Chapter 6.2 - The Devi-purana (study)

This work [devi-purana],¹⁰⁵ which is one of the most important of the

¹⁰⁵ The only printed edition of this work is that published in Bengali characters with a Bengali translation by the Vangavasi Press, Calcutta. Though, as we shall see presently, the Devi-purana is a work of great importance from different points of view, it has not yet been edited critically, nor is there a single edition printed in Devanagari characters. In fact, we know of no edition of this work other than that published by the Vangavasi Press, Calcutta. This want has forced us to use this edition in spite of its manifold defects. Manuscripts of the Devi-purana are not very numerous. Eggeling, in his Julius Eggeling, A Descriptive Catalogue, VI, pp. 1205-7 (No. 3362), describes a Manuscript (No. 380) of this Upapurana 'apparently consisting of some 7000-8000 slokas' and written in Bengali script by three different hands in 1804 A.D. Its contents are practically the same as those of our printed edition, and it ends thus: sadacarah subhaharah sarva-sanga-vivarjitah / vacayan satam ardham va pratyaham prapnuyac chivam / / etac cabhyudayam padam satais tribhir narottamaih / sahasrair dvadasair vatsa kathitam sarva-siddhidam / / ity adye devy-avatare sri-devipurane 'bhyudaya-padah samaptah | l (The second verse 'etac cabhyudayam padam etc.' does not occur in the printed edition. The final colophon also is different there). In his Bikaner Cat., p. 199 (No. 433) R. L. Mitra describes a paper Manuscript of the Devi-purana which is written in Nagara and has the same contents as those of the India Office Manuscript (No. 380) described by Eggeling. Like the latter, this Manuscript also ends with the verses 'sadacarah subhaharah' and 'etac cabhyudayam padam' (of which the former reads 'pratyaksam" for 'pratyaham' in the second half) but its final colophon runs as follows: adye devyavatare devipuranam samaptam iti. In the Calcutta Sanskrit College Library there are three complete paper Manuscripts (Nos. 176, 308 and 341) of the Devi-purana, of which the first (No. 176), written in Bengali script, contains 138 chapters and ends thus: sadacarah subhacarah sarva-sanga-vivarjitah / vacayan puranam etam tu sarva-kamam avapnuyat || ity adye devy-avatare devipuranam samaptam || The second Manuscript (No. 308) is written in Nagara characters and has been described by Hrishikesh Shastri and Shiva Chandra Gui in their Calcutta Sanskrit College Cat., IV, pp. 187-8.

Sakta Upapuranas, deals, in 128 chapters, mainly with the exploits and worship of Devi (the great goddess) who incarnated herself on the Vindhyas as a maiden mounted on a lion and became known as Vindhyavasini. In this Puranna Devi is not only called 'yoga-nidra' and the primordial and pre-eminent Energy (adya para sakti-7. 20) and at the same time said to be the same as Uma, the wife and Sakti (Energy) of Siva, but is also identified with the Saktis of other gods as well as with the Matrs and other female deities such as Daksayani, Kali, Candi etc. who are called Devi's different forms. It gives us important information about the different incarnations of Devi and her original nature and relationship with Siva and other gods; about Yoga and the method of its practice; about Sakta iconography and the Sakta vows and worships; about Saivism (as related to Saktism), Vais- navism, Brahmaism and Ganapatyaism; about warfare as well as the construction of towns and forts and the means of their protection; about the different Vedic schools as well as the Upavedas, Angas, Upangas etc.; about the science of medicine; about manuscripts-the method of copying, the script and materials to be used for the purpose, the characteristics of the scribes, and the method of making gifts of these; about holy places (in connection with which many It is fresh and undated and contains the text of the Devi-purana in 138 chapters. It ends with the same two verses and final colophon as those of the Bikaner Manuscript mentioned above. The third Manuscript (No. 341) is written in Bengali script and was copied on the eleventh lunar day of the bright half of Sravana in the Saka year 1743. Its final verse and colophon are the same as those of Manuscript No. 176 mentioned above. In his Notices of Sanskrit Manuscripts, Volume VI, pp. 180-5 (No. 2118) R. L. Mitra describes an old, undated, complete paper Manuscript written in Bengali characters and containing 128 chapters and 7371 slokas. The analysis of its contents as well as the final verse, as given by Mitra, shows that it is generally the same as our Devi-purana, chapters 1-127. Yet another Manuscript of the Devi-purana is described by Haraprasad Shastri in his Asiatic Society of Bengal (Calcutta) Cat., V, pp. 769-770, No. 4107 (Manuscript No. 4566 D). It is fresh, undated and incomplete and is written in Bengali script. Its final colophon, as given by Shastri, shows that it is the same as our Devi-purana, chapters 1-127. For other Manuscripts of the Devi-purana see Catalogue of the Sanskrit Manuscripts, p. 337 (in which there is mention of an incomplete Manuscript written in Devanagari) and p. 338 (mentioning a complete Manuscript written in Devanagari and dated '1609'); Chintaharan Chintaharan Chakravarti, p. 71 (where there is mention of an incomplete Manuscript written in Bengali script and dated Saka 1726); and Cat. of Sanskrit Manuscripts in Private Libraries in the North-west Provinces, Part I, Benares 1874, No. 452.

