I cannot rely upon [this doctrine as Paramārtha expounds it]” 疑信未決、未可即 依. This critical attitude towards Paramārtha is also evidenced in an explicit disagreement Gyōnen reports that Daoji expresses with Paramārtha’s understanding of what comprises the “substance” 體 of consciousness.

Here, then, we see a Shelun scholar and direct “dharma descendant” of Paramārtha himself, who is yet sceptical of the standard proof-texts used to support the doctrine of ninth consciousness and keen to fi nd alternative, less vulnerable proofs. Nonetheless, Daoji apparently cannot fi nd anything better than ZSL and the hot-off-the-press MSA. Moreover, he also does not seem to know any additional Paramārtha texts since lost, except the Shiqi di lun, which he knows only by hearsay and is sceptical of. 4.1.3.8 The *Vajrasamādhi sūtra and Wŏnhyo’s commentary (approx. 649-686) The *Vajrasamādhi sūtra (VSS) T273 and Wŏnhyo’s (元曉, 617-686) commentary on it, the Kŭmgang sammaegyŏng non (*Vajrasamādhi sūtra lun, T1730), contain extensive new developments in the doctrine of *amalavijñāna. Here, *amalavijñāna is given outright (apocryphal) warrant as buddavācana, and a creative synthesis is attempted between *amalavijñāna and other concepts important to East Asian Buddhism.

The VSS fi rst mentions *amalavijñāna twice each in the following two passages: (VSS<1>): “At that time, the Bodhisattva Non-Abiding asked the Buddha: ‘Lord! Through revolutionary transformation (轉 *parā/vṛt) by what inspiration 利 do all the affective consciousnesses 一切情識 of sentient beings undergo a revolutionary transformation 轉 so that they enter into the *amala[-vijñana]?’ The Buddha replied: ‘All the Buddhas, the Tathāgatas, constantly employ the one awakening 以一覺 to [effect a] revolutionary transformation in all consciousnesses, so that they will enter into the *amala[-vijñāna]. This is because the original awakening 本覺 of all sentient beings [works,] by means of the one awakening, to awaken those sentient beings, and [thus] to make the sentient beings all regain their original awakening, viz. to awaken them to the fact that all affective consciousnesses are empty, tranquil and unproduced. That is because it is an established fact 決定 that the original essence is originally without motion.” (VSS<2>): “[The Buddha said:] ‘One who is enlightened need not abide in nirvāṇa 不住涅槃. Why is this? One who awakens to 覺 original nonproduction 本無生 remains far removed from the maculations (mala, 垢) of sentient beings. One who awakens to the original lack of tranquillity 本無寂 remains far removed from the activity of nirvāṇa 離涅槃動. For one who abides 住 at such a stage, the mind abides nowhere. Free from both egress and access 無有出入, it accesses the amala consciousness 庵摩羅識.’ The Bodhisattva Non-Abiding 無住菩薩 asked: ‘If the amala-consciousness 庵摩羅識 has some place where it can be accessed, [does this mean it is] something that is attained (得, upalabdhi) ― that is, an attained dharma (處有所得是得法)?’ The Buddha replied: ‘No, it does not.’”

The VSS goes on to relate a parable of a prodigal son (VSS<3>), who carries gold coins in his hands, but does not know it. For fi fty years he roams in poverty and destitution, before his father fi nally tells him he has been in possession of gold all along. The moral of the story is spelt out thus: “It is just the same with the *amala[-vijñāna]. It is not something from which you have departed 出, and now, it is not ‘accessed’ 入. Just because you were deluded in the past, does not mean you did not have it; and just because you have realised now [that you have it], does not mean you have ‘gained access’ to it 入.”

VSS also mentions “ninth consciousness” twice: (VSS<4>): “The Buddha [said:] ‘Those who recite the Prāṭimokṣa precepts do so because of their unwholesome haughtiness, which is [like] waves and swells on the sea. If the sea, i.e. the ground of their mind in the eighth consciousness, is limpid, then the ‘[out-]fl ow’ (流, āsrava/ ogha) will be purified from the ninth consciousness. Where no wind moves, waves cannot arise. The precepts are by nature uniform and empty (śūnya) 等空; [those who] hold fast to them are deluded and confused.” (VSS<5>): “[The Buddha said:] Thusness is empty (śūnya) in its essence 真如空性. The fi re of the gnosis [that knows] this emptiness of essence 性空智火 completely burns up all fetters (結, saṃyojana). All is utterly uniform 平等平等, and the three stages of equivalent enlightenment 等覺三地 and the three bodies of sublime awakening 妙覺三身 are radiant, clear and pure within the ninth consciousness 於九識中䇷然明淨, [so that] there are no shadows.”

There is much that is new here. Astonishingly, the VSS is the fi rst time after Zhiyi in the later evidence that we see even the faintest allusion to āśrayaparāvṛtti, despite the centrality of it to Paramārtha’s doctrine of *amalavijñāna. We also see here a connection of *amalavijñāna to a kind of “other-power” doctrine, in which access to it is explained by the good works of the Tathāgatas (VSS<1>). This is also the first time we have seen *amalavijñāna associated with the doctrines of “non-abiding nirvāṇa” (無住涅槃, apratiṣṭhitanirvāṇa), “non-production” (anutpāda), and “neither coming nor going” (BUSWELL’s “free from both egress and access”). It is also the first time we have seen *amalavijñāna associated with “non-abiding” (VSS<2>). The parable (VSS<3>) is clearly modelled on the basic conceit of some of the nine parables of the Tathāgatagarbha sūtra, even though it matches none of those parables exactly ― by this means, the association between *amalavijñāna and tathāgatagarbha is clearly further strengthened. The metaphor of the “waves and the sea” (VSS<4>) derives from LAS, and is, of course, by this stage in Chinese Buddhist history, famously associated with AF. This is the first time we have seen it used in direct connection with *amalavijñāna. VSS<5> also draws an implicit connection between *amalavijñāna and the “taintless stage” (wugou di 無垢地 = *amalabhūmi) of the path to buddhahood, as it was expounded in the Chinese apocryphon the Pusa yingluo benye jing (菩薩瓔珞本業經, T1485).

Wŏnhyo develops these rich ideas even further in his commentary. As BUSWELL shows in his translation and study of Wŏnhyo’s commentary, Wŏnhyo picks up on these hints in the root text and elaborates them into a theory whereby *amalavijñāna becomes the key to a scheme of practice, by means of which the practitioner can progress from the ordinary to the awakened state. The key innovation and doctrinal goal in Wŏnhyo’s commentary is the attempt to bring out an “active” dimension of tathāgatagarbha, using a framework derived from AF (“original enlightenment” benjue 本覺, “acquired enlightenment” shijue 始覺, etc.). The main way Wŏnhyo achieves this is to map *amalavijñāna onto the path structure of the Pusa yingluo benye jing. Thus, “as Wŏnhyo interprets [the VSS], the enlightenment that is immanent in the mundane world . . . could actually be viewed as a practical catalyst to religious training.” Wŏnhyo constructs “a comprehensive system of meditative practice, focusing on the six divisions of contemplation practice that lead to the experience of ‘the contemplation practice that has but a single taste’,” which constitutes “a practical way of actually cultivating original enlightenment, rather than just passively acquiescing to it”. On this reading, “the Vajrasamādhi-sūtra provides a practical soteriology of original enlightenment by shifting the Awakening of Faith’s accounts of mind and enlightenment from ontology into the realm of actual practice.”411

The most important points in Wŏnhyo’s exposition of *amalavijñāna are as follows. In his introduction, Wŏnhyo says that as the result of the six practices advocated by the sūtra, the “ninth consciousness” appears by a revolutionary transformation 轉 (parā/vṛt). He calls this resulting ninth consciousness wugoushi 無垢識, and identifi es it with the dharmadhātu. This is the fi rst time we have seen *amalavijñāna associated with the dharmadhātu since SBKL<1>. The process of realisation continues with the revolutionary transformation (轉 once more) of the eight consciousnesses into the four wisdoms (ādarśajñāna, samatājñāna, pratyavekṣajñāna, anusthānajñāna), and the attainment of the three bodies (trikyāya) of the Buddha. We have not seen the idea of the attainment of all three bodies connected to *amalavijñāna before; the only place we have previously seen *amalavijñāna connected with the four wisdoms is in Daoji’s use of MSA. Wŏnhyo further says that in this state, gnosis and its object are nondual 境智無二.

Later, Wŏnhyo states explicitly that “original enlightenment” is identical to *amalavijñāna 本覺正是唵摩羅識; he therefore glosses the notion of “accessing *amalavijñāna” seen in VSS<1> as “attaining benjue” 得本覺. The ground for this move was obviously prepared for Wŏnhyo by the author of VSS in passage <1> above; Wŏnhyo is merely spelling out what is already there implicit. In glossing VSS<2>, Wŏnhyo explains that “accessing *amala[-vijñāna] means “leaving behind the two extremes” 離二邊, and he identifi es the attainment of *amalavijñāna with “returning to the fountainhead of the mind” 歸此心源. These claims reinforce the ties between *amalavijñāna and the AF framework.

In commenting on the parable of the foolish son, Wŏnhyo comments that the gold has the four qualities of permanence, bliss, self-identity and purity 常樂我淨. The use of these four well-known epithets of tathāgatagarbha strengthens still further the association between tathāgatagarbha and *amalavijñāna. He also assigns these epithets to benjue. He goes on to employ the LAS/AF fi gure of wind, waves and water, associating the underlying tranquil substratum (the sea) with prakṛtiprabhāsvaracitta 自性清淨心. Mention of this concept puts Wŏnhyo in a line that includes Huijun, MSA and Daoji before him, and Chengguan after.

Finally, in one other text, the Niepan zong yao 涅槃宗要, Wŏnhyo explicitly identifi es *amalavijñāna with Thusness, the substance of Buddha-nature 佛性體, and, most interestingly, jiexing 解性. This last notion derives from Paramārtha’s MSg, and seems itself to have been the centre of a process of reinterpretation in line with tathāgatagarbha doctrine as mediated by AF ideas. This is the fi rst time we have seen this idea associated with *amalavijñāna.

In sum, in the VSS and Wŏnhyo, *amalavijñāna is an important concept, and it undergoes a number of striking new developments. It is elaborated into a basis for practice; it is associated with the path structure from the Pusa yingluo benye jing; it is linked to the four wisdoms and the three bodies; it is tied much more closely to tathāgatagarbha, Buddha nature, and their four epithets of permanence, bliss, self-identity and purity; and it is associated still more with AF rubrics and concepts. These texts also seem to revert to some old and, from what the extant texts show us, authentic dimensions of Paramārtha’s doctrines, including the link to āśrayaparāvṛtti, and the identifi cation with the dharmadhātu, and the nonduality of perfect gnosis and its object.

