New History Of Tantric Lieterature In India

Author(s): J W De Jong
Publisher: J W De Jong
Catalog link: https://jainqq.org/explore/269595/1

Page #1

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA*
J. W. de Jong
Matsunaga Yūkei (1929- ) belongs to the younger generation of Tantric scholars in Japan. He has published many articles and several books, including a brief history of Tantrism (Mikkyo no rekishi, Kyoto, Heirakuji shoten, 1969) and a critical edition of the Sanskrit text of the Guh yasa mājatantra (Osaka, Tōho Shuppan Inc., 1978). Whereas in the past Japanese scholars studied only the Chinese translations of Tantric texts, Matsunaga's research takes into account Sanskrit manuscripts and Tibetan translations. Moreover,he is well acquainted with the work of Western and Indian scholars. In his recent book, entitled 'Historical studies on the formation of Tantric texts', Matsunaga has published the results of his studies on the formation of important Indian Tantric texts and rituals. This book is an important contribution to Tantric studies and contains much information on the development of Tantrism in India. It is therefore perhaps not superfluous to summarise its contents in some detail and to point out some of the results obtained by Matsunaga.
Matsunaga's book is divided into an introduction and four chapters, each subdivided into several sections. Below we will indicate the Introduction by the letter A, the chapters by Roman numerals and the sections by Arabic ciphers. Al deals mainly with terminological matters and the divisions of Buddhist tantras. In Japan the opposition between Mikkyo and Kengyo (the esoteric and the exoteric doctrines) goes back to Kūkai (774-835). However, Kūkai's definition of Mikkyo was not accepted by other monks such as Ennin (794864), and in later times the Mikkyo of the Tendai school (Taimitsu) and the Mikkyo of the Shingon school (Tomitsu) were opposed to each other. The expressions Mikkyo and Kengyo are found in a text translated by Pu-k’ung (705-774), the Sung-shih t'o-loni i-tsan (T. 902), but the attribution of this translation to Pu-k’ung is doubtful and it is not at all certain that the terms Mikkyo and Keng yo are based upon Indian terms. M. [=Matsunaga] points out that in the West such terms as esoteric Buddhism, Buddhist Tantrism and Tantric Buddhism are used. He examines the meaning of the word tantra in Indian non-Buddhist texts and in Buddhist texts, especially the Guh yasamāja,
* Matsunaga Yakei, Mikkyo kyoten seiritsushi-ron. Kyöto, Hözökan, 1980. 330+31 pp.

Page #2

92
J. W. de Jong
The term Buddhist Tantrism, he concludes, indicates a Buddhist variant of an Indian religious and cultural phenomenon, whereas the term Mikkyo points to the secret elements in Buddhist doctrine. In Japan, Tantric studies are concentrated on the doctrines of the Shingon and Tendai schools, whereas, in the West, more attention has been paid to Tantrism as a religion in India and Tibet and Tantric texts in Sanskrit and Tibetan have been carefully studied. According to M. both approaches have their valid points and should be combined in the study of Indian Tantrism. Considered in the light of both Western Tantric studies and the Japanese tradition of Mikkyo studies, Mikkyo can be regarded as encompassing within the boundaries of Buddhism with bodhi as its main goal, all the secret doctrines, ceremdies, spells and symbols which are derived from Tantrism. It is in this sense that the word Mikkyo is used by the author in this book. Henceforth, we will take the liberty to use in its place the term Tantrism.
In Japan the terms junmitsu (pure Tantrism) and zomitsu (mixed Tantrism) were used first by Kūkai. According to his terminology zömitsu indicated the Tantric doctrines not included in the systems of the Dainichikyo and the Kongochokyo (T. 848 and 865). It is only much later that these two terms were used by Eko (1666-1734) to indicate an opposition implying a value judgment. In Shingon studies zāmitsu is used as a deprecatory term to indicate ritual and dhāraṇi texts which have as their chief goal worldly goods and benefits, whereas junmitsu designates Dainichikyo and the Kongochōkyō texts which deal mainly with the attainment of bodhi. M. distinguishes three periods in the history of Tantric Buddhism, the early period, the middle period (the seventh century) and the later period (from the eighth century onward). Most of the zomitsu texts are products of the early period; the Dainichikyo and Kongochokyo texts date from the middle period. The so-called Tantras of the left-hand arose in India in the later period and the texts of this school have not been transmitted to Japan, M. points out that von Glasenapp wrongly used the terms junmitsu and zomitsu as corresponding to a pure Tantrism of the right hand and a Tantrism mixed with sākta elements of the left hand (cf. Buddhistische Mysterien, p. 41).
In Western publications on Tantrism many names of yāna-s are mentioned, but without references to sources. M. points out that in Tibetan historical works and in commentaries written by Tibetan scholars, Tantrism is designated by the name Snag-kyi theg-pa (Mantrayāna) in opposition to the Pha-rol-tu phyin-pa'i theg-pa (Pāramitāyāna). However, in the Guh yasamāja, the term Mantrayāna 'is not found, but only the term Mantranaya. In the Mahāvairocanasūtra the term mantracar yānaya is used. The eighth century commentator Buddhaguhya distinguishes a paramitānaya and a mantranaya. The same distinction is made in the eleventh and twelfth century by Advayavajra

Page #3

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
93
and Ratnākaraśānti who also make use of the term mantrayāna. Both mantrayana and mantranaya are found in the Sekoddeśaţika, a commentary of the Kalacakratantra. It is not possible to indicate exactly the meaning of these two words and how the term mantrayāna has arisen from the term mantranaya. However, when Advayavajra and Ratnākaraśānti use the term mantrayāna, it indicates a subdivision of Mahāyāna, the other subdivision being the Pāramitāyāna. Therefore, Tantrism is not, as is commonly assumed, a third yana opposed to both Hinayāna and Mahāyāna.
In a brief survey of the Tantric texts M. points out that most Chinese translations of Tantric texts date from the T'ang dynasty. In Tibet the first translations were made in the second half of the eighth century. Especially in the period of the later spread of the doctrine (Phyi dar) in the eleventh century and following centuries, a great many Tantric texts were translated into Tibetan. Many Tantric texts are only preserved in Tibetan translations, which have the advantage of being more accurate than Chinese translations.
In the thirteenth century Bu-ston divided Tantric texts into four groups : kriyā, caryā, yoga and anuttarayoga. This classification must have originated in India, for the eighth-century commentator Buddhaguhya had already divided Tantric texts into two groups, kriya and yoga, and subdivided the kriyatantra-s into general (spyi'i cho-ga) and special (bye-brag) tantras. In a work written in the tenth century, the Vajrajñanasamuccayatantra, a five-fold division is given : mahāyoga (rnal-'byor chen-po), dvaya (gñis-ka), carya, kriya and kalpa (rtog-pa). Several other systems of classification are mentioned by M., such as the seven-fold division found in Atiśa's Bodhimärgapradipapañjikā.
In A. 2 and A. 3 M. gives a short history of Tantric studies in Japan and Europe. In the final section of this chapter (A. 4) M. points out that Indian Tantrism can be stud. ied in two ways, by examining the literature and archaeological remains and by studying Tantrism as a living religion in Tibetan communities and in Japan. Both methods have their weak points and it is necessary to combine them in the study of Indian Tantrism.
Chapter 1 is entitled 'Ancient Indian culture and Tantrism'. In I. 1 M. points out that Buddhism has absorbed many elements of ancient Indian culture, such as spells and rituals which aim at obtaining worldly goals, although the attainment of bodhi and the reaching of Nirvāṇa were taught by the Buddha himself. In the early Tantric texts, worldly goals were more important than the pursuit of bodhi. However, in the middle period, the attainment of Buddhahood became the chief goal and many rituals were re-interpreted, sublimated and purified. On the other hand, abstract Buddhist Mahāyāna concepts such as that of the Void (śünya) and the Pure Light (prabhāsvara) came to be symbolised