THE SAKTA UPAPURANAS 37 countries and towns of historical interest have been named); about different kinds of gifts; about customs and usages; and so on. The Devi-purana begins abruptly with an opening verse 108 without any mention of the person to whom this verse is attributed. This unknown reporter further informs us that some sages request Vasistha to speak on various topics, viz., the killing of the past, present and future demons by Devi, Indra's recovery of the kingdom of heaven from the demons, the hoisting of Indra's banner, the sixty incarnations of Devi, and so on. 107 Consequently, Vasistha promises to narrate to them, in four sections (pada) according as Devi was incarnated in the different Yugas, 108 all the Purana-topics which, he says, were related (1) by Siva to Visnu, Brahma and others, (2) by Brahma to Matarisvan, and (3) by Manu, Atri and Bhrgu (who received these probably from Matarisvan) to Vasistha and others (including Agastya ?). He adds that it is Agastya's speeches on all these topics which would be widely circulated on earth through kings. 109 Vasistha further gives the titles and contents of the four sections (padas) thus: (1) The first Pada, which is called Trailokya-vijaya or simply Vijaya, deals with the origin of Devi at the beginning of creation. 110 (2) The second Pada, which is termed Trailokyabhyudaya, deals with the following topics: the story of Sakra 10ª namaskrtya sivam devin sarva-bhagavatam subham / puranam sampravaksyami yathoktam brahmana pura / / Devi-purana 1. 1. 107 All these topics have been treated of in the present Devi-purana 108 Cf. catuspada-vibhagena yathayuga-kramagata / devi sarva-sukhavaptim prayacchati prapujita / / Devi-purana 1.30. 10 Devi-purana 1. 29-32. Also Devi-purana 128.1. 11º Cf. utpatti-kirtanam srsteh prathamam samudahrtam / vijaye deva-pade tu rsinam pariprcchatam // Devi-purana 1.36. The first line may be interpreted in the following ways: (1) 'the origin (of Devi) before creation has been described', and (2) 'the first origin of creation has been described'. It is to be noted that in Devi-purana 127.24 f. the first origin of Devi before, or rather at the beginning of, creation has been described. J

38 STUDIES IN THE UPAPURANAS (i.e. Indra); the killing of Dundubhi; Ghora's rise, his receipt of a boon from Visnu, his overlordship through the practice of a spell, his son's entrance into heaven, Ghora's overcoming Maya, and his deception by Brhaspati; Devi's descent on the Vindhyas; Narada's muttering of the great Padamala-vidya with a view to deluding Ghora who was to take the form of the demon Mahisa; the killing of Maya, Khatva (a demon) and others by the gods after worshipping Rudra and the numerous forms of Devi. (3) The third Pada is called Sumbha-Nisumbhamathana. (4) The fourth Pada, which is not clearly designated (but which seems to be called 'Devasura-yuddha'), deals with the following topics: Andhaka's fight; war between the gods and demons; Andhaka's attainment of the status of Bhrngi by eulogising Hara; war between Taraka and Karttikeya; story of the birth of Karttikeya-loss of Kama's body, Visnu's worship of Siva for the good of Indra, and Karttikeya's birth and command over the army of gods; origin of Uma-Kali and her attainment of Siva as husband; Himalaya's giving his daughter in marriage; origin of Balakhilyas and other sages who acted as Hotrs; story of gods and sages mounted on the chariot of the Sun; various states to be attained through results of actions; origin of Mahasveta who was engaged to protect the Sun; story of Jambha and others as well as of the gods who were transformed into planets; performance of sacrifice to the planets by Brahma; and residence of the Matrs in different regions for the good of children¹¹¹ Next, after mentioning the results, rules and fit places of study of the Purana, Vasistha reports the interlocution between king Nrpavahana and his preceptor Citrangada without giving the least introduction about these two interlocutors. From this interlocution we understand that Nrpavahana¹¹ asked his preceptor Citrangada to declare 111 Devi-purana 1. 36-52. 112 The substitution of the name 'Nrpavahana' by 'Vidyesvara', 'Vidyapa' and

THE SAKTA UPAPURANAS 39 to him the different Vidyas on khadga, mala, anjana, gutika etc. which the latter had received from Indra. Consequently, Citrangada referred Nrpavahana to Agastya, who then narrated the Vidyas as revealed to Indra by Brahma who had received these from Siva through Visnu. Thus the topics of this Purana have been introduced in chap. 1. The contents of the remaining chapters are as follows:113 Chapter 2. After practising the kamika vidya king Nrpavahana entered the hermitage of Agastya (which is described in twelve verses and in which there were sages 'engaged in the worship of Brahma, Visnu, Siva, Skanda, Uma and Durga') and asked him about the Vidya which can make one immensely powerful and even give him final release. Consequently, after mentioning that this Vidya was handed down to Indra by Brahma who had received it from Visnu, who, in his turn, had received it from Siva, Agastya said that once, being eulogised by Indra for the sake of this Vidya, Brahma advised him to worship Adya Sakti and narrated to him the story of the demon Ghora in the following way. Ghora received from Visnu a boon that none but Siva Devi would be able to defeat him. He returned to his capital Candrasobhapura in Kusa-dvipa, managed the state affairs. according to the principles of the science of politics, and conquered the different quarters with the help of his son Vajradanda. Chapter 3.-Description of Ghora's conquest of the whole earth (with the seven dvipas and oceans); Kala and Vajradanda's invasion of Patala and Rasatala and conquest of the Nagas, Raksasas etc. living there; their return to the earth and approach to Sukra who was living at Udumbarapura in Madhyadesa, where the powerful 'Bhargava' (Sukracarya) resides permanently. Chapter 4. Though discouraged by Sukra, Vajra and 'Vidyadhara' (in Devi-purana 80.1, 92. 14, and 102.2 respectively) tends to show that Nrpavahana was a king of the Vidyadharas. 113 As not even a single edition of the Devi-purana has yet been printed in Devanagari characters, we have felt it necessary to give a somewhat detailed summary of its contents.