From this point in our analysis, as we venture into periods more distant from Paramārtha himself, the testimony of authors who mention *amalavijñāna/ninth consciousness tends only to become more uniform, more removed from anything Paramārtha himself said, and more derivate of intervening accounts. For these reasons, we will not treat individual the remaining authors to 800 so exhaustively, but rather, will only pick out the main trends in their accounts. There are two main groups of ideas we need to look at: those of the Faxiang school; and the beginning of a reaction against those ideas, as seen in the Vinaya authors Dingbin and Dajue, and in Chengguan. 4.1.3.9 Faxiang authors

We turn fi rst to Xuanzang’s Faxiang school. During roughly the same few decades when the VSS and Wŏnhyo’s commentary were composed, Xuanzang and his team would have been busy on their massive translation projects, and his disciples would have been producing the fi rst of their signifi cant body of commentarial literature. In this literature, they occasionally commented on Paramārtha’s doctrines, including *amalavijñāna. The three most important Faxiang authors to comment on *amalavijñāna and ninth consciousness are Kuiji, Wŏnch’uk and Tunnyun. We will here take them as representative, noting additional information supplied by other authors as necessary.

The battle against the notion of the so-called “ninth consciousness” was a key part of the struggle of the Faxiang school to roll back the ongoing synthesis of Yogācāra and tathāgatagarbha thought, which it saw as heterodox. This polemical setting exerts a signifi cant distorting infl uence over their presentation of the doctrine. This bias notwithstanding, the massive historical infl uence of the Faxiang position in the interpretation of Yogācāra/Vijñaptimātra doctrine in East Asia has arguably exerted an excessive infl uence over our understanding of *amalavijñāna, right down to modern scholarship.

Kuiji discusses *amalavijñāna in the context of a larger discussion of various rubrics numbering consciousness various ways. He begins by citing the LAS passage saying consciousness can be eightfold or ninefold. He then says: “On the basis of the Wuxiang lun and the Mahāyānâbhisamaya sūtra, this [verse, i.e. from LAS is interpreted to] [mis-]take 取 Thusness for a ninth consciousness, because it expounds the combination of two aspects, one true [[[consciousness]]] and eight worldly [[[consciousnesses]]]. Thus, [this interpretation] [mis-]takes 取 the eighth, fundamental consciousness 本識, when it is in a state of purity, for a ninth, expounding a distinction between the defi led and pure 染淨 fundamental consciousness. The “Sūtra of the Adornment of the Tathāgata by Merits” 如來功德莊嚴經 says: ‘The taintless consciousness 無垢識 of the Tathāgata Is the pure 淨 element/realm without ‘outfl ows’ (anāsravadhātu 無漏界); It is liberated from all obstructions (sarvâvaraṇa 一切障); It is conjoined with (*saṃyukta 相應) the cognition that is like a perfect mirror (圓鏡智, ādarśajnāna).’ “Since, here, the text speaks of “taintless consciousness” 無垢識 and “cognition that is like a perfect mirror” together, and since, further, the ninth is called *amalavijñāna 阿末羅識, we [therefore] know that [the text] is expounding a distinction between the defi led and pure [aspects of the] eighth consciousness, and taking [the latter] 以為 for the ninth.”

This passage was the key point of reference for later Faxiang writers discussing *amalavijñāna. We see here a number of hints that suggest Kuiji’s understanding is based quite closely on Daoji. Kuiji also adduces a new proof text for the concept, coming to him via none other than the (for his school) immensely authoritative Cheng weishi lun (the only source for the Rulai gongde zhuangyan jing verses).

In keeping with the Faxiang attempt to assert their orthodoxy over Paramārtha, we see here for the fi rst time the outright assertion that there is something wrong with the concept of ninth consciousness: it is based upon (grasping) misapprehension 取. Kuiji was elsewhere even more forthright about criticising this notion: “A former master set up [wugoushi] as a ninth consciousness. This is an error 古師立為第九識者非也.” In his commentary on the Cheng weishi lun (the Shuji 述記), he again cites the LAS verse, and then explains the nature of the error: “To say there is a ninth consciousness is superfl uous 增數. It is manifest that the other-dependent consciousness (*paratantra-vijñāna, 依他識) includes three types [of consciousness], when considered in general 略, and only 唯 eight, when considered in detail 廣. It is beyond increase and decrease 離於增減, and that is why we use the word ‘only’ 唯. The Laṃkâvatāra doubles up in its exposition[, saying] that if we take the eighth as defi led, and separate out its pure [aspect], we can speak of a ‘ninth consciousness’. This does not mean that there are nine [[[consciousnesses]]] in the substance 體 of the other-dependent consciousness 依他識, and it also does not mean that in terms of substantial kinds 體類, there is a separate ninth consciousness.”436

Apart from these passages, Kuiji only comments briefl y and inconsequentially on *amalavijñāna/ninth consciousness.437

In his famous commentary on the Saṃdhinirmocana sūtra, Wŏnch’uk438 also uses the term wugoushi in a manner that clearly attempts to recuperate the term as merely another name for ālayavijñāna. “The Trepiṭaka Paramārtha asserted a doctrine of nine consciousnesses on the basis of the Jue ding zang lun. “Nine consciousnesses” refers to: six consciousnesses (of the eye, etc.) . . . the seventh ādāna . . . and the eighth, ālayavijñāna, which is of three kinds . . . The ninth is *amalavijñāna 阿摩羅識, which here [in China] is called ‘taintless consciousness’ 無垢識. It takes Thusness for its substance 體, such that, in the same Thusness, there are two aspects 義: (1) the object [of gnosis] (‘noema’, 所緣境), which is termed Thusness, or the ‘limit of reality’ (bhūtakoṭi); (2) the subject [of gnosis] (能緣, ‘noesis’), which is termed ‘taintless consciousness’ 無垢識, and is also termed ‘original awakening’ (benjue, 本覺). This is as is explained by the Jiushi zhang, citing the “Ninth Consciousness Chapter” of the Jueding zang lun 具如九識章引決 定藏論九識品中說.”439

Later, Wŏnch’uk returns to the topic, arguing that wugoushi is correctly just another name for ālayavijñāna:

subject of the eponymous Cheng weishi lun. 436 T1830:43.239a12-16. 437 In one brief passage, Kuiji says that *amalavijñāna is only found at the stage of “the fruition of buddhahood” 唯在佛果; T1829:43.179a04-05. Other passages in which Kuiji comments are T1782:38.1001c26-29 and T1861:45.282c19-25. 438 We have seen that W ŏnch’uk (圓測, 613-696) (who is Kuiji’s senior in years, but as we shall see, depends upon him in the explication of *amalavijñāna), uses the transcription 阿末羅識 (seen for the fi rst time here and in Kuiji), as well as the standard 阿摩羅識. He also uses the term wugoushi to refer to *amalavijñāna. We also saw already that he refers to a specifi c text by P on the “nine consciousnesses”, which he calls “Essay on Nine Consciousnesses” (Jiushi zhang 九識章), saying it quotes the “Ninth Consciousness Chapter” of JDZL. 439 X369:21.240b20-c07. “[[[Ālayavijñāna]]] is also called ‘taintless consciousness’ (wugoushi), because it is utterly pure, and the basis for all dharmas ‘without outfl ow’ (anāsravadharmas, 諸無漏法). To explain: the Sanskrit word [for this term] is *amalavijñāna 阿末羅識, which here [in China] means ‘taintless consciousness’ (wugoushi). This is the state of [[[ālayavijñāna]] in?] sublime awakening 妙覺位. The substance of mind when it is conjoined 相應 with wisdom that is like a mirror (ādarśajñāna) is called wugoushi. It is utterly pure; all the dharmas of the path that are without ‘outflow’ (*anāsravamārgadharma), [such as] wisdoms (jñāna), states of absorption (samādhi) etc., take it as their basis. Thus the ‘Sūtra of the Adornment of the Tathāgata by Merits’ (Rulai gongde zhuangyan jing) says: ‘The taintless consciousness of the Tathāgata/ Is the pure element without ‘outfl ows’;/ It is liberated from all obstructions;/ It is conjoined with the cognition that is like a perfect mirror.’ This is explained in detail in the third juan of the commentary on the [[[Cheng]]]weishi [[[lun]]] 唯識疏.”

This explanation applies purely to the ālayavijñāna, even though it accepts (or speculates) that *amalavijñāna was the original term for wugoushi as cited in the Cheng weishi lun.

In one other brief comment in his commentary on the Sūtra of Humane Kings 仁王經疏 Wŏnch’uk overtly addresses a doctrine he ascribes to Paramārtha: “Further, Trepiṭaka Paramārtha propounded a total of nine consciousnesses, adding *amalavijñāna 阿摩羅識, which has as its essence Thusness-cum-original awakening 真如本覺. While it is in [a state of] bondage, it is called tathāgatagarbha; when it escapes bondage, it is called the dharmakāya 在纏名如來藏出纏名法身. Here [in China,] we call it ‘taintless consciousness’ 此云無垢識. This is as [it is expounded] in the Jiushi zhang 九識章. The remaining eight consciousnesses are roughly the same as in all the masters.”

The distinction here adduced between “in bondage” and “free from bondage” derives from a new proof-text in the history of *amalavijñāna doctrine, the seminal tathāgatagarbha scripture, the Śrīmālādevīsiṃhanāda sūtra, though the exact wording used by Wŏnch’uk appears in no Chinese sūtras, but rather, fi rst in the writings of this Faxiang generation and their successors. This is a new link between *amalavijñāna and tathāgatagarbha doctrine. Unfortunately, it does not seem there is any way of knowing whether Wŏnch’uk had it from some earlier text, and if so, what; or whether he perhaps added this fl ourish himself.