Page #4

94
J. W. de Jong
by Tantric rituals. Many Hindu gods were incorporated into the Buddhist pantheon as represented in the two mandalas, the Garbhamandala and the Vajradhātumandala. The same process continued in the later period in which the anuttarayogatantra-s were composed.
I. 2 deals with the acceptance of evil in Tantrism. Many Tantric rituals stand in oppo. sition to ordinary ethics, which explains the hostile reaction to both Hindu and Buddhist Tantrism in recent times. However, Tantrism does not blindly accept evil. Its presence in Tantrism is due, on the one hand, to the fact that Tantra yoga has as its goal union with the Absolute, and, on the other, to the fact that many ancient religious elements have been preserved in Tantrism. The opposition to ordinary ethics iš especially strong in anuttarayoga. In the Guhyasamāja, a representative text of the 'father' tantras of the anuttara yoga, the yoga system contains six members, omitting from the eight members of the system of the Yogasūtra 'prohibition' (yama) and ‘restraint' (niyama): In the sådhana-s of the anuttarayoga it is said that for the yogin in possession of 'wis. dom' (prajña) and ‘means' (upaya) evil does not exist. The anuttarayoga texts declare that ordinary rituals are useless for obtaining the highest siddhi. Most conspicuous is the opposition to ordinary ethics in the attitude to sexual acts. M. points to the relation between agricultural activities and procreation in agricultural communities. Quoting Eliade, M. states that for primitive peoples the sexual act was a sacrament expressing holy union. Physiological functions were thought to correspond to cosmic activities. In the fifth and fourth centuries B. C. non-Brahmanical religions such as Buddhism and Jainism developed in mercantile centres, where there was less use for the ancient agricultural magic. However, with the revival of Hinduism during the Gupta period, Hindu customs and old Indian magical practices were revived and adopted by Mahāyāna Buddhism, wherein they were transformed by mystical explanations, and by a process of symbolisation and puri. fication. The union (yuganaddha) of prajña and upaya symbolised the union of the masculine and feminine principles. In Hindu tantras the five makāras were taught (madya, mamsa, matsya, mudra and maithuna), but they were explained as representing the five elements. The practice of Tantric rituals was dangerous and could not be undertaken without correct preparation and the supervision of a teacher.
Tantric texts taught that it was sometimes even necessary to kill, and commentators such as Amoghavajra explain that one may kill those who want to destroy the doctrine and to harm living beings, and that one must abolish one's own bad thoughts. On the other hand, killing could be an act of mercy with the intention of thereby saving sinners and leading them to deliverance. The three kleśa-s, passion (rāga), hate (dveşa) and delusion (moha), which were called the three poisons, came to be valued in a positive

Page #5

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
way in Tantrism, which taught that one ought to transcend passion by means of passion. M. points out that Avalon and de La Vallée Poussin were the first Western scholars to explain correctly the symbolic character of Hindu and Buddhist Tantric doctrines.
II. 1 deals mainly with the dharani, its original meaning and later developments. According to M. mantra, ju 'spell' and dharani have almost the same meaning in the middle and later periods of Tantrism, but mantra is by far the oldest of them and already found in the Atharvaveda as a spell to ward off evil and to obtain happiness. The Chinese translation of the word vidya, ming-chou (Japanese myōju), combines the two meanings of science and magic. The Prajñāpāramitā texts are considered to be a great magical science, mahavidya, by means of which one is not only able to ward off evil but also to obtain the supreme bodhi and the wisdom of the Buddha.
1
95
The word dharani is derived from the verb dhr-and is translated into Chinese and Tibetan by words meaning 'to hold'. This meaning is related to the yoga exercise dhāraṇā which originally meant the concentration of the mind on one point. In early Mahāyāna texts the word dhāraṇī is used to indicate the memorizing of texts. In early Buddhism the words saccakiriya and paritta are used in the meaning of 'spell', but the word dharani is not found. The Bodhisattvabhumi divides dharani-s in four groups : dharma, artha, mantra and bodhisattvakṣāntilabha. The mantradhāraṇī is a spell capable of appeasing many plagues. The Mahaprajñāpāramitopadeśa divides the dharani-s into three groups: śrutadhara, vibhajyajñāna and ghoṣapravesa. M. compares these three to the dharma-, artha- and bodhisattvakṣāntilabhadhāraṇī-s mentioned in the Bodhisattvabhumi, and supposes that in the third or fourth century the mantradharani-s were added to the three other groups.
M. points out that dharani in the meaning 'spell' is not found in the earliest Mahāyana texts but in additions made to these texts in the third century A. D. The word dharani has also assimilated the two meanings of the word vidya, i. e. science and magic. In several texts the dharani is considered to incorporate the wisdom of the Buddha and the doctrine of the Buddha. In Mahāyāna Buddhism the doctrine of the Buddha is believed to be capable of warding off evil. According to M., by memorizing a dhāraṇī (in the sense of the concentration of the mind in meditation) one obtains the wisdom of the Buddha. By the power of memorizing the dhārant which incorporates the doctrine of the Buddha, and by the power of the wisdom of the Buddha, one is able
1
Matsunaga does not refer to the materials collected by Lamotte in his study on dharani (Le traité de la grande vertu de sagesse, tome IV, Louvain, 1976, pp. 1854-1864). Many texts discussed by Matsunaga have been studied also by Lamotte.

Page #6

J. W. de Jong.
to conquer external enemies such as the kleśa-s and calamities. On the other hand, if a monk is threatened by robbers, poisonous snakes, etc., it is possible for him to ward off such dangers by entering into meditation.
In Mahāyāna texts dharani is used in two meanings: the memorization of texts and a magic spell. In the many versions of the Anantamukhadhāraṇī from the beginning of the third century to the later half of the eighth century dharant is always used to indicate the memorization of texts, but in other Mahāyāna texts this meaning gradually disappears and the dharani becomes more and more a magic spell possessing the power attributed to paritta-s in early Buddhism.
The overwhelming majority of the early Tantric dharant texts are all magical, but in the Tantric texts of the middle period the problem of attaining Buddhahood becomes of major importance. Of the five versions of a dharani text (T. 1137, 1138a, 1138b, 1139 and 1140) three were translated during the Eastern Chin (317-420 A. D.), one by Bodhiruci in the beginning of the T'ang dynasty, and one by Fa-hsien in the beginning of the Sung dynasty. All versions mention the twenty advantages of reciting dharani-s, but only Bodhiruci's version mentions among these advantages that of rapidly obtaining the supreme bodhi. In the middle of the seventh century Tantric texts were translated by five translators, Chih-t'ung, Ch'ieh-fan-to-mo, Atigupta, Punyodaya and Hsuan-tsang. In texts, translated by Chih-t'ung, Atigupta and Punyodaya many spells are found, but the problem of attaining Buddhahood also becomes important. Of the ten texts translated by Hsüan-tsang, only one (T. 1034) does not make any mention of the attainment of Buddhahood, but the other nine texts repeatedly stress that this is one of the advantages derived from reciting dharant-s. However, in other versions of several of these texts, this is not mentioned at all, and it is therefore quite possible that Hsüang-tsang, who was not greatly interested in Tantrism, made alterations in the Tantric texts translated by him.
96
In the first section of II. 2 M. studies the development of Tantric ritual on the basis of Chinese translations of Tantric texts. Statues of Buddhas are already mentioned in a text translated at the end of the second century A. D. (T. 417), and the adoration of statues must have been practised in the third century. More detailed descriptions are given in texts translated in the fourth century and the beginning of the fifth century. Several texts translated in the same period describe the veneration of the seven Buddhas, the eight bodhisattvas and of Avalokitesvara. The simabandha (kekkai) and the homa (goma) rituals were probably practised by the beginning of the fourth century. The abhişeka (kanjo) is described in a text (T. 1331) the translation of which is attributed to Śrīmitra, who translated several texts in the period 317-322. M. remarks that the