40 STUDIES IN THE UPAPURANAS Kala attacked heaven; Indra sought the protection of Visnu. Visnu narrated to Indra the story of the demon Dundubhi who, for his lust for Siva's wife Uma, was burnt to ashes by Siva and from whose ashes was born the demon Ghora who was to reign supreme in Kusa-dvipa and was to be killed by Devi mounted on a lion when this Ghora would try to have her as his wife. Consequently, Visnu and Brhaspati pointed out the futility of war with this demon. Chapter 5.-Brhaspati, who was conversant with the political theories (nayah) promulgated (gitah) by Sambhu, Usanas, Visnu and Brahma, 4 sent Narada to Vajradanda and Kala with the message that the gods recognised their supremacy and allowed them to enjoy heaven until Devi (Vindhyacala-nivasini) and Siva came down to earth and killed them. Visnu and Brhaspati then went to Brahma. Chapter 6. Brahma saluted Visnu with the mention of the names of his ten incarnations (including Krsna, the redrobed Buddha, and the naked Kalkin) 115 and solicited his help. He also reminded him how in times of yore Camunda, being sent by Siva, saved Visnu from the rage of Kalagnirudra and his son Halahala, how Visnu knew this Camunda to be the powerful Sakti able to create, protect and destroy the universe, and how, being eulogised by himself and Visnu, Camunda promised to help them whenever necessary. Chapter 7.-So, Brahma advised Visnu to approach Siva. Visn, with Brahma, went to Siva and eulogised him for the destruction of Ghora and Vajradanda. Siva, the left side of whose body was found occupied by his Sakti, consoled Visnu by saying that Para Sakti would incarnate herself on the Vindhyas with a lion as her vahana, and have Brahma, Visnu and Siva as her servants, the four Vedas as her doorkeepers (pratiharyah), and many unmarried female deities (who have been described in verses 23 ff.) as her attendants, 114 The text of verse 2 runs as follows:sambhu-gita naya yasya ausanas ca tatha nijah / madiya brahma-gitas ca vetti yah sa vacaspatih | | 115 Kalkin is described as 'asvarudha', 'kali-dharma-pravartaka', 'digambara-dhara', "sudra-dharma-pravartaka', and 'melccha-varga-kuloccheda'.-Devi-purana 6. 6-7,

THE SAKTA UPAPURANAS 41 that her vahana would be composed of all the gods and goddesses and the worlds and stars etc., and that she would be furnished with a raksa-mantra" of Siva as uttered by Visnu. Siva then sent Devi with the aforesaid vahana and attendants to the Vindhya mountain. Chapter 8. In the meantime Vajradanda wanted to occupy heaven. So, Narada was sent to Kusadvipa for turning Ghora and his wife and subjects from the path of dharma. Narada succeeded in his mission by attracting him to sensual pleasures, 117 and informed him of the young girl (Devi) residing in the Vindhyas. Chapter 9. Ghora forsook the path of dharma and turned a debauch. His wife Candramati¹¹8 gave him good counsels by referring to the policies of statecraft and to the dangers from poison-girls (visa-kanya), naked Jain monks (nagna), washerwomen (rajaki), fisherwomen (Kaivarti) and others, and narrated the story of Sumedhasa (of Kraunca-dvipa) who outraged the modesty of the daughter of a sage of Puskaradvipa and went to hell. Fearing that these counsels might bring Ghora to his senses, Narada muttered the delusive Padamala-vidya¹¹ which Nandikesvara had received from Siva and which could accomplish 108 acts or even many more, if practised with the Vira-vrata. In this connexion it is said that with a view to acquiring the Padamala-vidya Bhargava entered Siva's body, came out unsuccessful through his generative organ, and was consequently known as Sukra; and the method of muttering the words of the Padamala-vidya and the results of performing homa with human flesh (mahamamsa) have been described. (Digression in chapters 10-12). 116 In the Raksa-mantra Siva is called 'mamsasava-vasa-meda-putanadi-rata' and is asked to remove all impediments created by mantra and yantra. 117 Narada referred to the 'vajikarana-yogas' as instructed by Nandikesvara, and to Visnu-tattva, Kama-tattva and Siva-tattva as proclaimed by the sage Kapila to Nandin. He also praised the enjoyment of a young girl of sixteen years of age. 118 Also named as Candravati and Candrabuddhi in Devi-purana 2.51 and 9.9 respectively. 119 This Padamala-vidya, which is undoubtedly Tantric, has Camunda as its presiding deity and is called 'Atharva-pada-dipani'. It is said to have been collected from 'crores of works' and then described in the Mula-tantra.

42 STUDIES IN THE UPAPURANAS Chapter 10. Narada is said to have received the Padamala-vidya from Sanatkumara (son of Brahma), who had received it from Nandikesvara. From this Sanatkumara Narada received instructions on (Saiva) yoga, with which the Vidya was to be accompanied for the sake of favourable results. (Here Saiva yoga¹20 is elaborately described). Chapter 11.-Brahma's receipt of the Aparajita-vidya and the Padamala-vidya from Siva. Introduction of the Padamala-vidya on earth by Dvaipayana who inherited it from Siva successively through Brahma, Angiras, Brhaspati, Surya, Yama, previous Indra, Vasistha, Sarasvata, Tridhaman, Trivrsa, Bharadvaja, Antariksa, Bahvrca, Aruna, Balaja, Krtanjaya and others (including Saktri, Parasara and Jatukarna). Praise of the Aparajita-vidya by narrating the story of the demon Hutagni who, after receiving a boon from Brahma, took his residence in Saka-dvipa, conquered heaven with the help of his son Canda, but was defeated by Visnu, who killed Canda by virtue of a banner of victory which Brahma had received for Visnu from Siva with the help of the Aparajita-vidya and which was resorted to by Durga, Mahadeva, Visnu, Yama, Indra, Agni and other gods as well as by the nine planets. Visnu's giving of this powerful banner to the previous Indra, after the latter had been defeated by a demon named Subala in the Bhautya Manvantara. Praise of the hoisting of this banner by kings. Chapter 12. Method of preparing Indra's banner (Indradvaja) and of hoisting it in Bhadra or Asvina during the Satakratu-mahotsava (i.e. Indra-dhvaja-mahotsava): - 120 In this chapter, Saiva yoga has been described in ten Paricchedas or sections. The main topics of these sections are the following:- description of the nature of Siva after the concepts of Purusa and Prakrti of the Samkhya system; rules to be observed during the practice of Saiva yoga; method of its practice, including the names of asanas (such as Padmaka, Svastika, Sthalika, Jalika, Pithardha, Candradanda and Sarvatobhadra -10. v. 4); realisation of Sabda-brahma and Para-brahma by forsaking the objects of enjoyment; method and results of meditation on omkara; yoga-dvara-darsana; upasarga-jaya; dharana; etc. (Devi-purana 10.i.16 says that a Yogin desiring success is not to devote himself too much to the acquirement of knowledge-atiprasamgo jnanesu na karyah siddhim icchata; Devi-purana 10.vi.8 characterises Sabda-brahma thus: 'antah-sarira-prabhavam udana-preritam ca yat / vag-uccaryam srotra-vrtti sabda-brahma tad ucyate //'; and in Devi-purana 10.x.1 'dharana' has been defined as follows: manaso hrdy avasthanad dharanety abhidhiyate).