In his Yuqie lun ji 瑜伽論記, Tunnyun445 provides us more new information: “Here, [[[Hui]]-]Jing [惠]景446 follows 擬 Master Paramārtha in establishing the doctrine of nine consciousnesses 九識義 on the basis of a citation from the ‘Ninth Consciousness Chapter’ 九識品 of the Jueding zang lun. However, in the portion of the Jueding zang lun corresponding 即 to the second part 分 of the present śāstra, there never was any ‘Ninth Consciousness Chapter’. Further, Master [Wen-]Bei [[[文]]]備447 says that an old tradition 昔傳 cites the [notion of] *amalavijñāna 阿摩羅識 from the Wuxiang lun 無相論 to prove that there are nine consciousnesses.448 The Wuxiang lun corresponds to the ‘Chapter on Absence of Essence (niḥsvabhāvatā) 無性品’ from the Xianyang [shengjiao] lun 顯揚論, but in that chapter, the term *amalavijñāna does not feature. Now, based upon the doctrine that there are nine consciousnesses from the Laṃkâvatāra sūtra etc., [we can say that] the ninth is called *amalavijñāna, which here [in China] would be said, ‘taintless’ 此云無垢. Master Ji 基 [i.e. Kuiji] says [of this]: [Tunnyun here quotes in full the long Kuiji passage translated above, p. 143.] “Divākara (地婆訶羅, fl . 676-688) says that there is also an interpretation 解, in Western lands 西方, that holds that a separate aspect of the sixth consciousness 六識 [i.e. manovijñāna] is called *amala 阿摩羅, because it has the excellent function 勝用 of eradicating ignorance and realising

cause it is cited in RGV: yo bhagavan sarvakleśa-kośakoṭi-gūḍhe tathāgatagarbhe niṣkāṇkṣaḥ sarvakleśa-kośa-vinirmuktes tathāgatagarbhakāye ‘pi sa niṣkāṇkṣa iti, JOHNSTON 79, 147; WAYMAN and WAYMAN 96. 445 We have already seen that Tunnyun ( 遁倫, d.u., Silla monk of the eighth century) uses the transcriptions 菴摩羅識, 阿末羅識 (citing Kuiji), and 菴末羅識 (citing Kuiji) as well as the ordinary 阿摩羅. We also saw that he reports that Huijing 惠景 traced the doctrine back to a “Ninth Consciousness Chapter” of JDZL. 446 Wenbei and Huijing were apparently late Shelun school fi gures; Y OSHIMURA (2002), 234. 447 See n. 446. 448 Given that Wenbei was a Shelun school fi gure, we can speculate, on the basis of the evidence to hand, that this is most likely a reference to Daoji. cessation 斷惑證滅. Dharma Master [Wŏn-]Hyo of Silla says that prakṛtiprabhāsvaracitta 自性清淨心 is called *amala 阿摩羅, and that it is of one substance with the eighth [[[consciousness]]], ālayavijñāna, but different in aspect 義別. Here, I follow (‘retain’, 存) this interpretation, which accords well 善順 with the sūtra [quoted] above.”

Tunnyun’s comments here stand out for the scholarly care with which he reports and evaluates various positions and traditions. He has also clearly taken pains to gather all the relevant information he could; but he is still entirely reliant upon relatively late, second-hand information. It also seems he could only base his assessment on parallels to Paramārtha’s texts in translations by Xuanzang; he apparently did not refer to Paramārtha’s own works on the topic.

Perhaps the most signifi cant thing about the treatment of *amalavijñāna and ninth consciousness/nine consciousnesses in these Faxiang authors is the overtly critical tone they adopt. Where early Sanlun authors (Jizang and Huijun) were also critical of Paramārtha, this was in a context in which they were critical of a number of schools, and for reasons somewhat tangential to Yogācāra concepts. Here, however, we encounter a head-on attack precisely on the concept of ninth consciousness itself, and the discussion is almost entirely governed by this polemic; such authors mention only those aspects of the doctrine that are necessary to refute it. In this polemical context, we should be alert for possible distortions of the doctrine, to make of it a straw man or a sitting duck.

Faxiang authors adduce a new proof text, the otherwise entirely unknown “Sūtra of Adornment of the Tathāgata by Merits”. Moreover, Faxiang authors implicitly also use Cheng weishi lun as a proof text (since their comments are often occasioned by pertinent passages in that text); the conjunction of the Mahāyānâbhisamaya and WXL as prooftexts imply they are probably reliant upon Daoji; and authors after Kuiji refer to Kuiji himself as an authority on the question. In addition, Faxiang authors refer to the follow- ing texts already referred to by their predecessors: LAS, WXL, and the Mahāyānâbhisamaya. The transcriptions they use seem to indicate that they have encountered the idea of *amalavijñāna in part through the writings of Tiantai Zhiyi.

By contrast, Faxiang authors make very little reference to Paramārtha’s own works: they only refer to the supposed Jiushi zhang, and to JDZL (with no signs that they had actually read JDZL itself). Given the reasons for caution in believing the traditional ascription of a Jiushi zhang to Paramārtha, there is very little to give us confi dence that the Faxiang authors were engaging with a textually grounded version of *amalavijñāna, traceable to Paramārtha himself.

This paucity of fi rm information is refl ected in the contents of the doctrine the Faxiang authors describe. They are more concerned to tell us what it is not, i.e. the “correct” understanding that in their view should be substituted for its mistakes; and what is wrong with it by contrast. They have very little to say about the actual content of Paramārtha’s doctrine: only that it counts *amalavijñāna as a ninth consciousness; that it associates *amalavijñāna with Thusness; (in Wŏnch’uk only) that it has two aspects, as object (Thusness etc.) and subject (benjue etc.); that it is the basis for anāsravadharmas; and that it has two states, after the manner of tathāgatagarbha, i.e. in and out of bondage.

Now, it seems highly likely from his citation of WXL and the Mahāyānâbhisamaya, and perhaps the reference to ādarśajñāna, that Kuiji is deriving his information about *amalavijñāna/ninth consciousness from Daoji. It is also apparent that Wŏnch’uk, already, is in part following Kuiji in his interpretation of the problem. We recall that Daoji was a teacher of Xuanzang, and also a student of Jingsong, to whom is ascribed a Jiushi xuan yi.456 These facts in combination make it possible that a Jingsong-Daoji version of the doctrine was the proximate source of Faxiang information about it, and the most immediate target of the Faxiang polemic.

Further, we have seen that an almost riotous variety of various ideas about *amalavijñāna/ninth consciousness was current in the century between Paramārtha and the Faxiang authors. By contrast, looking ahead towards the eighth century, we see that after Kuiji and Wŏnch’uk, the range of ideas we encounter is signifi cantly impoverished, and very often dependent on the Faxiang authors themselves.

It thus seems that we encounter here a bottleneck in the transmission of “ninth consciousness” lore: ideas possibly acquired via Daoji’s Shelun-school lineage are recast in a form most suitable for their treatment as the targets of a polemic, and this then becomes the dominant guise in which the lore is known to later generations. 4.1.3.10 Two Dharmaguptaka Vinaya authors in the early eighth century We turn next to Dingbin’s (定賓) Sifen lü shu shi zong yiji 四分律疏飾宗義記 (c. 703-705) and Dajue’s 大覺 Sifen lü xingshi chaopi 四分律行事抄批 (712), both in the Dharmaguptaka Vinaya 四分律 lineage. The main signifi cance of these two authors is that they disagree with Faxiang authors and side with Paramārtha. In so doing, however, they show that even scholars after the Faxiang authors but outside that school were heavily infl uenced by them in their understanding of the doctrine.

Dingbin cites a new proof text, translated only after the time of Kuiji and Wŏnch’uk: the Ghanavyūha-sūtra (Dasheng miyan jing 大乘密嚴經, T681), which held, in a manner reminiscent of LAS, that mind could either be of eight or nine kinds. He then cites an explanation of the ninefold system of consciousnesses, ostensibly from Paramārtha but most likely at best derived second-hand from the Wŏnch’uk passage above. This description begins with the six Abhidharmic consciousnesses, and then goes through seventh ādānavijñāna and eighth ālayavijñāna. It then comes to the ninth: “The ninth is called *amalavijñāna. The Tang Trepiṭaka says that here [in China], this is translated ‘taintless’ (wugou). This is also another name for the eighth consciousness. When one becomes a Buddha, the eighth consciousness undergoes a revolutionary transformation (轉, parā/ vṛt), and becomes taintless. There is [thus according to Xuanzang] no separate ninth consciousness. [This is explained] in detail in the *Vijñaptimātra (唯識, prob. Cheng weishi lun), the Mahāyānasaṃgraha etc.”

Dingbin goes on elaborate further on how Xuanzang and his schoolmen disagreed with this doctrine. However, he then refers to the Ghanavyūha again, as proof that it is in fact the Faxiang understanding that is incorrect: “The ninth here is what is described in the last juan of the Ghanavyūha as follows: ‘The pure garbha of the Tathāgata/ Is also called *amalajñāna 無垢智.’ Thus we can see that [[[Paramārtha’s]] understanding] is not in error 不謬.” He then goes on to describe the concrete contents of the doctrine of ninth consciousness thus: “Trepiṭaka Paramārtha says that there are two kinds of *amalavijñāna: (1) as object of gnosis (noema, 所緣), meaning Thusness; (2) “original awakening” 本覺, i.e. *tathatājñāna, 真如智. [This] subject of gnosis (noesis, 能緣) is identical with the “non-empty” (aśūnya) [facet of the] matrix ([[[tathāgata]]-] garbha); the object (noema) is identical with the “empty” (śūnya) [facet of the] matrix ([[[tathāgata]]-] garbha). According to the same (?通 ? for 同) treatise, both facets take Thusness as their substance 體.”465

This characterisation of *amalavijñāna as of two kinds is clearly derived from similar comments made by Wŏnch’uk, or from the same source Wŏnch’uk relied upon. However, it is not entirely derivative; it is the fi rst time we have seen *amalavijñāna associated with the concept of what is “non-empty”, another epithet of the tathāgatagarbha. In implicitly adducing the Śrīmālādevī as a proof-text, however, Dingbin may also be taking his cue from Wŏnch’uk, who we saw above refers to other concepts from the same text.

Dajue explicitly says he gets his information from “the Tang Trepiṭaka” (唐三藏, i.e. Xuanzang). He links *amalavijñāna to ādarśajñāna (like Kuiji’s Rulai gongde zhuangyan jing, probably following Daoji’s use of MSA), and says that ninth consciousness is called *amalavijñāna 阿摩羅識 in the “Ninth Consciousness Chapter” of JDZL (thus following Wŏnch’uk). Echoing Dingbin and Wŏnch’uk, he then says, “Paramārtha explains: ‘This [*amalavijñāna] is of two kinds: (1) as object of gnosis (noema, 所緣), meaning Thusness; (2) what is termed ‘original awakening’ (benjue), which is the gnosis itself (jñāna) qua subject of gnosis (能緣, noesis). The manovijñāna 意識 and this [[[amala]]] consciousness unite noema and noesis, which thus, in their unity, comprise the substance of this consciousness 合為此識體.”

The last (very cryptic) sentence here is new, but otherwise, Dingbin’s comments are entirely derivative of either the Faxiang authors or Dingbin’s use of Faxiang characterisations. Elsewhere, Dajue repeats verbatim a lengthy passage from Dingbin, showing that he was certainly aware of what Dingbin had to say about *amalavijñāna.470 It is thus diffi cult to be sure to what extent Dajue worked directly from Wŏnch’uk and Kuiji, and to what extent he had even their ideas indirectly via Dingbin.

The most important of these two scholars is clearly Dingbin, whose comments are more detailed, and earlier. Although Dingbin disagrees with the Faxiang position, he clearly has much of his information from Faxiang scholars, and reads Paramārtha’s doctrine through the lens of Faxiang views. It is also noteworthy that he has to fall back, for a proof text, on the Ghanavyūha, a new text translated even after Xuanzang’s era. Neither Dingbin nor Dajue seem to have direct access to Paramārtha’s texts. 4.1.3.11 Chengguan

Later comments by Chengguan (澄觀, 738-839) echo Dingbin and Dajue closely. Chengguan also sides quite openly with Paramārtha (as he perceives him) against Xuanzang.