Page #7

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
authenticity of this work has been called into question but that the passage describing the abhiseka is certainly of Indian origin. The meditation on Buddhas (buddhanusmộti, kanbutsu) is first mentioned in a text translated in the beginning of the fourth century (T. 643). In a text translated in the first half of the sixth century (T. 1007), one can find the early stage of the mandala, and descriptions of Buddhas and bodhisattvas with many arms. In it mudra-s are mentioned for the first time. A text translated by Chih-t'ung about 650 A. D. (T 1103b) describes in detail the meditation on Tārā by means of the bija trum, which is transformed into a blue lotus, which in turn transforms itself into the bodhisattva Tārā. In several texts which are considered to be precursors of the Mahavairocanasūtra (Dainichikyo) the three rituals of warding off evil (sokusai), increasing happiness (zo yaku) and exorcising demons (chobuku) are described (T. 893, 895 and 897).
In the second section of II. 2 M. examines the history of the two mandalas (Vajradhātuand Garbha-) with their two groups of five Buddhas ; Garbhamandala : Vairocana in the centre, and in the East, the South, the West and the North respectively Ratnaketu, Samkusumitarāja, Amitāyus and Dundubhisvara; Vajradhātumandala : Vairocana in the centre, and in the East, the South, the West and the North respectively Akşobhya, Ratnasambhava, Amitäyus and Amoghasiddhi. Already in the fourth century there existed a group of four Buddhas : Akşobhya, Ratnaketu, Amitāyus and Dundubhisvara. This group continued to exist in the same form until the end of the seventh century. It is only in the beginning of the eighth century that groups of five Buddhas with Vairocana in the centre first appear in Chinese translations. In a text translated by Bodhiruci (T. 1092) two groups of five Buddhas are described, one with Sākyamuni in the centre, and in the four directions (E, S, W, N) Akşobhya, Ratnasambhava, Amitabha and Lokeśvararāja, and the other with Vairocana in the centre, and in the four directions Akşobhya, Ratnasambhava, Avalokitesvararāja and Amoghasiddhi. Another text translated by Bodhiruci (T. 951) describes a group of five Buddhas with Sākyamuni in the centre, and in the four directions Ratnaketu, Samkusumitarāja, Amitābha and Akşobhya. In the Dainichikyo this group developed into two different groups, one with Vairocana in the centre, and in the four directions Ratnaketu, Samkusumitarāja, Amitāyus and Acala, and the other with Issaisekensaisontokushin (Skt. ?) in the centre, and in the four directions Ratnaketu, Samkusumitarāja, Amitāyus and Dundubhisvara. M. considers the first group to be older and the second a later development. From the eighth century in India, the Garbhamandala did not change much, but there were important changes in the Vajradhātumandala. As to the pictorial representations of the Garbhamandala, two different traditions exist, one going back to Subhakarasimha, and the other to Amoghavajra and Buddhaguhya.

Page #8

98
J. W. de Jong However, it is doubtful that Amoghavajra and Buddhaguhya were responsible for the second tradition, which is better described as the non-Subhakarasimha tradition.
Chapter three deals with the formation of some of the most important Tantras of the middle period, such as the Dainichikyo (T. 848), the Kongochokyo (T. 865) and the Rishukyo (T. 243). The texts of this period can be distinguished from those of the first period by the following five characteristics. 1. The goal of the Tantric rituals is no longer worldly gain but the attainment of Buddhahood (jõbutsu ; abhisambodhi). 2. In the early tantras, mudra-s, dhāraņi-s and samadhi-s were not systematically arranged. In the middle period they are grouped together systematically as the three secrets (san. mitsu) of body, word and mind (kāyavākcitta). 3. The content of many early tantras was magical, with no direct relation to Buddhist teaching. In the tantras of the middle period, Mahāyāna ideas are incorporated in Tantric ritual. M. proposes to call this the ritualisation(gikika) of Mahāyāna ideas. 4. A special characteristic of the tantras in the middle period is the formation of mandalas, namely the Mahākaruņāgarbhodbhavamandala, which is based on the Dainichikyo and the Vajradhātumandala which is described in the Shinjitsushokyo (Sarvatathāgatatattvasamgrahasūtra, T. 882). 5. In the early Tantras the preacher is Sākyamuni, but in the tantras of the middle period the preacher is the Tathāgata Mahāvairocana. According to the tradition of Japanese Tantrism, the preaching of Mahāvairocana is that of the dharmakāya as opposed to the preaching of the nirmāņakaya by Sākyamuni.
Mahāyāna ideas are found above all in the first chapter of the Dainichikyo, which has as its chief theme the explanation of the sarvajña-jñāna, the knowledge of the Buddha and the highest reality. Madhyamaka influence is visible in the explanation of the bodhi and the citta as free from all ideas (vikalpa), void (śünya), without marks (alakşaņa) and unperceivable (anupalabdha). Connection with the Tathāgatagarbha system is shown by the conception of the mind as originally pure (prakstiviśuddha). According to Buddhaguhya's commentary, the bodhicitta is of two kinds, the mind whichis directed towards bodhi, called pranidhicitta and prasthānacitta, and the mind which has bodhi as its essence. The idea that the essence of the bodhicitta is identical with the supreme bodhi is not found in ordinary Mahāyāna Buddhism. The influence of Mahāyāna ideas can also be traced in the Sarvatathāgatatattvasamgraha, and espeially in the five-fold meditation described in the first chapter. This meditation, which later is called the Gosājoshingan, culminates in the identification of one's body with the body of the Buddha. In this meditation such Mahāyāna ideas as that of the Void, of the luminosity of the mind and of the symbolism of the vajra are expressed in the ritual to be performed. M. points out that by incorporating these ideas in this ritual they are