THE SAKTA UPAPURANAS 43 Selection of trees (viz., dhava, arjuna, priyangu, udumbara, asvakarna, candana, amra, sala or saka) for the preparation of the pole of the banner; lengths of the main pole and of those representing the daughters of Indra; rites to be performed before cutting down the selected tree; auspicious and inauspicious dreams; method of cutting down the tree; the king's good or evil as indicated by the direction in which the tree falls down; preparation of the pole out of the tree; the bringing of the pole to the place of worship in a procession attended with songs of public women and beat of drums; the covering of the pole with pieces of cloth of different hues and furnishing it with a flag; erection of the pole by means of machines, with the citation of mantras of Visnu, Siva and Indra; worship of the banner. Performance of this ceremony of flag-hoisting by kings and their subjects in towns, cities and khetas. Chapter 13. Narada deluded Ghora and his ministers and others by muttering the Padamala-vidya, so much so that Ghora's queen became devoted to the Digambaras and gave herself up to the performance of Vratas and all other religious duties of the Pasandas. With the intention of enjoying mountain girls Ghora started with his army in an auspicious moment and experienced various kinds of bad omens including the sight of a red-robed person. At Narada's advice Ghora went to the Vindhya mountain which was watered by the Narmada and inhabited by learned sages as well as by various Mleccha tribes such as Varvaras, Pulindas, Savaras, Tankas and Kapalins 121 who were descended from the son (Nisada) of Vena 122, and who lived on roots and fruits, put on barks, and decorated their dark bodies with gunja fruits, garlands, feathers of parrots, and leaves worn 121 Devi-purana 13. 51 b-52 arenu-(vena-?)sambhava-santana-sambhuta varvaravarah || pulindah savaratamka-kapali-mleccha-jatayah || From the reference to the Kapalins in this verse it seems that the Kapalikas were originally a wild tribe (like the head-hunters of Assam?). 122 In Devi-purana 13.51 b-52 a the wild tribes are called 'renu-sambhava-santana-sambhutah'. But the reading 'renu-'must be a mistake for 'vena-'. We know from the Visnu (I. 13) and other Puranas that Nisada, who was the predecessor of the Nisadas, Savaras etc. living in the Vindhyas, was born of Vena's thigh.

44 STUDIES IN THE UPAPURANAS round their waists. In this mountain Ghora saw Devi in the form of a girl and was bent on having her. He sent Durmukha, but the latter was killed by Vijaya. Ghora saw various kinds of bad dreams towards the close of the night. Chapter 14.-Description of the fight of the demons with Devi's attendants Jaya, Vijaya, Ajita and Aparajita; and Jaya's killing of the demons Kala, Bhairava and others. Chapter 15. Indra's eulogy of Jaya for killing Kala and other demons. Jaya's fight with and killing of the demons Vajradanda and Yamantaka. Chapter 16. Susena praised Ghora as a paramount and unrivalled ruler, and consequently Ghora prepared for war against Devi. In the meantime Narada eulogised Devi who, though originating from Rudra, creates and protects Brahma, Visnu and Siva, and who is known under different names, viz., Durga, Sakambhari, Gauri, Vindhyavasini, Katyayani, Kausiki, Kaitabhesvari, Mahadevi, Mahabhaga, Mahasveta, Aparna, Kapala, Ekapatala, Savitri, Gayatri, Laksmi, Sarvatobhadra, Yoganidra, Jaya, Vijaya, Ajita, Aparajita, Bhadrakali, Camunda, Candi, Candini, Parvati, Daksayani, Narayani, Revati, Simhavahini, Vedamata, Sarasvati, Bhogavati, Sati, Satyavati, Bhima, Dhumra, Ambika etc. Devi was pleased to manifest herself by mounting a lion, carrying carman, sword, bow, naraca, sula, khatvanga, tusk, lute, bell etc. in her hands, and wearing a leopardskin, and promised to kill Ghora. Chapter 17. Finding that Ghora oppressed all the gods and was about to kill even the female deities, Siva eulogised Devi with the Mahadandaka hymn, in which he called her 'Siva-vaktrodbhava', 'pramatha-laksa-vrta', 'hema-varna', 'smasana-priya', 'kumari', 'Samkhya-yogodbhava', Vetali, Kapalini, Bhadrakali, Mahakali, Kali, 'maharuksa-ghantaravodgiti-karnotsava' etc. [In this hymn Devi is identified with Brahmi, Kaumari, Mahendri, Mahesvari, Vaisnavi, Varahi, Sarasvati, Arundhati, Jahnavi and other female deities and is said to dwell in Hemakuta, Mahendra, Himadri, Vindhya, Sahya and Srigiri mountains as well as on mountain-tops and in caves.] -