In a fi rst extended discussion,471 Chengguan says that Paramārtha called the pure aspect of eighth consciousness, which Chengguan refers to as “the pure consciousness of the Buddha” 佛淨識,472 a ninth consciousness, and named it *amalavijñāna 阿摩羅識. Xuanzang said that this term should be translated wugou, and that it results from the revolutionary transformation (轉, parā/vṛt) of the eighth consciousness into a taintless consciousness (wugoushi) upon the attainment of buddhahood; but that it is not a ninth consciousness. Chengguan then gives the two Ghanavyūha-sūtra citations fi rst seen in Dingbin,473 saying they support Paramārtha in establishing a ninth consciousness. Chengguan then repeats verbatim Dingbin’s assertion, which says that *amalavijñāna is of two kinds, corresponding to the “non-empty” and “empty” aspects of

緣、 即是真如。 二名「本覺」、 即能緣智。 意識、 此識、 通能所緣、 合為此識體也。」 X736:42.876b19-22. X736:42.1019b20-23; passage discussed n. 22. T1736:36.323c03-17. This rare term would seem to derive from XZ’s translation of the *Buddhabhūmi sūtra śāstra: T1530:26.293a29, 311b01, 327c24-25. (Perhaps Chengguan is trying to hoist XZ with his own petard . . .) See above n. 459, n. 463.

tathāgatagarbha. Chengguan relates this doctrine to the two aspects 門 of mind from AF, i.e. the Thusness aspect 真如門 and the saṃsāric aspect 生滅門,475 quoting as a proof-text a section of AF saying that mind is only one, and is Thusness. He then says that whatever we call the mind in this liberated state, there are important differences between it and ordinary consciousness 凡識, namely that pure consciousness 淨識 creates the four wisdoms and the three bodies (probably following Wŏnhyo).

Later in the same text, Chengguan returns to the topic, in the course of glossing the notion of rushi xin (如實心, *yathābhūtacitta, “mind that is adequate to reality”, apparently a close relative of yathābhūtajñāna, 如實智). He identifi es this concept with prakṛtiprabhāsvaracitta (自性清淨心, “aboriginally luminous mind”), for which he refers again to the Śrīmālādevīsiṃhanāda sūtra and AF, and also cites a passage from MSA comparing prakṛtiprabhāsvaracitta with pure water that has been tainted by mud and dirt, which returns to its original purity when the taints are removed. He then argues that there can be no essence of mind separate from the Thusness of mind, and therefore, that purity of mind is merely a matter of removing adventitious defi lements. Thus, mind is identical with Thusness, and prakṛtiprabhāsvaracitta is identical with tathāgatagarbha, and also with original pure consciousness 本來淨識. Having thus prepared the ground, he introduces Paramārtha’s notion of *amalavijñāna, which he calls a ninth consciousness. The remainder of this passage repeats much of the argument he already laid out in the fi rst passage cited above.

There is much that is new in Chengguan’s discussion here. We have not seen *amalavijñāna related to “the pure consciousness of the Buddha”. We have seldom since Paramārtha seen such a close association between *amalavijñāna and prakṛtiprabhāsvaracitta, though the link does appear in Huijun, MSA and Daoji, and again in Wŏnhyo. It is also, to my knowledge, the fi rst time since Paramārtha that anyone has said so directly that *amalavijñāna is obscured by adventitious defi lements. In some respects, then, it is as if Chengguan is returning to aspects of Paramārtha’s original doctrines; and yet, despite the meticulous way he specifi es his sources, we have no indication that he has direct knowledge of Paramārtha’s works.

In Chengguan, then, it seems we see a strengthening of an initial reaction against the Faxiang rejection of Paramārtha’s ideas seen fi rst in Dingbin and Dajue. However, Chengguan elaborates this understanding in a creative way that is most reminiscent, if anything, of the mode of doctrinal development that we see begun in VSS and built upon by Wŏnhyo in his commentary. The *amalavijñāna/ninth consciousness as articulated in these texts has little to do with Paramārtha’s own doctrine. If *amalavijñāna, now fi rmly identifi ed with a ninth consciousness, is by the time of Chengguan on the verge of winning for itself a secure place in East Asian Buddhism, it is in a form that has declared almost complete independence from its original author.

The Faxiang authors articulated a very infl uential vision of *amalavijñāna doctrine, as we have seen. At the same time, they made the confl ict with Paramārtha so sharp it was almost a matter of “you’re either with us or against us”. In the long run, this may ironically have hastened the demise of the doctrine they opposed to Paramārtha’s. Apparently scholastics began to decide that they were “against them” ― the ideas ascribed to Paramārtha, even as Kuiji and Wŏnch’uk (inaccurately) described them, proved too attractive to reject entirely, and too well supported in a range of proof texts (none of them, by this stage, Paramārtha’s own!). Perhaps, then, we hear here one stroke of the death knell of Xuanzang’s “orthodox” Faxiang line against the “sinifi ed” line represented by the Yogācāra-tathāgatagarbha-Buddha nature-AF synthesis that eventually won out in mainstream East Asian Buddhism.482

the passage shared with Dingbin on the two kinds of *amalavijñāna, relating to the non-empty and empty aspects of tathāgatagarbha; and the reference to AF’s two aspects of mind. 482 For the sake of completeness, we should note that this survey has omitted the following later evidence of relatively marginal importance: (1) The Da foding rulai miyin siuzheng liaoyi zhu- pusa wanxing shoulengyan jing 大佛頂如來密因修證了義諸菩薩萬行首楞嚴經 T945, translated by Pāramiti/Pramiti (? 般剌蜜帝), who arrived in Canton in 705 (DEMIÉVILLE [1952], 43 n. 2), 4.1.4 Summary

Before we compare *amalavijñāna doctrine in later witnesses with that of Paramārtha himself, it will be helpful to identify some general trends in this later material.

First, the materials seem to fall into two main periods. A fi rst period lasts from approximately the 580s, or the early Sui, to the formation of Xuanzang’s Faxiang school. In this period, we see quite various impressions of *amalavijñāna doctrine, with little apparent centre of gravity. A second period lasts from approximately the middle of the seventh century, when Xuanzang’s school became active, until around the end of the eighth century, when we ended our survey. Understanding of *amalavijñāna from this period is dominated by the Faxiang authors, and their preoccupation with proving that eight, not nine, is the correct count for kinds of consciousness. In this same period, however, we see a second strand of material, represented mainly by VSS, Wŏnhyo, Dingbin and Chengguan, in which the gathering tendencies are to accept *amalavijñāna, sometimes by overtly rejecting the Faxiang position; to associate *amalavijñāna more and more overtly with tathāgatagarbha; and to creatively connect *amalavijñāna to a range of other ideas and texts.

contains one very brief reference to *amalavijñāna 菴摩羅識; T945:19.123c15. This is primarily of interest only because it is another instance of the term fi nding its way into a sūtra (cf. VSS) or Indic text (cf. also MSA). (2) Li Tongxuan (李通玄, 635-730) apparently mentions *amalavijñāna in his Xin Huayan jing lun 新華嚴經論 T1739, but his comments are notable mostly for their outlandishness. Li mistakenly refers to this consciousness as ādānavijñāna or “ninth consciousness”. According to Li, this doctrine is taught in the Saṃdhinirmocana sūtra! 解深密經 (T676:16.692c02-04, referring to XZ T676), but the actual content of this doctrine sounds more like a cross between LAS and rumours of P; T1739:36.722c22-23, 723a06-14, 723a23, 723b05-09, 723b12, 736a20-b02, 741b29-c01. (3) The Shi moheyan lun 釋摩訶衍論 (a commentary on AF, probably written sometime in the late seventh or eighth centuries; see YAMAMOTO Kazuhiko 山本和彦, “Shaku makaen ron”, s.v. Daizōkyō zen kaisetsu daijiten) mentions *amalavijñāna 唵摩羅識 in discussing a ninth consciousness, and quotes part of the passage I called “VSS<1>” above; T1668:32.611c22-27. (4) The She Moheyan lun shu 釋摩訶衍論疏 (said to have been compiled 集 by 法敏 Famin [579-645], but this seems a clear anachronism), a sub-commentary on T1668, quotes T1668 quoting the same VSS passage; X771:45.800c18-22. (5) We see brief mentions in Śubhākarasiṃha (善無畏, 637-735, arrived in China 716), Amoghavajra (不空金剛, 705-774) and Amoghavajra’s disciple Huilin (慧琳, d. 820); T906:18.913c07, T1177a:20.757c14-18, T2128:54.604c20. (6) Zhanran (湛然, 711-782) discusses ālayavijñāna as 了因 (jñāpakahetu) and “*amala 菴摩羅” as direct cause 正因

(kāraṇahetu?) (see n. 345), thus echoing Huijun and Li Shizheng (see n. 426). Moreover, as we move further away from Paramārtha, the trail runs cold in the hunt for genuine new information. In the earliest references to Paramārtha, it is diffi cult to be sure whether we are seeing the result of accurate doxography, creative endeavour, or inaccurate hearsay. Later, however, we fi nd authors clearly repeating their predecessor’s views. This suggests strongly that over time, *amalavijñāna lore became increasingly like a chamber of echoes or a game of “Chinese whispers”.

Throughout the period we have surveyed, our authors very seldom refer to any works from the extant Paramārtha corpus as evidence for their characterisation of *amalavijñāna/ninth consciousness: only Huijun, T2805 (whose direct knowledge of JDZL is established by its quote therefrom), T2807, Daoji, Kuiji (whose reference however seems possibly second-hand via Daoji), Wŏnch’uk as reported by Taehyŏn, and Wenbei as reported by Tunnyun. Verbatim citation of a known Paramārtha text in the discussion of this doctrine is even rarer, only occurring in Huijun, T2807 and Daoji.

Otherwise, where authors purport to refer to works by Paramārtha, they refer mostly to the mysterious special work on nine consciousnesses he is supposed to have composed. However, as we have seen, confusion seems to reign supreme over the exact title, location, nature and contents of this work, and there is little sign that any of the authors who refer to this work had themselves seen or read it. Apart from the supposed treatise or chapter on nine consciousnesses/ninth consciousness, the other main supposedly Paramārthian source authors refer to is the apocryphal AF.

Instead of making reference to Paramārtha’s works, extant or otherwise, authors reach for many other sources to piece together a picture of the doctrine, and *amalavijñāna gradually gets woven into a fabric of allusions to an ever-shifting range of new proof-texts. Thus, authors claim to fi nd the origins of the doctrine in MSg, the Viṃśatikā, LAS, the Mahāyānâbhisamaya, the elusive Rulai gongde zhuangyan jing, the Cheng weishi lun, the Pusa yingluo benye jing, the Tathāgatagarbha sūtra, MSA, the Śrīmālādevī, the Ghanavyūha, and even, in the singular case of Li Tongxuan, the Saṃdhinirmocana sūtra. Of course, it is the usual task of exegetes to fi nd or forge links like these between texts and doctrines they interpret and doctrines in other texts. At the same time, we are certainly justified in wondering why such diligent textual scholars would almost uniformly turn to such sources, and almost entirely overlook Paramārtha himself, if they had the choice of consulting Paramārtha’s own works directly.