Page #9

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
99
gradually transformed in the process of the so-called ritualisation of Mahāyāna ideas. He explains how this process is further developed in two representative anuttara yogatantra-s, the Guh yasamajatantra and the Hevajratantra.
In the Tantric texts of the middle and later periods the chief goal is the union of the individual with the Buddha, the nyűga-ganyű, the merging of the Buddha with the self and the merging of the self with the Buddha. Several rituals which aim at obtaining this goal are described in the texts, for instance the so-called shusanzongan or jiingyökan in which the union with the Buddha and bodhisattva is obtained by meditating on his bija (a Sanskrit syllable), his symbol (samaya) and finally on his external form. This ritual is first described in a text translated by Chih-sung in the beginning of the seventh century (not eighth as said on p. 151) which describes the meditation on Tārā. Another ritual is the gojigonshingan, the ritual of the five syllables a, va, ra, ha and kha, which is found in the Dainichikyo and related texts. These five bija-s possess five colours and five forms and represent the five elements. They are located in meditation in five places in the body. Another ritual found in the Dainichikyo is the four-fold recitation (jāpa, nenju, tib. bzlas-briod). The Dainichikyo belongs to the caryā class of Tantras. In the Tattvasamgraha, a yogatantra, the gosõjóshingan, the fivefold meditation in which one's body is united with the Buddha is described. Many rituals are found in anuttarayoga texts and it is impossible to designate any one which is representative of this class of tantras. As an example of a ritual, M. summarises the catur angasadhana which is found in the Guhyasamāja, consisting of sevā, upasādhana, sadhana and mahasādhana. In the Uttaratantra sevā is divided into two parts, the fourfold vajra ritual (vajracatuşka) and the yoga in six members (şadanga yoga). The first forms the basis of the ut pattikrama and the second of the utpanna- or sampannakra. ma. These two krama-s are described in the Pindikstasadhana and the Pancakrama, which were edited by de La Vallée Poussin in 1896 (Études et textes tantriques. Pancakrama). Finally, M. points out that the ritual of the identification of the self with the Buddha (nyugaganya) is found in a rather complete and standardised form in Japan, but that in India many different forms existed, based upon different texts, rituals and schools. However, all these rituals have as their final goal the union of the microcosmos (the individual) with the macrocosmos (the Buddha), the unio mystica.
Sections 2, 3 and 4 of chapter three examine the history of three Tantric texts which are of great importance for Sino-Japanese Tantrism : the Dainichikyo, the Kongochokyo
2
R. Tajima, Étude sur le Mahavairocana-satra (Paris, 1936), pp. 117-118.

Page #10

100
J. W. de Jong and the Rishukyo. M. first gives some information on the Chinese and Tibetan versions of the texts and their commentaries, and then examines the problem of the randatsu, the inversion of words or passages. According to the tradition these inversions in the text were made purposely so that it would not be possible to understand the text without the aid of the oral tradition. M. rejects this traditional explanation, and by comparing the Chinese translation of the Dainichikyo with the Tibetan translations of this text and of Buddhaguhya's commentary, he shows that the theory advanced in this commentary (T. 1796) of an inversion in the mandala is not confirmed by other texts. This inversion is not found in the original text of the Dainichikyo, and therefore goes back to Subhakarasimha or to the oral tradition handed down by him. Andther case of randatsu is found in the two groups of four Buddhas in the Garbhamandala. The first group comprises Acala (i. e. Akşobhya) and the second Dundubhisvara. The commentary tries to explain this difference away by considering Acala not to be a proper name. M. explains that the text of the Dainichikyo is not well arranged and is incomplete. There are many differences between the four versions of the text, i. e, in the Chinese and Tibetan translations of the text and the Chinese and Tibetan translations of the text as found in the two commentaries (T. 1796 and the Tibetan translation of Buddhaguhya's commentary). It is obvious that the text of the Dainichikyo has been much changed in the course of time, and that different traditions have been established. The so-called randatsu-s are due to conflicting traditions and not to a conscious effort to hide the meaning of the text.
Previous scholars assumed that the Sanskrit text of the Dainichikyo was written ei. ther in 660—670, or about 650 or slightly earlier. M. points out that the archaeological evidence shows that Padmapāņi and Vajrapāņi are already found together at the end of the sixth century in the cave-temples of Aurangabad. The three basic kula-s of the Dainichikyo are represented by the Buddha, Padma and Vajra. It is therefore possible that this text had already been written in the first half of the seventh century. The problem remains why the text had not been brought to China by the middle of the seventh century. Omura Seigai thought that the Dainichikyo was written in Nālandā, but Toganoo Shoun opted for Lāță in Western India. M. rejects Toganoo's arguments and points out that the text was obtained by Wu-hsing in Nālandā. Archaeological evidence also points to the same region.
The Sanskrit text of the Dainichikyo has so far not been found, but several quotations have been traced, one in the first Bhavanakrama by Kamalaśila, fourteen
3
Cf. Tajima, op. cit., p. 54, n. 1.

Page #11

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
101
and a half verses in a text discovered in Java, three quotations in the Pradipodd yota. na, a commentary on the Guh yasamaja, and one quotation each in the Guh yasa māja and in the Sadhanamala. Wogihara Unrai identified the fourteen and a half verses which were quoted in the Sang hyang Kamahayanikam and compared the Sanskrit text with the Chinese translation. M. gives the text of these versions and in the notes refers to both the Chinese and Tibetan translations.
Kongochokyo is the name of the translation of the Sarvatatathāgatattvasamgraha (hence Tattvasamgraha) by Amoghavajra (T. 865), but is also the name of a collection of eighteen texts preached in eighteen assemblies. This extensive text, which is said to contain 100.000 ślokas (i. e. 3.200.000 syllables), is described by Amoghavajra (T. 869). Of the eighteen texts described by him the first is the Tattvasamgraha, and the sixth and fifteenth correspond to the Rishukyo and the Guh yasamajatantra. However, at the time of Amoghavajra these last two texts were much shorter than those now extant. The same must be true of the other texts described by him, with the exception of the first, the Tattvasamgraha. According to tradition, Vajrabodhi lost the larger recension of the Kongochōkyō during his sea voyage to China and was able to save only a smaller recension in 4.000 ślokas. This tradition is not to be relied upon. However, it is possible that at the time of Amoghavajra there existed several texts belonging to the Kongochokyo and many rituals which, together, can be considered as a large recension of the Kongochokyo.
The Sanskrit text of the Tattvasamgraha has been published recently by Horiuchi Kanjin. The Tibetan translation by Sraddhākaravarma and Rin-chen bzan-po dates from the eleventh century. The Tattvasamgraha was translated three times into Chinese, by Vajrabodhi (T. 866), by Amoghavajra (T. 865) and by Shih-hu (T. 882). The last is the most complete text and corresponds to the Tibetan translation. It consists of 26 sections and is divided into five chapters. Amoghavajra's translation corresponds to the first six chüan of Shih-hu's translation in thirty chüan. The manuscript brought from India by Amoghavajra contained four of the five chapters of the recension translated by Shih-hu, but he was unable to translate more than the first chapter. The text translated by Vajrabodhi represents a much shorter recension than the one translated by Amoghavajra.
Shih-hu arrivd in China in 980, and the text translated by him must have been written in India at the latest in the middle of the tenth century. Tibetan versions of three
4 See also J. W. de Jong, 'Notes on the sources and the text of the Sang hyang Kamahāyānan Mantranaya', Bijdragen tot de Taal-, Land- en Volkenkunde, 130 (1974), pp. 465-482.