THE SAKTA UPAPURANAS 45 Chapter 18. As a result of this eulogy, Devi waged war against the demons, and Jaya killed Susena. Chapter 19.-Destruction of demons by Jaya, Ajita and Aparajita. Chapter 20. Fate and perseverance as moulding the destiny of men. Ghora's assumption of the form of a buffalo (mahisa) and fight with Devi. Devi's killing of Ghora as well as of those demons who grew out of his body. Chapter 21. After the fall of Mahisa (i.e. Ghora) the gods worshipped Devi with the offer of buffaloes and he-goats. Praise of the worship of Devi (Mahisasura-mardini) with the offer, at midnight, of these animals on the Suklastami and Sukla-navami Tithis of Asvina. (These Tithis are called Mahastami and Mahanavami respectively). Chapter 22. Description of the method of Devi-worship as performed annually in a golden, silvern, earthen or wooden image or in a sword or trident by the members of the four castes as well as by women. This method consists of various operations, of which the following are noteworthy, viz., beginning of Devi-worship on Asvina-sukla-pratipad; performance of homa and Siva-worship, and feeding of virgin girls daily; construction of nine houses or of at least one house for Devi-worship, fasting, and animal sacrifice on the Astami Tithi; severing of the figure of an enemy prepared with powdered rice, 123 and performance of Devi's carfestival (ratha-yatra) on the Mahanavami Tithi. Chapter 23. Results of worshipping Devi for nine days from Asvina-sukla-pratipad to Navami, and of having on the banner the figure of Devi mounted on a lion, a monkey, a bull, a swan, a peacock, Garuda, a buffalo, an elephant, a lotus, or a dead body. Chapter 24.-Determination of different kinds of samkranti (viz., manda, mandakini etc.) and of their parts in which religious rites can be performed. Good effects of these 13 Devi-purana 22.16- .satrum krtva tu pistajam / khadgena ghatayitva tu dadyat skanda-visakhayoh || The printed ed. of the Devi-purana wrongly reads 'sakram' for 'satrum' in the first line.

46 STUDIES IN THE UPAPURANAS samkrantis to the members of the four castes and to thieves, vintners, Candalas, Pukkasas, Pisacas, actors, dancers and others. Chapter 25. Benefits of vasu-dhara-dana¹24 to Devi in fire daily, fortnightly, monthly or at longer intervals. Chapter 26. Method of vasu-dhara-dana:-Preparation of the sacrificial pit (kunda) with its girdle (mekhala); construction of the necessary implements such as a pot (made of gold, silver or brass) for containing ghee, a ladle with a hole, and so on; performance of homa meant for the nine planets, the ten Lokapalas, and the Matrs with the citation of the mula-mantra, hrdaya-mantra, astra-mantra etc. or according to the direction of the Atharva-veda; citation of a hymn of praise to Agni; feeding of Brahmins, actors, dancers, harlots, virgin girls, widows, poor men and others. Chapter 27. The sacrificer's good or evil as known from the colour, sound, smell or general appearance of the fire in which homa is performed. Results of the use of different objects for the performance of homa. The sacrificer's abstinence from speaking with Pasandas, deformed persons, and others. The different occasions (viz., Asvina-maha- stami, Asvina-mahanavami, Karttiki purnima, Skandasasthi, Sakra-dhvajotsava, Pusya-snana etc.) on which the vasu-dhara should be given. Chapter 28.-Devi-worship in temples, pasture-lands, mountains etc.; feeding of virgin girls; offer of gold, clothes etc. to those Brahmins who are devotees of Devi; and reading of works on Devi-mahatmya.-All these are to be performed in Devi-vratas, Vasu-dhara-dana, Dhvajocchraya etc. Chapter 29.-Origin of the visible world as well as of gods, Vidyas, mantras, Tantras, mudras etc. from Devi who is the Para Sakti. Chapter 30.-Results of worshipping the images of Devi 124 'Vasu' means 'ghee'. The word 'vasu-dhara' is explained in Devi-purana 27.21 as follows:vasu dravyam ghrtam ajyam amrtam havih kamikam / tasya dhara sada deya vasor dhara hi sa mata / /

THE SAKTA UPAPURANAS 47 made of different materials, viz., stone, sapphire, gold, silver, copper, earth, crystal, pearls etc. Chapter 31. Method of performing Devi's Ratha-yatra (car-festival), in which Devi is to be worshipped with the mention of her different names (viz., Durga, Katyayani, Vindhyavasini etc.) and women are to utter auspicious words and scatter flowers, durva, aksata etc. while following the procession. General amnesty to be observed on the occasion of Devi's Ratha-yatra. Chapter 32-Results of the worship, on different occasions and in different places (such as river-banks, forts, burning grounds etc.), of Devi, who is said to be always present in the Ganges, the Narmada, the Vindhya mountain, Ujjayini, Arvuda, Himalaya, Nisadha and Drona. Directions for the construction of images of Devi having ten, two or twentyeight hands and seated on a lion. (In these images Devi should be exhibited as surrounding, with a snake, the body of the demon issuing out of the buffalo beheaded by herself). Consecration of these images by those (Brahmins) who are versed in Devi-sastras and have knowledge of the Matrmandala (the group of Matrs) and of works on bhuta-tantra, graha, vyala, and garuda-vidya. Chapter 33. Sukra's eulogy of Siva on the Kailasa mountain by mentioning the manifold benefits of Sivaworship as well as of acquiring knowledge of Siva and becoming a wandering mendicant according to the rules of the Saivas. Siva's description of the method of performing the Devi-vrata (also called Durga-vrata) on the Suklastami Tithis of different months beginning with Sravana. (In this Vrata, which is said to have been described in the Tantra, Devi is to be bathed and worshipped with the mention of her different names, viz., Katyayani, Uma, Narayani, Jaya, Vijaya, Ajita, Aparajita, Carcika, Mahisaghni etc., various articles are to be offered and incense of different kinds including Turuska-dhupa is to be burnt, Brahmins and virgin girls are to be fed, and the worshipper is to avoid all contact with Pasandas). [In verse 4 there is mention of a 'stotra' recited by a