Parallel to this process of weaving *amalavijñāna into a larger intertextual fabric, the doctrine also progressively becomes associated with more and more concepts. These include: dharmakāya; the “Thusness aspect of mind” of AF; Buddha nature; “neither increase nor decrease”; “not coming or going”; “original awakening” (benjue); jiexing; “non-abiding Nirvāṇa” (apratiṣṭhitanirvāṇa); the LAS/AF fi gure of the wind, the waves and the water; the epithets of tathāgatagarbha “eternal, blissful, self-identical, pure”; the dharmadhātu; wugoushi 無垢識; the “gnosis that is like a mirror” (ādarśajñāna); the tathāgatagarbha idea of “in bondage” and “free from bondage”; the non-empty and empty aspects of tathāgatagarbha; “Buddha consciousness”; various technical doctrines of causation as it relates to liberation; the “pure Buddha consciousness”; the “pure garbha consciousness of the Tathāgata”; and, of course, tathāgatagarbha itself.

This lengthy survey (§4.1) has thus shown that there is very much about *amalavijñāna in later sources that is never found in Paramārtha’s extant works. On the other hand, then, how much overlap is there with Paramārtha’s documented doctrine of *amalavijñāna? 4.2 What later sources say that agrees with Paramārtha

We saw above (§3) that there seem to be two quite distinct doctrines of *amalavijñāna in Paramārtha’s corpus. Here, however, I will treat all these text as a single unit, for purposes of comparison with later texts.

We find that there really is very little overlap between *amalavijñāna in Paramārtha and in later authors. This is in part a function of the wide variation of later authors among themselves. However, even if we take all the later sources as a unit for the purposes of comparison, it is remarkable how seldom they concur with our extant evidence about the doctrine they were ostensibly discussing. The only areas of frequent overlap are: 1) The term *amalavijñāna itself. However, some later sources only use the term “ninth consciousness”, or wugoushi, both terms that are not used in association with *amalavijñāna doctrine in Paramārtha himself; and transcriptions widely diverge from Paramārtha.

2) Discussion is at least about some kind of “pure” consciousness, as the term would lead us to expect. 3) *Amalavijñāna is a state of consciousness that attends liberation, and is attained through some transformation or purifi cation of ālayavijñāna. 4) The connection between *amalavijñāna and Thusness. However, Thusness also features prominently in AF, whose categories loom so large in the attempts of later authors to come to grips with *amalavijñāna. It is difficult to determine, therefore, whether this agreement is a function of accurate reporting of Paramārtha’s ideas from the later authors, or of the application of the AF lens. 5) The identification between *amalavijñāna and prakṛtiprabhāsvaracitta is touched upon in Huijun (actually citing SBKL), MSA, Daoji, Wŏnhyo (also reported second-hand by Tunnyun) and Chengguan. This link is thus the specifi c component of Paramārtha’s actual doctrine that most frequently recurs in later authors. It suggests that to the extent that the later tradition did base itself upon accurate information, it was working not from (reports of) JDZL, but SBKL. It is interesting to note that no Faxiang author notices this aspect of the doctrine.

There is also some reference in the later tradition to the following dimensions of Paramārtha’s doctrine, but it is slender. In many cases, we fi nd ourselves in a grey zone, where agreement could be a result of coincidence: 1) The idea that *amalavijñāna is related to āśrayaparāvṛtti is only mentioned explicitly by Zhiyi. It is also hinted at slightly in some authors. 2) The association between the attainment of *amalavijñāna and the realisation of a state “without ‘outfl ows’” (anāsrava), or the association between *amalavijñāna and anāsravadharmas, features briefly in the Rulai gongde zhuangyan jing verse cited by Kuiji, Wŏnch’uk and Tankuang. 3) The statement that *amalavijñāna is “permanent” appears only in Jingying Huiyuan, and does not then reappear in later sources until the loose association of *amalavijñāna with the four epithets of tathāgatagarbha in Wŏnhyo. (The related Paramārthian notion that *amalavijñāna is “true” because free from change [SWXL<1>] is never found later.) 4) The association between the attainment of *amalavijñāna and the attainment of power over body and lifespan is only weakly echoed in the identification of *amalavijñāna and dharmakāya in Jizang, Fali, Li Shizheng and Wŏnch’uk; and in the association between the attainment of *amalavijñāna and the three bodies (trikāya) in VSS<5>, Wŏnhyo, and Chengguan.

5) The identification of *amalavijñāna and the “per fected nature” (pariniṣpannasvabhāva) is only weakly hinted at in Jizang. 6) The idea that *amalavijñāna is obscured by adventitious defilements is only found in Huijun; by association in MSA; and in Chengguan. 7) Very little is made of the notion that *amalavijñāna is characterised by a nondualism of subject and object. We fi nd this notion refl ected directly only in Wŏnhyo. In Jizang, Dajue and Dingbin, the nondual also seems to feature, but it has a curiously different emphasis. 8) The relationship between delusion and language, or the relationship between the attainment of *amalavijñāna and the escape from language, is only refl ected in Huijun and Jingying Huiyuan. (We also do not fi nd much emphasis on Paramārtha’s related assertion that *amalavijñāna is free from error.) 4.3 What Paramārtha says that later sources do not

Comparing Paramārtha’s extant corpus our later sources, we fi nd that the following aspects of Paramārtha’s doctrine are never mentioned at all: 1) The association or identification of *amalavijñāna with the counteragents (pratipakṣa) of ālayavijñāna. 2) The idea that attainment of the *amalavijñāna entails a transformation of the relationship to the skandhas. 3) The association between the problematic of the attainment of *amalavijñāna, and liberation specifically understood as a process of evading rebirth, and thereby escaping future suffering. 4) The related overtones, found in Paramārtha, of the old Nikāya/Āgama doctrine of consciousness as the subject of transmigration and liberation; and indeed, any sign of a relationship between *amalavijñāna and the vijñānaskandha. 5) The identifi cation of *amalavijñāna with a “higher” stage of Vijñaptimātratā/weishi (weishi as the object of “perfect insight”, zhengguan weishi etc.), beyond the weishi “in practice” (fangbian weishi) that obviates only external objects but not the ordinary perceiving consciousness. 6) The notion that *amalavijñāna is free of “badness” (dauṣṭhulya), which, as we saw, is connected to its close association with āśrayaparāvṛtti. 7) The idea that *amalavijñāna is a basis for transcendent (lokôttara) dharmas. 8) Any association whatsoever between *amalavijñāna and the idea of the “continuum” (saṃtāna), either the ordinary continuum of the pṛthagjana before liberation, or the “continuum produced by lokôttaradharmas” of JDZL<3>. 9) The idea that *amalavijñāna is without a basis (依, āśraya), or even the very question of its relationship to a basis. 10) The “Mahāyāna Abhidharma” framework that is so key to the exposition of *amalavijñāna in JDZL. 11) The identification of *amalavijñāna with emptiness (SBKL).

I argued (§3) that there are two separable *amalavijñāna doctrines in Paramārtha, and that JDZL probably preserves the version closest to the original. We see here that some details of the version of the doctrine refl ected in the other group of texts (SWXL, SBKL, ZSL) are refl ected, if weakly, in later texts. However, the JDZL doctrine sinks almost without a trace.

This almost total silence on the actual content of JDZL forms a striking contrast to the fact that so many of the later sources claim to trace the notion of *amalavijñāna back to a putative “Ninth Consciousness Chapter” found precisely in JDZL.

5. Conclusions

There are very few areas of real overlap between *amalavijñāna doctrine in extant Paramārtha texts and in later sources. Of course, traditional bibliographies report that Paramārtha wrote many more texts than we have received. At least on the evidence of the extant texts, however, it seems that the tradition inherited from Paramārtha only a very basic idea of a pure, post-liberatory consciousness, in some relationship of contrast to ālayavijñāna, which had a close relationship or identity with Thusness.

Recognising the virtual certainty that some of Paramārtha’s texts and ideas have indeed been lost to our record, the possibility cannot be ruled out that some of what the tradition reports was in fact part of the doctrine of Paramārtha or his group. In particular: 1) We saw (§4.1.2.2, §4.1.2.3) that we cannot be sure that Paramārtha did not author a text especially on “ninth consciousness” (Jiushi zhang etc.), or that he did not teach a ninth consciousness or a system of ninefold consciousness. 2) We cannot be sure that he never associated *amalavijñāna with tathāgatagarbha, and may indeed therefore have been attempting thereby to effect some kind of rapprochement between Yogācāra and Tathāgatagarbha thought.

Testimony that Paramārtha’s teaching had these features is early and widespread. We have no contradictory evidence intervening between the earliest witnesses and Paramārtha’s group, which might allow us to cast doubt on this testimony. Thus, it is entirely possible, if not certain, that Paramārtha taught that *amalavijñāna was a separate, ninth kind of consciousness, associated with tathāgatagarbha. If such important aspects of Paramārtha’s original doctrine may indeed have been lost, it reminds us that we must also be aware of the possibility that we have an incomplete picture when we attempt to study and characterise Paramārtha’s thought more generally.

However, the fact remains that later authors only received a very vague and pareddown version of Paramārtha’s doctrine. Subsequent authors then often took the concept as raw material for their own constructive projects, or, in the interests of attacking or defending the notion, wove it into complex new networks of proof texts and various concepts. The result, as we have seen, is that the bulk of what was said about *amalavijñāna by later authors was new. We have little grounds for confidence that these authors were well acquainted with any works by Paramārtha, upon which they based their comments.

Despite some excellent studies, modern scholarship has still tended to accept too readily the image of *amalavijñāna found in the later tradition, rather than to examine closely what Paramārtha’s texts had to say about it. I hope that this study has shown that those sources reveals a surprising profi le of Paramārtha’s genuine attested doctrine of *amalavijñāna.

To summarise, the major fi ndings of this study were: 1) It is reasonable to think that Paramārtha coined the term amoluoshi at least in part on the basis of the term amalavijñāna in AK 5.29. 2) Paramārtha’s extant works contain not one but two separable doctrines of *amalavijñāna: one in JDZL, and the other in ZSL, SWXL and SBKL. The JDZL doctrine is most likely earlier, and more likely to be authentic (though the other version of the doctrine may also be authentic). 3) The rich details of these original doctrines have been insufficiently known in modern scholarship. They were also almost unknown to later authors in the tradition. 4) Later authors propose a riot of extremely varied ideas about *amalavijñāna and ninth consciousness, little of it traceable with any confidence to Paramārtha or his group. 6. Directions for future research

If we have tended to overlook the original content of Paramārtha’s own *amalavijñāna doctrine, that implies at least three agendas for further research.

First, it is important to look for the sources of Paramārtha’s attested *amalavijñāna doctrine.