Page #12

102
J. W. de Jong commentaries written by Buddhaguhya, Sākyamitra and Anandagarbha in the eighth century are found in the Tanjur. The text must therefore have been in existence in its final form at the end of the eighth century. The recensions translated by Amoghavajra and Vajrabodhi date from the first half of the eighth century and the end of the seventh century. At the end of the seventh century the original recension of the Tattvasamgraha was already in existence. Until the end of the eighth century the text was continually expanded. As to its place of origin, several traditions seem to confirm that the Tattvasamgraha was written in the South of India.
In Japan the name Rishukyo is normally used for Amoghavajra's translation, (T. 243) which is regularly recited in Shingon temples. However, there are six Chinese translations, by Hsüan-tsang (T. 220, no. 10; 660—663 A. D.), Bodhiruci (T. 240 ; 693 A. D.), Vajrabodhi (T. 241 ; the attribution of this translation to Vajrabodhi is doubtful), Amoghavajra (T. 243; 765—771 A. D.), Shih-hu (T. 242; 991 A. D.) and Fa-hsien (T. 244 ; 999 A .D.). There are three Tibetan translations, Sriparamādya nāma mahāyānakalparāja and Sriparamadya mantrakalpakhanda nāma (Tōhoku nos. 487—488), Srivajramandalalamkāra nāma mahātantrarāja (Tohoku no. 490) and Āryaprajña pāramitānayasata pañcaśatikā (Tōhoku no. 489). Finally, there is a text written partly in Sanskrit and partly in Khotanese (cf. E. Leumann, 'Die nordarischen Abschnitte des Adhyardhaśatikā-Prajñāpāramitā', Taisho daigaku gakuho, 1930, pp. 47-87). The translation by Fa-hsien and the first two Tibetan translations are designated as the larger recension, and all the other texts as belonging to the smaller recension. The relations between these different translations and the development of the text have been studied first by Toganoo Shoun in his book Rishukyo no Kenkyū (Köyasan, 1930; cf. Biblio.' graphie bouddhique, IV–V, 1934, no. 450), and, in recent years, by Nasu Seiryū, Nagasawa Jitsudo, Kanaoka Shūyū and Fukuda Ryōsei. All scholars agree on two points, The oldest and most primitive recension is the one translated by Hsüan-tsang. Out of this recension developed the one translated by Bodhiruci. However, there is no agreement at all with regard to the text's further development. Some think that the larger recension is older and the smaller an extract from it. Others think that the larger recension has developed out of the smaller one. Still others suppose that the original text was different from both the larger and smaller recensions.
M. recapitulates the different theories proposed and subjects them to a critical examination. He points out that it is wrong to assume that the different texts can be supposed to have developed in a direct line one out of the other because there are too many discrepancies between them. In the first half of the seventh century a brief quotation from the Prajñaparamita-ardhaśatika or Dvyardhasatikā is made by Candrakīrti in his

Page #13

103
A NEW HISTORY OF TANTRIC LITERATURE IN INDIA Prasanna pada : śūnyaḥ sarvadhar mā niņsvabhāvayogena "all dharmas are empty as they are without essence" (pp. 238. 8, 278. 14, 444. 8,500.11 and 504.7). Sakai Shinten has traced this quotation to Amoghavajra's translation (T. 243, p. 785a 25). Important information on the development of the Rishukyo is found in the introduction of Jñānamitra's commentary, which is preserved in Tibetan translation (Tōhoku no. 2647). M. re-examines this passage, which was previously used by Toganoo and Nagasawa, but with whose interpretations M. does not agree. According to this passage, an ācārya named Kukure or Kukupa, who was versed in the eighteen great texts, the Sarvabuddhasamayoga, etc., came to the kingdom of Zahor and taught Sakrabhūti, the son of king Indrabhūti, the traditions (lun ; Skt. āgama) and the secret instructions (man-nag; Skt. upadeśa). Sakrabhūti extracted for his daughter Govadevi the Tshul brgya-Ina-bcu-pa (Nayaśata pañcaśatika) from the Dpal-dam-pa (Śriparamādi ?). Govadevī developed the traditions and the secret instructions, and so created the text which is now extant. Toganoo supposed that the Sriparamādi was identical with the larger version, called Śriparamadya (cf. Tōhoku nos. 487---488), and that the Adhyardhaśatikā, the smaller version, was extracted from it. According to Nagasawa the Sriparamādi is not identical with the Sriparamadya but represents the original text, from which the smaller recension was extracted and which was expanded to form the larger recension. M. disagrees with this conclusion. He quotes another passage from the introduction of Jñānamitra's commentary which refers to the Dpal dam-pa phren-ba as a Prajñāpāramitā text. This text must be closely related to the texts which are called Sriparamădya. According to M. the Dpal dam-pa refers to the larger recension and not to the original text. Moreover, M. has doubts regarding the correctness of the Sanskrit reconstruction of Śriparama for Dpal dam pa, because the Tibetan translation of the Sriparamadya translates parama by mchog and not by dam-pa, and because the Sriparamadya mantrakalpakhanda distinguishes between dam-pa and mchog. Furthermore, the eighteen texts mentioned in Jñānamitra's commentary are not the eighteen texts of the Kongochōkyō as supposed by Toganoo and Nagasawa, but the eighteen tantras which are transmitted by the rÑin-ma-pa school. Finally, M. points out that Jñānaśrīmitra's commentary mentions not only the Dpal dam-pa but also traditions (agama)and secret instructions (upadeśa). According to M. the ācāryas developed different traditions and secret instructions. In a later period these traditions and instructions were written down in the form of sūtras and in this way different texts were created. Jñānamitra explained that
5 Matsunaga remarks that dam pa can translate agra, vara, uttama, para, etc. However, the reconstruction Sriparama for Dpal dam-pa is the most likely one.

Page #14

104
J. W. de Jong the Nayaśata pañcaśatika was extracted from the Dpal dam-pa. A text of this name existed already in the first half of the seventh century because it is mentioned by Candrakirti. Therefore, it can be considered nearly certain that in the course of the development of the Rishukyo from a Prajñāpāramita text an important role was played by a text called Sriparamadya or Satapancaśatika. Already Bodhiruci's translation is Tantric in style, and this shows that this development of a Prajñāpāramitā text into a Tantric text was already well advanced in the seventh century.
The larger recension, Sriparamadya, according to the Tibetan translation, (Tohoku nos. 487-488) consists of three parts. The first part corresponds to the smaller recension but contains moreover descriptions of constructions of mandala-s and rituals connected with mandala-s. Probably the original text was expanded by means of all kinds of sādhana-s and rituals which were practised in connection with it. The larger recension did not yet exist at the time of Amoghavajra because his analysis of the contents of the Rishukyo (cf. T. 869) has no relation with the existing larger recension. Three texts translated by Amoghavajra (T. 1119, 1120 and 1123) are considered to be extracted from the Sriparamadya. However, their main theme is the idea of the gohimitsu (the five secrets represented by the five Vajra Bodhisattvas). At the time of Amoghavajra the Sriparamādya was only a short text, which was later expanded into the larger recension by incorporating all kinds of traditions and secret instructions.
In the process of the transformation of a Prajñāpāramitā text into a Tantric text important changes took place. In the texts translated by Hsuan-tsang and Bodhiruci the preacher is the Bhagavat himself or the Bhagavat who takes the aspect of different Ta. thāgatas, but in the text translated by Amoghavajra the text is preached by different Tathāgatas. In sections 3 to 10 of the main text which comprises 17 sections the preaching is repeated by eight great bodhisattvas who summarise the teaching in the form of a bija, a spell consisting of one syllable. These bodhisattvas are the sambhogakaya-s of Mahāvairocana. Spells are not found in Hsüan-tsang's translation. In Bodhiruci's translation at the end of each section it is only said that the Tathāgata pronounced the following spell. However, in Amoghavajra's translation the spell is considered to express the essence of the teaching in each section. In early Tantric texts spells were used to ward off evil and to obtain happiness. In the middle period the motivation of the spell is turned inwards (naimenka). The third change concerns the description of the merits of the recitation of the text. The merits are mentioned in greater detail at the end of the first section than at the end of the other sections. At the end of this section Hsüan-tsang's version mentions the attainment of Buddhahood, but this is not found in the texts translated by Bodhiruci and Amoghavajra, which stress respectively the attainment of the