48 STUDIES IN THE UPAPURANAS Gana named Puspadanta, and verses 15 and 19 describe females wearing 'cinamsuka'.] Chapter. 34. Results of decorating and cleansing the house in which Devi is worshipped and of honouring those persons who are versed in Devi-sastra. Story of a fisherman (Kaivarta) who attained the results of dhvaja-dana by suspending his fishing net from a tree before entering a Devi-temple on the Vindhya mountain. Chapter 35. Method of preparing flags and furnishing Devi-temples with these. (During dhvaja-dana Devi is to be worshipped, and Brahmins and virgin girls are to be fed). Chapter 36. A Devi-stotra uttered by Siva at the request of Sukra. (In this stotra Devi is called 'nada-bindu-svarupini', 'siva-sakti-para-svarupini', etc.). Chapter 37. The reasons for Devi's different names, viz., Gauri, Narayani, Durga, Katyayani, Raudri, Vindhyavasini, Jayanti, Ajita, Vijaya, Aparajita, Mahisaghni, Kali, Kapali, Kapalini, Camunda, Nanda, Kausiki, Kaitabhesvari, Mahasveta, Mahabhaga, Bhavani, Jyestha, Brahmacarini, Aparna, Ekaparna, Patala, Tara, Vama, Citi, Revati, Ekanamsa, Laksmi, Kandavarini, Maya, Ganga, Yamuna, Rajani, Himsa, Bala, Sankuveni, and many others. Though one, Devi appears as many through guna. Chapter 38. Enumeration of Devi's different forms which are worshipped in different places.-Devi is worshipped as Mangala in the tract of land lying between the Vindhya and the Malaya mountain, as Jayanti on the sea-coast between the Vindhyas and Kuruksetra, as Nanda between Kuruksetra and the Himalayas, as Kalika, Tara and Uma in the mountains, as Bhairavi in Kiskindhya etc., as Rudrani in Kusasthali, as Bhadrakali in Jalandhara, as Mahalaksmi in the Kola mountain, as Kala-ratri in the Sahya mountain, as Amba in Gandhamadana, as Ujjani in Ujjayani, 125 and as Bhadrakalika in Vaideha. 125 This 'Ujjayani' (also called 'Ujjayini') is the same as the ancient city, popularly known as 'Ujani", which comprised the modern villages of Kogram, Mangalkot (Sanskrit 'Mangalakosthaka') and Aral situated on the bank of the river Ajaya in Katwa subdivision in the district of Burdwan in West Bengal. It has been mentioned, sometimes ---

THE SAKTA UPAPURANAS 49 Chapter 39.-Names of holy places in which Devi was worshipped in her different forms 126 by different gods, sages and kings (including Jamadagnya, Bhauma's son, Vibhisana and Rajasena) viz., Puskara, Himalaya, Naimisa, Malayaparvata, Kiskindhya-parvata, Kasikasrama, Veda-parvata, Kamakhya-giri-kandara 127, Sarasvati-tata, Purva-sindhu, Lanka, Kailasa, Kanvasrama, Dharmaranya, Mahakala, Kotitirtha, Bhadravata, Kaveri-sangama, Gokarna, Dan- daka, Gandaki-sangama, Mahasona, Mahodaya, Mundipitha, Malaya-parvata, Saka-dvipa, Kusa-dvipa, Krauncadvipa etc. Praise of the power of mantras by narrating the story of the demon Bala, who, being deluded by Visnu by means of the Mohini Vidya received from Siva, offered his body for the fulfilment of the sacrifice of gods. Story of the death of the demons Subala (son of Bala), Dundubhi and others at the hand of the Parama Sakti (also called Mahavidya) who took the form of an extremely old and emaciated woman named Ksemankari and was accompanied by the eight Vidyas (asta-vidya-samanvita). It is said that these Vidyas, divided hundredfold, are worshipped in different forms by kings and women according to the 'Daksinacara marga' and by Pulindas, Savaras and others according to the 'Vamacara marga', and that they are specially popular among harlots, cowherdesses, Tudus, Hunas and Khasas and in Himavat-pitha (Tibet ?), Jalandhara, Vidisa, Mahodaya, Varendra, Radha, Kosala, Bhottadesa, Kamakhya, Kiskindhya, Malaya, Kolu, Kanci, Hastinapura and under its popular name 'Ujani", in the works of Bengal as one of the mahapithas created by the fall of the different parts of Sati's body severed by Visnu, and Devi is said to reside here under the name of 'Ujani' or 'Ujjani'. For mention of this sacred city see Brhaddharma-purana I. 14. 14, in which Devi says: "ujjayinyam tatha puryam pitham mangalakosthakam / subha mangala-candy-akhya yatraham vara-dayini / /" See also Trikandasesa, Tantra-cudamani, Pitha-nirnaya (alias Mahapitha-nirnaya), Siva-carita, Bharata-candra's Annada-mangala (Pitha-mala, verse 22), and so on. For an excellent account of the Sakta Pithas (including the said 'Ujjayani' or 'Ujani' of Bengal) see Journal of the Asiatic Society of Bengal, XIV, 1948, pp. 1-108. 120 viz., Nanda, Siva, Amba, Kamakhya, Mahesvari, Kotimunda, Maharaja, Khandamunda, Aghora, Kalika, Vijaya, Canda, Dhutima (?), etc. 137 'yajed bhaumatmajo devim kamakhye giri-kandare.-Verse 6 b.