Second, what has proven true for *amalavijñāna may prove true of Paramārtha’s thought more generally. Paramārtha’s actual ideas may have been buried under what was made of them by his successors ― enthusiasts as much as enemies. Those ideas may therefore constitute a missed chapter the development of East Asian Buddhism. We may need to bracket out what we think we “know”, from the image of Paramārtha constructed by the later tradition, and study Paramārtha’s own texts more carefully.

Finally, Paramārtha’s ideas are signifi cant in part because of the place they hold in our usual narratives of the so-called “sinifi cation” of Buddhist concepts. If the general image of *amalavijñāna has been inaccurate to date, then part of the general understanding of the process of sinifi cation may have been built on sand. It will also be important, therefore, to reassess the place of *amalavijñāna in relation to the problem of sinifi cation, in light of a more accurate picture of *amalavijñāna and its actual sources. Abbreviations AKBh Abhidharmakośabhāṣya BBh Bodhisattvabhūmi Bh Bhāṣya Ch. Chinese D Derge D Dīgha nikāya DBZ Dai Nippon Bukkyō zensho DN Dīgha nikāya FXL Foxing lun 佛性論 T1610 IBK Indogaku bukkyōgaku kenkyū インド学仏教学研究 It Itivuttaka JDZL Jueding zang lun 決定藏論 T1584 LAS Laṃkâvatāra-sūtra M Majjhima nikāya MAV Madhyântavibhāga MAVT Madhyāntavibhāga-ṭīkā MPNS Mahāparinirvāṇa sūtra MSA Mahāyānasūtrâlaṃkāra MSg Mahāyānasaṃgraha MSgBh Mahāyānasaṃgrahabhāṣya P Paramārtha PTS Pāli Text Society RGV Ratnagotravibhāga S Saṃyutta nikāya SBKL Shiba kong lun 十八空論 T1616 SdhN Saṃdhinirmocana sūtra Skt. Sanskrit Sth Sthiramati SWXL San wuxing lun 三無性論 T1617 T Taishō shinshū daizōkyō 大正新脩大藏經 (CBETA version) Tib. Tibetan TrBh Triṃśikabhāṣya VP LA VALLÉE POUSSIN, trans., Abhidharmakośabhāṣya WXL Wuxiang lun 無相論 X Shinsan dai Nippon zokuzōkyō 卍新纂大日本續藏經 (CBETA version) XSL Xianshi lun 顯識論 T1618 XYSJL Xianyang sheng jiao lun 顯揚聖教論 T1602 XZ Xuanzang 玄奘 YBh Yogācārabhūmi-śāstra ZSL Zhuanshi lun 轉識論 T1587 Bibliography

Note: In this bibliography, I have followed Chicago History style. However, at Zinbun’s request, in the footnotes I have adopted name-date citation to save space. To ease cross-reference, I therefore list multiple items by a single author in chronological order. Where I am not using the fi rst edition of an item, I have tried to give the original date of publication in square brackets, followed by the date of the edition or printing I used, thus: [1984] 1998. Citations in footnotes refer to the date of the version I used. AHN, J. “Taehyŏn.” Digital Dictionary of Buddhism, http://buddhism-dict.net/cgi-bin/ xpr-ddb.pl?59.xml+id(‘b5927-8ce2’). Accessed October 3 2008. ALBAHARI, Miri. “Against No-Ātman Theories of Annattā.” Asian Philosophy 12, no. 1 (2002): 5-20. ANACKER, Stefan. Seven Works of Vasubandhu, the Buddhist Psychological Doctor. Delhi: Motilal Banarsidass, [1984] 1998. BODHI, Bhikkhu, trans. The Connected Discourses of the Buddha: A Translation of the Saṃyutta Nikāya. Somerville MA: Wisdom Publications, 2000. BRAARVIG, Jens. Akṣayamatinirdeśasūtra. Vol I: Edition of Extant Manuscripts with an Index; Vol II: The Tradition of Imperishability in Buddhist Thought. Oslo: Solum Forlag, 1993. BUSWELL, Robert E. Jr. The Formation of Ch’an Ideology in China and Korea: The Vajrasamādhi-Sūtra, a Buddhist Apocryphon, Princeton Library of Asian Translations. Princeton: Princeton University Press, 1995.

trans. Cultivating Original Enlightenment: Wŏnhyo’s Exposition of the Vajrasamādhi-sūtra (Kŭmgang Sammaegyŏng Non). The International Association of Wŏnhyo Studies’ Collected Works of Wŏnhyo, Volume 1. Honolulu: University of Hawai’i Press, 2007. CARPENTER, J. Estlin, ed. The Dīgha-nikāya vol. III. Lancaster: Pali Text Society, 2006. CHEN, Jinhua. Monks and Monarchs, Kinship and Kingship: Tanqian in Sui Buddhism and Politics. Kyoto: Scuola Italiana di Studi sull’Asia Orientale, 2002. CHEN Yinque 陳寅恪. “Liang yi Dasheng qi xin lun wei Zhikai xu zhong zhi zhen shiliao 梁譯大乘起信論偽智愷序中之真史料 [1948].” In Chen Yinque wenji zhi san: Jinming guan cong gao er bian 陳寅恪文集之三　金明館叢稿二編, 147-152. Shanghai: Shanghai guji chubanshe, 1980. CH’OE Yŏnsik 崔鈆植. “Daijō shi ron gen gi ki to Kudara bukkyō『大乗四論玄義』と百済仏教.” Komazawa daigaku Bukkyō gakubu ronshū 39 (2008): 13-28. Chinese Buddhist Electronic Text Association. Shinsan dai Nippon zoku zōkyō 卍新纂大日本續藏經. Tokyo: Kokusho kankōkai, 1975-1989. CBReader v. 3.7, 2008.

Taishō shinshū daizōkyō 大正新脩大藏經, ed. Takakusu Junjirō 高楠順次郎 and Watanabe Kaigyoku 渡邊海旭. Tokyo: Taishō shinshū daizōkyō kankōkai/ Daizō shuppan, 1924-1932. CBReader v. 3.7, 2008. DEMIÉVILLE, Paul. “Sur l’authenticité du Ta tch’eng k’i sin luoen.”Bulletin de la maison français-japonais 2, no. 2 (1929): 1-78. Reprinted in Choix d’études bouddhiques 1929-1970, 1-79.

Le concile de Lhasa: une controverse sur le quiétisme entre bouddhistes de l’Inde et de la Chine au VIII. siècle de l’ère chrétienne. Paris: Impr. nationale de France, 1952. EDGERTON, Franklin. Buddhist Hybrid Sanskrit Grammar and Dictionary. 2 vols [1953]. Delhi: Motilal Banarsidass, 1998. FRAUWALLNER, Erich. “Amalavijñānam und Ālayavijñānam.” In Beiträge zur indischen Philologie und Altertumskunde, Walter Schubring zum 70. Geburtstag dargebracht, 148-159. Hamburg 1951. “Amarashiki to araiyashiki 阿摩羅識と阿頼耶識.” Translated by Kumoi Shōzen 雲井照善. Ōtani gakuhō 大谷学報 32, no. 2 (1952): 54-71. FRIEDMANN, David Lasar. Sthiramati Madhyāntavibhāgaṭīkā: Analysis of the Middle Path and the Extremes [Utrecht: 1937]. Oregon: Canon Publications Photographic Reprint, 1984. FUKAURA Masafumi 深浦正文. Yuishikigaku kenkyū 唯識学研究. 2 vols. Tokyo: Nagata Bunshodō, 1954, 1976. FUNAYAMA Tōru 船山徹. “Ryō no Sōyū sen Satsubata shi shiden to Tōdai bukkyō 梁の僧祐撰『薩婆多師資伝』と唐代仏教.” In Tōdai no shūkyō 唐代の宗教, ed. YOSHIKAWA Tadao 吉川忠夫, 325-354. Kyoto: Hōyū shoten, 2000.

“‘Kanyaku’ to ‘Chūgoku senjutsu’ no aida ― Kanbun butten ni tokuyū na keitai wo megutte「漢訳」と「中国撰述」との間 ― 漢文仏典に特有な形態をめぐって .” Bukkyō shigaku kenkyū 仏教史学研究 45, no. 1 (2002): 11-28.

“Shintai sanzō no chosaku no tokuchō ― Chū-In bunka kōshō no rei toshite 真諦 三蔵の著作の特徴 ― 中印文化交渉の例として.” Kansai daigaku tōsaigakujutsu kenkyūjo kiyō 関西大学東西学術研究所紀要 38 (2005): 97-122. “The Work of Paramārtha: An Example of Sino-Indian Cross-cultural Exchange.” Translated by Benjamin Brose. Forthcoming in the proceedings volume of an April 2007 Vienna conference panel “Early Chinese Buddhist Texts”. GIMELLO, Robert M. “Chih-yen (602-668) and the Foundations of Hua-yen Buddhism.” Ph.D. dissertation, Columbia University, 1976. HAHN, Michael. Nāgārjuna’s Ratnāvalī. Bonn: Indica Et Tibetica Verlag, 1982. HAKAMAYA Noriaki 袴谷憲昭. “Viniścayasaṃgrahaṇī ni okeru ārayashiki no kitei Viniścayasaṃgrahaṇī におけるアーラヤ識の規定.” Tōyō bunka kenkyūjo kiyō 79 (1979): 1-79. HAKEDA, Yoshito S. The Awakening of Faith, Attributed to Aśvaghosha. New York: Columbia University Press, 1967. HARVEY, Peter. “Consciousness Mysticism in the Discourses of the Buddha.” In The Yogi and the Mystic, ed. Karel Werner, 82-102. London: Curzon, 1989. HIRAI, Shun’ei 平井俊英. Chūgoku hannya shisōshi kenkyū: Kichizō to Sanron gakuha 中国般若思想史研究 ― 吉蔵と三論学派. Tokyo: Shunjusha, 1976. HIRAKAWA, Akira. Index to the Abhidharmakośabhāṣya. Tokyo: Daizō shuppan, 1973. Huimin (Keibin), Bhikkhu 釋恵敏. Shōmon ji ni okeru shoen no kenkyū『声聞地』における所縁の研究. Tokyo: Sankibō busshorin, 1994. ITŌ Takatoshi 伊藤隆寿. “Miroku kyō yūi no gimonten 『弥勒経遊意』の疑問点.” Komazawa daigaku Bukkyō gakubu ronshū 4 (1973): 59-75.

“Hōshōin zō Miroku jōge kyō yūi jūjū ni tsuite 宝生院蔵『弥勒上下経遊意十重』について.” IBK 25, no. 2 (1977): 323-326. IWATA Ryōzō 岩田良三. “Shintai no amarashiki ni tsuite 真諦の阿摩羅識について.” Ōsaki gakuhō 大崎学報 122 (1967): 176-177.

“Amala-jñāna to amarashiki setsu ni tsuite / Amala-jñānaと阿摩羅識について.” IBK 19, no. 2 (1971): 136-137.

“Shintai no amarashiki setsu ni tsuite 真諦の阿摩羅識説について.” Suzuki gakujutsu zaidan nenpō 鈴木学術財団年報 8 (1972[a]): 46-56.