Page #15

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
105 vajrakaya of the Buddha-Tathāgata-s and the attainment of the states of Tathāgata and Vajradhara. Although in the tantras of the middle period the attainment of Buddhahood is the goal, this is not often expressly stated. According to M. the reason is that one strives to obtain union with the Absolute, for instance with Mahāvairocana. This implies the attainment of Buddhahood.
The text translated by Amoghavajra is a Tantric text, but it still shows traces of the fact that it developed out of a Prajñāpāramitā text. Amoghavajra's version contains no important ritual elements, and the philosophical contents are more important. However, from the eighth century onwards the text was greatly transformed by the addition of many rituals. The larger recension represents a ritualisation of the text. At the time of Amoghavajra there were already rituals relating to the Rishukyo but they were not yet incorporated in the text. The complete transformation into a Tantric text was achieved by the creation of the different texts belonging to the larger recension by the addition of many rituals.
The fourth chapter deals almost entirely with the Guhyasa mājatantra and related texts. The G. [=Guhyasamāja] belongs to the anuttarayoga tantras which are divided into three groups: 1. the yuganaddha or advaya tantras ; 2. the prajñā or mother (ma) tantras ; 3. the upaya or father (pha) tantras. To the first group belong the Kalacakra tantras. The second and third group are divided into seven and six families. The G. belongs to the Akşobhya family of the upāya or father tantras. It has eighteen chapters, of which seventeen form the Mulatantra and the eighteenth the Uttaratantra. The G. was translated into Tibetan by Sraddhākaravarma and Rin-chen bzan-po in the beginning of the eleventh century and by Shih-hu into Chinese in 1002. The G. was very popular in India and Tibet but not in China.
The Malatantra can be divided into two parts, the first comprising chapters 1-12 and the second chapters 13–17. There are two important traditions relating to the G., the Jñānapāda school which goes back to Jñānapāda, and the 'Phags-lugs which goes back to Aryanāgārjuna. In one of the oldest commentaries (Ākhyānatantra), the Samdhivyākaraņa, there is as yet no clear influence of one of these two schools. The Samdhivyakarana explains only the first 12 chapters of the G. This is an indication that this part of the G. is older than the second part. The same conclusion emerges from an examination of the G. itself. A commentary of the Dge-lugs-pa school on the Pradipoddyotana, a commentary on the G., distinguishes three traditions regarding the construction of mandala-s: the Indrabhūti tradition with thirteen Venerables, the Jñānapāda tra6 The word 'Venerable translates Japanese son, which is always used to indicate the different persons (Buddhas, bodhisattvas, Vidyārājas, etc.) in a mandala.

Page #16

106
J. W. de Jong dition with nineteen Venerables and the 'Phags-lugs with thirty-two Venerables. The first is described in Indrabhūti's Jñanasiddhi and agrees entirely with the mandala found in the first chapter of the G. (five Buddhas with Akşobhya as central figure, four śakti-s and four krodha-s) The mandala of the Jñānapāda tradition comprises a group of five Buddhas with Mañjuvajra as central figure, four śakti-s and ten krodha-s, adding six krodha-s to the four in the previous mandala. However, the names of these six krodha-s are not found in the first part of the G., but only in the second part. The mandala of the 'Phags-lugs is not found in the G. itself and must have developed in a later period. The Uttaratantra mentions the fourfold upaya (sevāsadhana, upasadhana, sadhana, mahāsādhana) and declares that chapters 2-17 of the G. correspond to these four upayas, cf. below p. 108. However, the original form of these four upāya-s is already found in the twelfth chapter. There is also a great difference in content and size between the first and the second half of the G. Chapters 13–17 are much longer and, moreover, describe many rituals. Finally, the Uttaratantra recapitulates the teachings and the rituals in fifty-two questions and answers. Everything mentioned here is already found in the first half of the G. It is therefore reasonable to suppose that, originally, the first half of the G. was a complete text in itself.
With regard to the doctrines, the sadhana-s and the construction of the mandalas of the anuttarayogatantra-s, both the father and mother Tantras are based upon the Tattvasamgraha. An intermediary position between the Tattvasamgraha and the G. is occupied by the Māyājalatantra, which is extant in a Chinese translation (T. 890) and in two Tibetan translations (Tohoku nos. 466 and 833). The second translation represents a different and later text. A commentary on the G. (Tōhoku no. 1909) quotes from the Guhyagarbha (Tōhoku no. 832), which is based upon the Māyājalatantra. This indi.. cates a close relation between the G. and the Mayājalatantra. The mandala of the Mayajalatantra comprises a group of five Buddhas with Vairocana as the central figure, whereas in the G. the central figure is Akşobhya as in the Tattvasamgraha, a yogatantra. The four śakti-s of the mandala of the Maya jalatantra are not found in the Tattvasamgraha, but their names are mentioned in other yogatantras. The mandala of the Māyājalatantra comprises also two groups of four krodha-s. Ten krodha-s are found in a text translated in Chinese in the Sung period (T. 891) and in a collection of mandalas composed in the eleventh century, the Nişpannayogāvali. Both the four śakti-s and the ten krodha-s are found in the G. The Mayājālatantra clearly represents a transitional stage between the Tattvasamgraha, a yogatantra, and the Guhyasamaja, an anuttarayogatantra.
In the past the G. and the Tathāgataguhyasūtra were considered to be identical,

Page #17

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
107
an error that has persisted for a long time, although it was pointed out long ago by Thomas Watters in a note to Bendall's edition of the Sikşāsamuccaya (St. - Petersburg, 1897—1902, p. 274 n.). Bhattacharya assumed wrongly that the G. was written by Asanga in the fourth century. In recent years the date of the G. has been studied by Wayman, Tucci and Hadano. According to the tradition of the Shingon school in Japan the G. corresponds to the fifteenth text of the Kongochok yo as described by Amoghavajra (T. 869). However, a comparison of Amoghavajra's description with the G, shows that the correspondence is limited to the title, the place of preaching and the teachings found in chapter five of the G. At the time of Amoghavajra (700—750) the G. had not been composed yet. The G. was translated into Chinese in 1002 and into Tibetan at about the same time. The lower limit for its composition is around 1000 A. D. In order to determine its date more precisely it is necessary to study the works of Jñānapāda who, according to Hadano, lived in the period 750-800. Although Jñānapāda wrote works relating to the G., this does not allow us to assume the existence of the G. at that time, becausethe relations between a text and the sadhana-s and commentaries belonging to a particular school are not easy to determine. In some cases the text was written first and the sadhana-s and commentaries later. It also happened that on the basis of several sādhana-s a text would be written. Finally, a commentary expressing the opinions of a certain school was sometimes used as a basis for the fabrication of a text. In Jñānapāda's works are found three verses of the G, which express some of its fundamental ideas. It is not possible to determine whether these verses were quoted by Jñānapāda from the G. Although it is therefore not certain that by 800 A. D. the G. had already taken its present form, it must have existed in that period as an independent text, or at least in such a form that it provided sufficient basis for the formation of the text.
In several respects the teachings of the Uttaratantra are more advanced than those of the Malatantra. The two krama-s are not mentioned in the Malatantra but are first found in the Uttaratantra (verse 84). The Uttaratantra adds a fourth abhiseka to the three abhiseka-s (kalaśa, guhya and prajñajñāna). The doctrine of the fourfold upaya is further developed in the Uttaratantra. In the Uttaratantra, the vajracatuṣka (śünyatabodhi, bijasamhsta, bimbanispatti, akşaranyäsa) and the yoga in six members (sadangayoga: pratyahara, dhyāna, prāmāyāma, dhāraņā, anusmrti and samadhi) form the basis of the practice of the anuttarayoga.
There have been two translations of the Uttaratantra, an old and a new. The Kanjur contains only the new translation, but most of the verses of the old translation are quoted in Viśvamitra's commentary. There are differences in terminology between the two translations, but the old translation must have been partially revised on the basis of the