50 STUDIES IN THE UPAPURANAS Ujjayini. 128 Offer of fish, meat and wine in the worship of Ksemankari is prescribed with the mention of Damani, Padamala, Srighosa-vajrasasana and Pratyangira-devi, and the priest, who is not to be a naisthika brahmacarin, is required to worship the goddess after the kula-marga. The giving of one's own self, wife and the entire property to the consecrator of Devi's image. Praise of the Vidyas, which are said to be sixty-four in number. Chaps. 40-41.-Story of Mahadharmasura's war with the gods, and his death at the hand of Devi invoked by Brahma. Chapter 42. Gods' eulogy of Devi by mentioning the names of places where she always remains in different forms, viz., Jalandhara Pitha, Malaya, Sahya, Vindhya, Himalaya, Citragopa, Narakala (v. 1. 'nave kale' for 'narakale'), Nicaksaparvata (Nicakhya-parvata ?), Lanka, Udra-desa, Strirajya, Kasikavana, Kamarupa, Kanci, Campa, Vaidisa, Varendra, Uddiyana, Manaksa, Kusasthala, Cola, Simhala, Venudanda, Kanyakubja, Nava-durga-sthala, 129 etc. etc. At Vaidisa Devi remains in the middle, seated on a 'simhasana', and is widely known as 'urdhvajayavaha mahakali'. At Jambukanatha she is famous as Bhadrakali. Chapter 43.-Praise of the Puspakhya-vidya, with the narration of the story of Amayasura who was attracted by the physical beauty of Sumati, beloved of Gajanana, and whom Gajanana killed with the help of his friend Parasurama. (In this chapter there are names of more than forty-three kinds of vyuhas, 130 into which Gajanana arranged Parasurama's army). Chapter 44. Gajanana went to his own residence in the Malavya mountain, and Parasurama established Kalika 128 Devi-purana 39. 141-145. 129 The text reads: "simhale venudande ca kanyakubje'tha vaidise / nava-durga-sthale krtva trimunda tatra kirtita | /" (Devi-purana 42.10). According to Devi-purana 44. 4-5, the nine Durgas were established at Mahodaya. 130 Of these, seven (named as Kaksa, Paksa, Urasya, Danda, Abhoga, Mandana and Samghata) are 'prakrta', and the rest (viz., Pradadha, Drdha, Kosadya, etc.) are not so (prakrtai rahitan).-See verses 22-34. Bri

THE SAKTA UPAPURANAS 51 at Ayodhya. Results of the worship of the nine Durgas (nava-durga) at Mahodaya and of the other forms (viz., Kamika, Ambika, Tara, Vindhyatavi, etc.) of Devi situated at different places (viz., Malaya-parvata, Mandaksa-parvata, Candra-parvata, Kiskindhya-parvata, etc.). Chapter 45. Selection of proper day and time for the performance of sacrifices to the planets and the deities such as Skanda, Bhanu, Visnu, Uma, Vinayaka and others. Results of such sacrifices. Chapter 46. Origin of the universe and gods from Devi. Names of the eleven Rudras, twelve Adityas, eight Vasus, two Asvins, ten Visvedevas, fourteen Manus, fourteen Indras, seven Maruts, three Agnis, five kinds of years, sixty kinds of samvatsaras, and so on. Determination of the nature of influence of planets on different parts of India according to kurma-vibhaga. (In connexion with kurma-vibhaga the names of Tirabhukti, Anga, Vanga, Samatata, Vardhamana, Kamarupa, Nepala etc. have been mentioned, and Narmada has been called the centre of the earth-mahi-madhyam). Chapter 47.-Division of time into 'truti', 'dina', etc. Names of planets, and their classification into 'mandalagraha' (viz., Surya and Candra), 'chaya-graha' (viz., Rahu) and 'tara-graha'. The parentage, relative position, and dimensions of Planets and Stars. Names of the seven regions (loka) and their presiding deities. Position of these regions. Chapter 48.-Determination of the two kinds of purnima (viz., raka and anumati) and two kinds of amavasya (viz., sinivali and kuhu). Description of the condition of the moon, whose fifteen kalas are eaten up by gods, Pitrs and others during the Amavasya. Chapter 49.-Solar and lunar eclipses, which are caused by Rahu in the forms of shadows of the moon and the earth respectively. Chapter 50.-Classification of the sixty forms of Devi into three groups: (1) twenty sattvika forms constituting the first grade, viz. Mangala, Vijaya, Bhadra, Siva, Santi, Dhrti, Ksama, Rddhi, Vrddhi, Unnati, Siddhi, Tusti, Pusti, Sri,

52 STUDIES IN THE UPAPURANAS Uma, Dipti, Kanti, Yasa, Laksmi and Isvari; (2) twenty rajasa forms constituting the second grade, viz., Brahmi, Jayavati, Sakti, Ajita, Aparajita, Jayanti, Manasi, Maya, Diti, Sveta, Vimohini, Saranya, Kausiki, Gauri, Vimala, Rati, Lalasa, Arundhati, Kriya and Durga; (3) twenty tamasa forms of the third grade, viz., Kali, Raudri, Kapali, Ghantakarna, Mayurika, Bahurupa, Surupa, Trinetra, Ripuha, Ambika, Mahesvari, Kumari, Vaisnavi, Sura-pujita, Vaivasvati, Ghora, Karali, Vikata, Aditi and Carcika. Description and measurements of the images of Mahadevi to be worshipped in common houses or in palaces. Benefits of constructing images with gold, silver or copper. Directions regarding the construction of temples for Deviworship; decoration of these temples with paintings, flags etc. Method of consecration and worship of the images of Devi. Description of the images of each of the above-mentioned sixty forms 131 of Devi, with the mention of the particular years in which these forms are to be worshipped. Description of the images of Ganesa, Matrs, eight Vidyas, six Seasons, eleven Rudras, Mahalaksmi, Kusmanda (seated on a human corpse), and Varvara. Worship, in images or mandalas, of the Sun in different forms under different rasis; materials with which these images are to be made. Signs to represent the deities. Worship of Siva, Durga, Surya, Ganesa, Brahma and Hari in mandalas 132. Ringing of bells during worship. Method of performance of homa and Devi-worship. Chapter 51. Aman, who is versed in Devi-sastras and has knowledge of the Matr-mandala but is not a naisthika brahmacarin, is eligible for performing Devi-worship, no matter whether he is a Brahmin, a Ksatriya, a Vaisya or a Sudra. A Devi-worshipper is to worship Siva (according to Siva-siddhantas), Surya (according to the 'Matharokta 131 Of these forms, Durga is to be seated on an intoxicated quarter-elephant, and Vimala, Bahurupa and Carcika on human corpses, and Kumari, Karali and Vikata are said to be fond of honey, wine and meat. 132 Devi-purana 50. iv. 55 ff.