“Shintai no sanshō setsu ni tsuite 真諦の三性説について.” IBK 21, no. 1 (1972[b]): 355-58.

“Shō daijō ron to kushiki setsu ni tsuite 摂大乗論と九識説について.” IBK 20, no. 2 (1972[c]): 302-305.

“Shintai no sanshō setsu ni tsuite (zoku) 真諦の三性説について(続).” IBK 22, no. 1 (1973): 107-110.

“Shintai no sanshō sanmushō setsu ni tsuite 真諦の三性三無性説について.” Suzuki gakujutsu zaidan nenpō 鈴木学術財団年報 10 (1974): 26-43. IWATA Taijō 岩田諦静. Shoki yuishiki shisō kenkyū 初期唯識思想研究. Tokyo: Daitō shuppansha, 1981.

“Seshin zō Shō daijō ron no kan’yaku keitai ni tsuite 世親造『摂大乗論釈』の漢訳形態について.” IBK 33, no. 2 (1985[a]): 73-78.

“Shintai yaku Shō daijō ron Seshin shaku ni okeru hen’i no yakugo ni tsuite 真諦訳『摂大乗論世親釈』における変異の訳語について.” Ōsaki gakuhō 大崎学報 140 (1985[b]): 17-34.

“Shintai yaku no Shō daijō ron Seshin shaku ni okeru sanshō setsu ni tsuite 真諦 訳の『摂大乗論世親釈』における三性説について.” In two parts. 上, Hokke bunka kenkyū 法華文化研究 13 (1987): 25-48; 下 Hokke bunka kenkyū 法華文化研究 15 (1989): 27-52.

“Shintai yaku Shō daijō ron Seshin shaku ni okeru ‘Shi kai mushiji’ ge to saiseijō hōkai ni tsuite 真諦訳『摂大乗論世親釈』における此界無始時偈と最清浄法界について.” In Suguro Shinjō hakushi kokei kinen ronbunshū 呂信静博士古稀記念論文集, ed. Suguro Shinjō hakushi kokei kinen ronbunshū kankōkai 勝呂信静博士古稀記念論文集刊行会, 117-134. Tokyo: Sankibo busshorin, 1996.

“Shintai yaku Shō daijō ron Seshin shaku ni okeru ariyashiki setsu ni tsuite 真諦訳『摂大乗論世親釈』における阿黎耶識説について.” IBK 45, no. 2 (1997): 203-209.

“Shintai yaku Shō daijō ron Seshin shaku ni okeru kōzō bubun no kentō ― ‘Shaku eshi shō sō hon Shochie shō’ 真諦訳『摂大乗論世親釈』における増広部分の検討　釈依止勝相品所知依章.” Minobu ronsō 身延論叢 5 (2000): 97-118.

“Hokke gen gi ni okeru anmarashiki setsu ni tsuite ― Shintai yaku Shō daijō ron Seshin shaku to no kanren 『法華玄義』における菴摩羅識説について　真諦訳『摂大乗論世親釈』との関連.” Hokke kyō kenkyū 法華経研究 13 (2001[a]): 371-392.

“Shintai no yuishiki setsu no tokushoku ni tsuite 真諦の唯識説の特色について.” IBK 50, no. 1 (2001[b]): 173-180(L).

“Shintai yaku Shō daijō ron Seshin shaku ni okeru anmarashiki setsu ni tsuite 真諦訳『摂大乗論世親釈』における菴摩羅識説について.” IBK 49, no. 1 (2001[c]): 58-64.

“Shintai yaku Shō daijō ron Seshin shaku ni okeru shūji setsu ni tsuite 真諦訳

『摂大乗論世親釈』における種子説について.” In Taga Ryūgen hakushi kokei kinen ronshū: Bukkyō shisō bukkyōshi ronshū 田賀龍彦博士古稀記念論集 ― 仏教思想仏教史論集, ed.Taga Ryūgen hakushi kokei kinen ronshū kankōkai 田賀龍彦博士古稀記念論集刊行会 , 107-124. Tokyo: Sankibō busshorin, 2001(d).

“Shintai yaku Shō daijō ron Seshin shaku ni okeru dōgo no shinnyo ni tsuite 真諦訳『摂大乗論世親釈』における道後の真如について.” In Sasaki Kōken hakushi kokei kinen bukkyōgaku bukkyōshi ronshū 佐々木孝憲博士古稀記念仏教学仏教史論 集, ed. Sasaki Kōken hakushi kokei kinen bukkyōgaku bukkyōshi ronshū kankōkai 佐々木孝憲博士古稀記念仏教学仏教史論集刊行会, 3 - 1 8 . T o k y o : S a n k i b ō busshorin, 2002.

Shintai no yuishiki setsu no kenkyū 真諦の唯識説の研究. Kyoto: Sankibō bus shorin 山喜房佛書林, 2004. JAMPSAL, L., R. CLARK, J. WILSON, L. ZWILLING, M. SWEET and R. THURMAN, trans. The Universal Vehicle Discourse Literature (Mahāyānasutrālaṃkāra). Tanjur Translation Initiative Treasury of the Buddhist Sciences series. New York: American Institute of Buddhist Studies, Center for Buddhist Studies, Tibet House US, 2004. JOHNSTON E. H., ed. The Ratnagotravibhāga Mahāyānottaratantraśāstra [1950]. In The Uttaratantra of Maitreya, ed. H. S. Prasad. Bibliotheca Indo-Buddhica Series No. 79. Delhi: Sri Satguru Publications, 1991. JONES, J. J., trans. The Mahāvastu. 3 vols. London: Pali Text Society and Routledge & Kegan Paul, 1949-1956, 1978, 1987. KANNO Hiroshi 菅野博史 . “Ekin Daijō shi ron gen gi ki no sanshu shakugi to Kichizō no yonshu shakugi 慧均『大乗四論玄義記』の三種釈義と吉蔵の四種釈義.” In Kimu- ra Kiyotaka hakushi kanreki kinen ronshū: Higashi Ajia bukkyō ― sono seiritsu to tenkai 木村清孝博士還暦記念論集　東アジア仏教 ― その成立と展開, 87-100. Tokyo: Shunjusha, 2002.

“Daijō shi ron gen gi ki no kisoteki kenkyū『大乗四論玄義記』の基礎的研究 .” Unpublished conference paper delivered at the 59th annual conference of the Japanese Association of Indian and Buddhist Studies 印度学仏教学会, Aichi Gakuin Daigaku, September 2008. KATSUMATA Shunkyō 勝又俊教. Bukkyō ni okeru shin-i-shiki setsu no kenkyū 仏教における心意識説の研究 [A Study of the Citta-Vijñāna Thought in Buddhism]. Tokyo: Sankibō Busshorin, 1961. KENG, Ching. PhD dissertation, Harvard University. Forthcoming, 2009. LAI, Whalen. “The Meaning of ‘Mind-Only’ (wei-hsin): An Analysis of a Sinitic Mahāyāna Phenomenon.” Philosophy East and West 27, no. 1 (1977): 65-83. LAMOTTE, Étienne, ed. and trans. Saṃdhinirmocana Sūtra: Explication des mystères. Louvain: Bureaux de Receuil, 1935.

La somme du grand véhicule d’Asaṇga (Mahāyānasaṃgraha). Louvain-la-Neuve: Institut Orientaliste, Université de Louvain, 1973. LA VALLÉE POUSSIN, Louis de. Bouddhisme, études et matériaux: Théorie des douze causes. Gand [Ghent]: Librarie Scientifi que E. van Goethem, 1913.

Vijñaptimātratāsiddhi: La siddhi de Hiuan-Tsang. 2 vols. Paris: P. Geunther, 1928-29.

L’Abhidharmakośa de Vasubandhu [1923-1926]. Brussels: Institut belge des hautes études chinoises, 1980. LÉVI, Sylvain, ed. Mahāyāna-sūtrālaṃkāra. 2 vols. Paris: Librairie Honoré Champion, 1907, 1911.

Vijñaptimātratāsiddhi: Deux traités de Vasubandhu: Viṃśatikā (La vingtaine) accompagnée d’une explication en prose, et Triṃśikā (La trentaine) avec la commentaire de Sthiramati. Paris: H. Champion, 1925.

Matériaux pour l’étude du système vijñaptimātra, un système de philosophie bouddhique. Paris: H. Champion, 1932. LIEBENTHAL, Walter. “Notes on the Vajrasamādhi.”T’oung Pao 44 (1956): 347-386. LINDTNER, Christian. “From Brahmanism to Buddhism.” Asian Philosophy 9, no. 1 (1999): 5-37. LÜ Cheng 呂澂. Lü Cheng foxue lunzhu xuanji 呂澂佛學論著選集. Jinan: Jilu chubanshe, 1991. LUSTHAUS, Dan. Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Philosophy and the Ch’eng Wei-shih Lun. London: RoutledgeCurzon, 2002. MASEFIELD, Peter, trans. The Itivuttaka. Oxford: Pali Text Society, 2001. MATSUDA Kazunobu. “Darai Rama 13 sei kizō no ichiren no nepāru kei shahon ni tsuite: Yuga ron ‘Shō kettaku bun’ bonbun danpen hakken ki ichi ダライラマ13世寄贈の一連のネパール系写本について ―『瑜伽論』摂決択分梵文断簡発見記 ―.” Nihon Chibetto gakkai kaihō 日本チベット学会会報 34 (1988): 16-20. MITSUGIRI Jikai 三桐慈海. “Ekin no sanron gaku 慧均の三論学.” In Sanron kyōgaku no kenkyū 三論教学の研究, ed. HIRAI Shun’ei 平井俊栄, 223-236. Tokyo: Shunjusha, 1990. MOU Zongsan 牟宗三. Foxing yu bore 佛性與般若 [1977]. Taipei: Xuesheng shuju 學生書局, 1997. MULLER, A. Charles, ed. Digital Dictionary of Buddhism. <http://buddhism-dict.net/ ddb>. Edition of 30/09/2008. NAGAO, Gadjin M. Madhyāntavibhāga-bhāṣya: A Buddhist Philosophical Treatise Edited for the First Time from a Sanskrit Manuscript. Tokyo: Suzuki Research Foundation, 1964. ÑĀṆAMOLI, Bhikkhu, and Bhikkhu BODHI, trans. The Middle Length Discourses of the Buddha: A New Translation of the Majjhima Nikāya. Boston: Wisdom Publications, 1995. NANJIO, Bunyiu, ed. The Laṇkāvatāra Sūtra. Kyoto: Otani University Press, 1923. O’BRIEN, Paul Wilfred. “A Chapter on Reality from the Madhyantivibhagacastra.” Monumenta Nipponica 9, no. 1-2 (1953): 277-303. ŌCHŌ E’nichi 横超慧日. “Shinshutsu shiryō: Shi ron gen gi no ‘Sho shō chū ka gi’ 新出資料・四論玄義の初章中假義.” IBK 7, no. 1 (1958): 131-134. OGIHARA, Unrai 荻原雲来. Kanyaku taishō bonwa daijiten漢訳対照梵和大辞典 [1964-1974]. Tokyo: Kōdansha, 1986. [See also WOGIHARA.] ŌTAKE, Susumu 大竹晋. Jūji kyō ron 十地経論. Shin kokuyaku issaikyō 新国訳一切経, Shaku kyōron bu 釈経論部 16. Tokyo: Daizō shuppan, 2005.