Page #18

J. W. de Jong.
new translation. There are also several discrepancies in the teachings of the two translations. In the new translation each of the four upaya-s is preached in four chapters, the mahāsādhana in chapters 5, 9, 17 and 13, the sadhana in chapters 4, 16, 8 and 12, the upasadhana in chapters 6, 2, 10 and 14 and the seva in chapters 7, 3, 11 and 15 (verses 25-28). In the old translation the upasadhana is preached in chapters 6, 2, 15 and 14 and the seva in chapters 7, 1, 3, 10, 11 and 5. In the old translation chapter 1 is mentioned and chapter 5 occurs twice, at the beginning and at the end. The second difference concerns the definition of the word tantra (verses 34-35). In the old translation the three divisions of the tantra into adhara, prakrti and as amharya are not considered to be upaya, hetu and phala. It divides the tantra into 'adhara, prakṛti and prabheda and these three are considered to indicate divisions in the content of the G. Finally, there is a great difference between the two translations in the treatment of the two krama-s, which do not have in the old translation the same importance as in later periods. Differences between an old and a new translation are not always due to the fact that the text used by the translators of the old translation is older than that used by the translators of the new translations. The differences between the two are sometimes caused by differences in the traditions of the schools. However, in the case of the Uttaratantra the differences between the old and the new translations are certainly due to the fact that the text used to make the old translation represents an earlier stage of development. This is shown clearly by the testimony of Bu-ston, by the terminology, by the fact that Viśvamitra quotes several tantras which were current at the time of the old translation and, finally, by the fact that the old translation gives less weight to the four upaya-s and the two krama-s which later were considered very important in anuttarayoga Tantrism.
108
The next section (IV. 3) is entitled the Ākhyānatantras and the Ārya school ('Phagslugs). Among the Akhyānatantras of the G. the most important are the following four: Samdhivyakaranatantra, Vajramalatantra, Caturdevtpariprechatantra and the Vajrasamuccayatantra. To the 'Phags-lugs belong two sadhana-s, the Pindtkrtasadhana and the Pañcakrama which describe the utpatti- and sampannakrama-s. The Sanskrit text of both texts was published by de La Vallée Poussin (Études et textes tantriques. Pañcakrama. Gand-Louvain, 1896). To the 'Phags-lugs belongs also a commentary on the G. written by Candrakirti, the Pradipoddyotana of which the Sanskrit text has been preserved.
The exegetical principles of the Pradipoddyotana are called the seven alamkaran: 1. upodghata (samjña, nimitta, karty, prama, prayojana); 2. nyaya (santana, nidana, nirukti, hetu); 3. jatkotika(neyartha, nitärtha, samdhyabhasa, nasandhya,

Page #19

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
109 yatharuta, naruta) ; 4. caturvidhākhyāyika (akşarārtha, samastanga, garbhin, kolika); 5. dvividhabheda (satravyākhyāna, śişyavyakhyāna); 6. pañcapudgala (ut pala, pundarika, padma, candana, ratna); 7. satyadva yavinirnaya. Of special importance are the third and the fourth principles, which have recently been studied by Ernst Steinkellner. M. briefly explains the seven alamkāra-s and then attempts to explain why this system was developed and why it was considered so important in the 'Phags lugs. The 'Phags-lugs school was created some time after the formation of the Malatantra and taught many doctrines and practices which are not found in the Guh yasamājatantra. The system of the seven alamkāra-s was created in order to legitimise the interpretation of the Guhyasamājatantra in the 'Phags-lugs. Moreover, in the later Tantric period many Hindu elements and especially śākta doctrines were incorporated into Buddhist Tantrism. The system of the seven alamkara-s served not only to proclaim the superiority of the exegesis of the 'Phags-lugs but also to demonstrate the orthodox Buddhist character of the teachings of the tantras. M. mentions that in recent years Tantrism has more and more come to be considered as the necessary consequence of Mahāyāna Buddhism. It is in this connection important to see how in Buddhist Tantrism, Buddhism and Tantrism were joined together, and in order to understand this it is necessary to give some attention to the system of the seven alamkara-s.
One of the Ākhyānatantras of the Guhyasamāja, the Vajramalatantra, contains 68 chapters. Of special interest is the relation between the Vajramalatantra and the Pañcakrama. A comparison of chapters 1-67 of the Vajramalatantra with the Pañca. krama shows that there is no relation between the two and that the Vajramalatantra is therefore not the scriptural basis for the system of practice described in the Pancakrama. However, the 68th chapter of the Vajramālātantra is closely related to the Pancakrama. Several verses of the Pancakrama are also found in the Vajramalatantra (Pañcakrama IV, 8, 9, 11; IV, 19–21; I, 16–24 ; not identical but similar: 1, 4-6). M. shows that it is very likely that the verses IV, 19–21 were taken from the Pañcakrama by the author of the Vajramālātantra. According to the present text of the Pañcakrama, verses I, 16-24 are a quotation from the Vajramalatantra. In the oldest commentary on the Pañcakrama, the Caryāmela paka pradipa, only verses 19 -23 are explained. Moreover, these same verses are found in the Sambarodayatantra. The Caryāmela pakapradipa also does not explain the verses quoted from the Catur. devipariprcchatantra and the Advayasamatavija yamahayogatantra in chap
7
'Remarks on Tantric Hermeneutics, Proceedings of the Csoma de Körös Memorial Symposium (Budapest, 1978), pp. 445-458.

Page #20

110
J. W. de Jong ter I of the Pañcakrama. Another commentary, the Vajrajapatika. does not explain the verses quoted in chapter 1 from the Samdhivyakaranatantra, but only explains the mantra syllables om, aḥ and hūm (cf. Pañcakrama I, 42—43). The original text of the first chapter of the Pancakrama must have been expanded later by several quotations from the Samdhivyakaranatantra (12—13), the Vajramalatantra (16-18 and 24), the Advayasamatāvijaya (59—66), etc., The 68th chapter of the Vajramālatantra must have been composed after the Pancakrama. Although the Ākhyānatantras are generally considered more authoritative than sadhana-s and commentaries, this example shows how an Akhyanatantra is fabricated in order to legitimise the teachings of an already existing sadhana text.
In the next subsection (IV. 3. 5) M. examines the relation of the Vajrajnanasamuccayatantra (V.) with the Pancakrama and the Pradipoddyotana. In explaining the three jñana-s, V. is less complete than the Pañcakrama and does not mention all the original forms (praksti) found in the Pañcakrama. However, this does not mean that V. is later than the Pancakrama. The author of V. must have known the complete system of the Pancakrama. He mentions only some representative prakrti-s and refers to the others by adding 'etc.'. The second chapter of the Pancakrama deals with the three jñāna-s but does not quote V. V. must have been composed after the Pancakraта.
The Pradipoddyotana is based upon the sādhana system of the Phags-lugs as found in the Piņņikstasādhana, the Pancakrama, etc. It was therefore composed in a later period. The main theme of the Pradipoddyotana is the group of seven alamkara-s. In the first half of V., the most important of these seven, the third and fourth, are explained, and the names of all seven and their subdivisions are found in the second half of y. However, V. is never quoted by name in the Pradipodd yotana. V. is a very brief text containing only four and a half leaves. The explanations of the third and fourth alamkara-s in the first half of V. are found to be incomplete when compared with the explanations found in the Pradipoddyotana. The second half of V. is more detailed than the Pradipoddyotana with regard to the names of the seven alamhara-s and their subdivisions. The first chapter of V. has no colophon but the second chapter contains a colophon. The second chapter must have been added later, which lead to the omission of the colophon of the first chapter. It is obvious that the first half of V. was composed before the Pradipodd yotana. However, the first half of V. did not contain a detailed explanation of the seven alamkara-s, and for this reason the second half was later added. This part must have been composed after the Pradipoddyotana was written. The first half of V. was composed not long before the Pradipoddyotana and this ex

Page #21

A NEW HISTORY OF TANTRIC LITERATURE IN INDIA
111 plains why this text is not quoted in the Pradipoddyotana as a scriptural authority.
The Śrijñanavajrasamuccaya (Tohoku no. 450) is similar to V. but much more detailed. It contains many teachings which are not found in V. It must therefore have been composed after V. M. summarises the results of his studies in the texts of the 'Phags-lugs by indicating the order in which the texts are composed : 1. Vajramālātantra chapters 1-67; 2. Pindikštasadhana and the Pancakrama in its original form; 3. Vajramalatantra chapter 68; 4. The Pañcakrama in its present form ; 5. The first half of the Vajrajñānasamuccayatantra ; 6. Pradipoddyotana ; 7. The second half of the Vajrajñanasamuccaya ; 8. Srijñanavajrasamuccaya.
The Mūlatantra and the Uttaratantra of the Guhyasamaja were composed around 800 A. D. The 'Phags-lugs flourished around 1000 A. D. Between 800 and 1000 many new doctrines and practices arose in the 'Phags-lugs, and in order to legitimise them it was necessary to fabricate Ākhyānatantra-s or to make additions to existing Akhyanatantra-s.
In the last section of Chapter IV (IV. 4) M. studies the Mañ juśtimülakalpa (Ma.), the Sanskrit text of which was published in three volumes by Ganapati Sâstrî (Trivandrum, 1920, 1922 and 1925). Ma. was characterised by Jean Przyluski in the following way: "C'est une sorte d'encyclopédie qui traite, sous forme de sermons, des sujets les plus variés : iconographie, rituel, astrologie, etc....” The Sanskrit text contains 55 chapters, the Tibetan translation 37 and T'ien-hsi-tsai's Chinese translation (T. 1191) 28. Moreover, there are Chinese translations of several other chapters (41, 50, 51 and 55). M. compares the different texts in Sanskrit, Tibetan and Chinese.
Ma. has been studied by Indian and Western scholars. M. relates briefly the opinions of Bhattacharya, Snellgrove and Dutt on the date of Ma. before discussing in greater detail the opinion expressed by Przyluski. Przyluski compared three Chinese translations of a text relating to the mantra of Mañjuśri (T. nos. 1181, 1182 and 956) with chapter 14 of Ma. However, chapter 9 is much closer to T. 1181 and 1182, which were translated in 702 and 703. This chapter must therefore already have been in existence at the end of the seventh century. The 41st chapter of Ma. (Garuda patalaparivarta) was translated into Chinese by Amoghavajra. However, the second half of Amoghavajra's trans
8 "Les Vidyārā ja. Contribution à l'histoire de la magie dans les sectes Mahāyānistes', BEFEO 23
(1923), p. 301. Przyluski's study is based on the first volume only of Ganapati Sastri's edition (cf. pp. 302—303). This has been overlooked by Matsunaga, who reproaches him for not having studied the relations between the Sanskrit text of chapter 41 and the corresponding Chinese translation

Page #22

112
J. W. de Jong
lation is full of Tantric expressions which are not found in other translations of that period. Moreover, in Kūkai's catalogue the Chinese translation is said to consist of only three folios. Amoghavajra translated only the first half, and the second half was added later. The Sanskrit text of Ma. comprises three chapters (50—52) dealing with rituals relating to Yamāntaka. The corresponding Chinese translations (T. 1215, 1216) and another Yamāntaka text translated by Amoghavajra (T. 1214) must be compared with the Sanskrit text and the Tibetan translation of chapters 50—52 of Ma. The comparison shows that the first Chinese translation (T. 1215) is different in character from the other two (T. 1216 first part and second part) and from T. 1214. The three Chinese translations are attributed to Amoghavajra, but only the first (T. 1215) can have been translated by him. In the other two Tantric characteristics are much more prominent. There are also differences between all four Chinese texts with regard to the person of the preacher and the person or persons to whom the text is preached. Chapters 50-52 of Ma. must be later than chapters 4-46 and T. 1214, and must have been composed in different periods. They constitute a later addition to Ma. The historical chapter 53 (Rajavyāka. raņa-parivarta) mentions the Pāla king Gopāla, who probably reigned about 750 A.D. This chapter must have been written in the same period.
Finally, M. points out that it is necessary to study carefully the contents of Ma. in order to determine its position in the history of Tantrism and, in a wider perspective, in the history of Indian culture. However, at present there are many uncertainties which make such an undertaking practically impossible. Furthermore, Ma. contains many materials relating to geography, history, astronomy, mathematics, etc., and it will be a task for the future to study the text from the points of view of the natural and human sciences. Studies of this sort will also contribute to the clarification of the historical de velopment of the text.
It is of course impossible to give a complete survey of the contents of a book in which so many difficult problems are studied. Matsunaga's book shows clearly how complicated the historical development of Tantric rituals and texts is. In the past, many scholars have not hesitated to express very definite opinions about the historical development of Tantrism in India and about the dates of texts without undertaking a detailed study of the available evidence. It is impossible to study the history of Tantrism in India without consulting the Chinese translations of Tantric texts. For the study of relations between texts it is absolutely necessary to take into account Sanskrit texts, Chinese translations and Tibetan translations. It is the great merit of this book that it has made full use of all the materials available. In these studies of the formation of Tantric texts, Matsunaga

Page #23
- A NEW HISTORY OF TANTRIC LITERATURE IN INDIA 113 has concentrated his remarks on some of the most important tantras. It is to be hoped that his work will stimulate research along the same lines on other Tantric texts. The amount of Tantric literature is enormous. Many Sanskrit manuscripts have not yet been edited. Several hundreds of texts were preserved in Chinese translation and many more in Tibetan translation. It is still impossible to gain an exact idea of the amount of exegetical literature written by Tibetan scholars since the introduction of Tantrism into Tibet up to the present day. Little is known about the relations between Buddhist Tantrism and Hindu Tantrism in India and about the development of Tantrism in Tibet. Since the publication of Louis de La Vallee Poussin's Bouddhisme. Etudes et Materiaux (London, 1898), Tantric studies have made much progress, thanks to the work done by scholars in Japan, India and the West, but we still have a long way to go before a true picture of Tantrism in all its complexity can emerge. We must be grateful to Matsunaga for having made such an important contribution towards this distant goal. Australian National University J. W. de Jong.