Yuishiki setsu wo chūshin to shita shoki Kegon kyōgaku no kenkyū: Chigen, Gishō kara Hōzō e 唯識説を中心とした初期華厳教学の研究　智嚴・義湘から法蔵へ. Tokyo: Daizō-shuppan, 2007(a).

“Shintai ‘Kushiki giki’ itsubun 真諦『九識義記』佚文.” Unpublished paper circulated in the seminar “Shintai sanzō to sono jidai 真諦三蔵とその時代,” Kyōto University, 2007 (b).

“‘Kushiki shō’ kanjin senjutsu setsu ni taisuru jakkan no gimon (Chūgoku hen) 『九識章』漢人撰述説に對する若干の疑問（中国篇）.” Unpublished paper circulated in the seminar “Shintai sanzō to sono jidai 真諦三蔵とその時代,” Kyōto University, 2007(c).

“‘Kushiki shō’ kanjin senjutsu setsu ni taisuru jakkan no gimon (Indo hen)『九識章』漢人撰述説に對する若干の疑問（インド篇）.” Unpublished paper circulated in the seminar “Shintai sanzō to sono jidai 真諦三蔵とその時代,” Kyōto University, 2007(d). PAUL, Diana Y. Philosophy of Mind in Sixth-Century China: Paramārtha’s ‘Evolution of Consciousness’. Stanford: Stanford University Press, 1984. PRADHAN, Prahlad, ed. Abhidharmakośabhāṣya of Vasubandhu. Patna: K. P. Jayaswal Research Institute, 1967. RADICH, Michael. ” The Somatics of Liberation: Ideas about Embodiment in Buddhism from Its Origins to the Fifth Century C.E.” PhD dissertation, Harvard University, Nov. 2007. RAHDER, Johannes, ed. Daśabhūmikasūtra. Leuven: J.-B. Istas, 1926. RUEGG, David Seyfort. La théorie du Tathāgatagarbha et du Gotra. Paris: École française d’Extrême-Orient, 1969. SAKUMA, Hidenori. Die Āśrayaparivṛtti-theorie in der Yogācārabhūmi. 2 vols. Alt- und NeuIndische Studien 40. Stuttgart: Franz Steiner Verlag, 1990.

佐久間秀範. “Yugashi-ji ron ni okeru tenne shisō『瑜伽師地論』における転依思 想.” IBK 39, no. 1 (1991): 66-74. SATO Testuei 佐藤哲英. Tendai daishi no kenkyū: Chigi no chosaku ni kansuru kisoteki kenkyū 天台大師の研究　智顗の著作に関する基礎的研究. Kyoto: Hyakkaen, 1961. SCHMITHAUSEN, Lambert. Der Nirvāṇa-Abschitt in der Viniścayasaṃgrahaṇī der Yogācārabhūmi. Wien: Hermann Böhlaus Machf., 1969.

Ālayavijñāna: On the Origin and Development of a Central Concept in Yogācāra Philosophy. 2 vols. Studia Philologica Buddhica Monograph Series IVb. Tokyo: The International Institute for Buddhist Studies, 1987. Shengkai 圣凯. Shelun xuepai yanjiu 摄论学派研究. Beijing: Zongjiao wenhua chubanshe, 2006.

“Amoluoshi yu benjue: Shelun shi de ‘amoluoshi’ sixiang 阿摩罗识与本觉 ― 摄论师的 ‘阿摩罗识’ 思想.” Dunhuangxue jikan 敦煌学辑刊 2007, no. 3: 106-118. STACHE-ROSEN, Valentina. Dogmatische Begriffsreihen im älteren Buddhismus II: Das Saṇgītisūtra und sein Kommentar Saṇgītiparyāya. 2 vols. Berlin: Akademie-Verlag, 1968. STANLEY, Richard. “A Study of the Madhyāntavibhāga-bhāṣya-ṭīkā.” PhD dissertation, Australian National University, 1988. STCHERBATSKY, Th., trans. Madhyānta-Vibhanga: Discourse on Discrimination Between Middle and Extremes, Ascribed to Bodhisattva Maitreya and Commented by Vasubandhu and Sthiramati. [[[Bibliotheca Buddhica]] vol. 30, 1936;] New Delhi: Oriental Books Reprint Corporation, 1978. SU, Gongwang 蘇公望. “Zhendi liuhua nianpu 真諦留華年譜.” Originally published in instalments in Weimiao sheng 微妙聲, 1936-1937. Republished as “Zhendi sanzang nianpu 眞諦三藏年譜.” In Xiandai fojiao xueshu congkan 現代佛教學術叢刊 13, Han Wei liang Jin Nanbeichao pian (xia) 漢魏両晋南北朝篇(下), ed. ZHANG Mantao 張曼濤. Taibei: Dasheng wenhua chubanshe, 1978. SUZUKI, Daisetz Teitaro, trans. The Lankāvatāra Sūtra: A Mahāyāna Text [1932]. Delhi: Motilal Banarsidass, 1999.

An Index to the Lankavatara Sutra [1934]. New Delhi: Munshiram Manoharlal, 2000. TAKAKUSU Junjirō 高楠順次郎 and MOCHIZUKI Shinkō 望月信亨, eds. Dai Nippon Bukkyō zensho 大日本佛教全書. Tokyo: Yūseidō shuppanbu 有精堂出版部, 1934. TAKASAKI, Jikidō. A Study on the Ratnagotravibhāga (Uttaratantra), Being a Treatise on the Tathāgatagarbha Theory of Mahāyāna Buddhism. Rome: Serie Orientale Roma, 1966. THANISSARO Bhikkhu. The Mind Like Fire Unbound. Barre, MA: Dhamma Dana Publications, Barre Centre for Buddhist Studies, 1993. Tōkyō daigaku bungakubu Indo tetsugaku Indo bungaku kenkyūshitsu 東京大学文学部インド哲学インド文学研究室, ed. Deruge-ban Chibetto daizōkyō ronsho bu: Tōkyō daigaku bungakubu shozō: Yuishiki bu sems tsam デルゲ版チベット大蔵経論疏部東京大学文学部所蔵：唯識部 SEMS TSAM, sDe dge Tibetan Tripiṭaka bStan hgyur Preserved at the Faculty of Letters, University of Tokyo. Tokyo: Sekai seiten kankō kyōkai, 1980. UI, Hakuju. Indo tetsugaku kenkyū 印度哲学研究 . Tokyo: Kōshisha Shobō 甲子社書房 , 1926-1932. Vol VI. WALDRON, William S. The Buddhist Unconscious: The Ālaya-vijñāna in the Context of Indian Buddhist Thought. London: RoutledgeCurzon, 2003. WALSHE, Maurice, trans. The Long Discourses of the Buddha: A Translation of the Dīgha Nikāya. Boston: Wisdom Publications, 1995. WAYMAN, Alex. “Notes on the Three Myrobalans.” Phi Theta Annual 5 (1954/55): 63-77. WAYMAN, Alex, and Hideko WAYMAN, trans. The Lion’s Roar of Queen Śrīmālā: A Buddhist Scripture on the Tathāgatagarbha Theory. New York: Columbia University Press, 1974. WIJESEKERA, O. H. de A. “Vedic Gandharva and Pāli Gandhabba.” University of Ceylon Review 3, no. 1 (1945): 73-107.

“The Concept of Viññāṇa in Theravāda Buddhism.” Journal of the American Oriental Society 84, no. 3 (1964): 254-259. WINDISCH, Ernst, ed. Iti-vuttaka. London: Pali Text Society, 1975. WOGIHARA, Unrai, ed. Bodhisattvabhūmi: A Statement of Whole Course of the Bodhisattva, Being the Fifteenth Section of Yogācārabhūmi. Tokyo: Sankibo, 1971. [See also OGIHARA.] YAMAGUCHI, Susumu 山口益, ed. Madhyāntavibhāgaṭīkā: Exposition Systématique du Yogācāravijñaptivāda 安慧阿遮梨耶造中邊分別論釈疏. Nagoya: Librairie Hajinkaku 破塵閣, 1934. YE, Ayue (Yō Agetsu) 葉阿月. Yuishiki shisō no kenkyū ― konpon shinjitsu toshite no sanshō setsu wo chūshin ni shite 唯識思想の研究 ― 根本真実としての三性説を中心にして. Tainan: Gaochang yinshuju, 1975. Yinshun 印順 . “Lun Zhendi sanzang suochuan de amoluoshi 論真諦三藏所傳的阿摩羅 識.” In Yi fofa yanjiu fofa 以佛法研究佛法, Miao yin ji 妙音記 vol. 3, part 4, 269-300. Taibei: Zhengwen chuban, 1972. YOKOYAMA, Koitsu and Takayuki HIROSAWA, eds. Dictionary of Buddhist Terminology Based on Yogācārabhūmi : Sanscrit-Tibetan-Chinese & Tibetan-Sanscrit-Chinese / Bukkyōgo jiten : Yugashiji ron ni motozuku Bon-Zō-Kan taishō, Zō-Bon-Kan taishō. Tokyo: Sankibō Busshorin, 1997. YOSHIMURA, Makoto 吉村誠. “Shōron gakuha no shinshiki setsu ni tsuite 摂論学派の心識説について.” Komazawa daigaku Bukkyō gakubu ronshū 駒澤大学仏教学部論集 34 (2002): 223-241.

“Shintai no amarashiki setsu to Shōron gakuha no kushiki setsu 真諦の阿摩羅識説と摂論学派の九識説.” IBK 56, no. 1 (2007[a]): 177-183.

“Den Shintai ‘Kushiki shō’ itsubun 傳真諦『九識章』佚文.” Unpublished materials circulated in the seminar “Shintai sanzō to sono jidai”, 2007(b).

“Tendai bunken ni mirareru Jiron, Shōron gakuha no shinshiki setsu ― Chigi to Tanzen no chosaku wo chūshin ni 天台文献に見られる地論・摂論学派の心識説 ― 智顗と湛然の著作を中心に.” Forthcoming, IBK 57. YŪKI, Reimon 結城零聞. “Shina yuishikigaku shijō ni okeru Ryōgashi no chii 支那唯識学史上における楞伽師の地位.” Shina bukkyō shigaku 支那仏教史学 1, no. 1 (1937): 21-44. ZIMMERMAN, Michael. A Buddha Within: The Tathāgatagarbhasūtra, The Earliest Exposition of the Buddha-Nature Teaching in India. Bibliotheca Philologica et Philosophica Buddhica VI. Tokyo: The International Research Association for Advanced Buddhology, Soka University, 2002.