THE SAKTA UPAPURANAS 53 vidhi'), Brahma (according to the Vedic method) and Visnu. Necessity of utensils in worship, and the results of making these utensils with different materials (such as gold, silver, copper, tin, lead, iron, stone, wood etc.) under different conditions. Chaps. 52-54.-Methods of performing graha-yaga and Matrka-puja in mandalas in the different months of the year. Chapter 55. Enumeration of the different kinds of divya, antariksa and bhauma Utpatas (or Adbhutas), and the means (viz., Matrka-puja, graha-puja, Surya-puja, laksa-homa, koti-homa, dana, Brahmanadi-bhojana, etc.) of preventing their bad effects. Chapter 56.-Detailed description of the method of performing homa. Chapter 57. Description of a form of Devi which, when worshipped according to the prescribed rules along with Matrs, Karttika, Ganesa, Brahma, Visnu, Siva, Lokapalas and others, gives immediate results. Identification of Devi with Siva. Chapter 58.--Story of king Bhagya who worshipped Hari in a mandala on the Bhagya-dvadasi Tithi and also HariHara and Uma-Mahesvara in images, and thus attained great fortune in the Surottama-kalpa. Chapter 59. Method and result of worshipping Devi in the different months of the year. (In this chapter, mention is made of the worship of Brahma, Agni, Uma, Ganesa, Naga, Skanda, Surya, Matrs, Yama, Siva, Visnu, Kama, Rudra, Indra and others in Sravana, of Devi-worship and Devi-ratha-yatra in Asvina- suklastami, of 'dipa-dana' and 'gajasvadi-nirajana' in Karttika-krsna-caturdasi and Amavasya, of worship of Candika in Phalguna, and so on). Chapter 60.-Method and results of performing vrsotsarga and 'go-vivaha'133 along with Devi-worship. Results of 133 In this festival, which is to be performed in Magha or Phalguna, four, two or one heifer is to be duly married to a bull of the 'nilavrsa' type,

54 STUDIES IN THE UPAPURANAS repairing old temples, and of worshipping Uma-Samkara, Hari-Hara, Ardha-narisvara, and the Trimurti. Chapter 61. Proper Tithis for, and results of, the worship of Brahma, Uma, Siva, Agni, Devi-and-Samkara (who are to be placed in a swing), Ganesa, Nagas, Skanda, Bhaskara, Matrs, Devi Mahisamardini, Dharmaraja, Vrsa, Visnu, Kamadeva-and-Rati, Siva, and Indra-and-Saci. Chapter 62. Results of worshipping Siva with different kinds of flowers in different months of the year. Chapter 63. Names of sixty-eight holy places in which Siva remains present under sixty-eight names. (Among the names of places there is mention of Puskara, Ujjayini, Gokarna, Rudra-mahalaya, Devadaruvana, Turanda, Sukalanta, Ekamra, Dipta-cakresvara, Kasmira, Nepala, Gangasagara, etc.). Chapter 64. Description of Goratna-vrata, in which a cow and a bull are to be decorated and given to a devotee of Siva after worshipping Siva and Uma. Chapter 65. Method of worshipping Siva-and-Uma on the Suklatrtiya Tithi after painting their figures with 'gorocana', saffron etc. on two pieces of cloth. Devi-worship in Pradipta-navami-vrata and Guggula-vrata. Method of performing Pusya-snana (which is meant for kings): selection of proper place (viz., Himalaya, Sahya-parvata, Vindhyacala, sea-coast where ships can remain safely, and so on); description of dreams which indicate good or evil 134; drawing of mandalas with padmas in them in the prescribed manner; placing of deities in these mandalas; worship of these deities; use of wine and meat in the worship of Pisacas, demons and Raksasas. Chapter 66.-Description of Pusya-snana-vidhi continued: account of the origin of pitchers during the churning of the ocean; names of the nine pitchers in which the water for the king's bath is to be kept and which are said to contain the A 'nila-vrsa' has been defined as follows: lohito yas tu varnena samkha-varna-mukho vrsah / langula-sirasas caiva sa vai nila-vrsah smrtah // Devi-purana 60. 7 b-8 a. 134 A Nirgrantha, seen in a dream, indicates evil.---Devi-purana 65.52.

THE SAKTA UPAPURANAS 55 seven oceans, seven continents, all the planets and stars, all the Kula-parvatas, all the holy rivers (such as the Ganga, Sarasvati, Sindhu, Subhaga etc.), and all the holy places (including Eka-hamsa, Pundarika, Gangasagara etc.); names of the presiding deities of these pitchers; placing of these pitchers in different directions. Chapter 67. Description of Pusya-snana-vidhi continued: placing of auspicious articles in the pitchers; decoration of a throne and a number of beds (the latter being meant for the royal visitors); bathing of the king by Brahmins; worship of different deities including the Matrs; donations to Brahmins and others; amnesty to prisoners. Chapter 68. Special baths at particular places for the accomplishment of particular objects (such as removal of barrenness of women). Places favourable for the success of mantras. Proper places for drawing mandalas. Chapter 69.-Description of Vinayaka-yaga (worship of Vinayaka) to be performed by kings for averting the effects of bad dreams. (In this Gana-yaga, Surya and Ambika also are to be worshipped). Chapter 70.-Description of the raksa-mantra called Vinayaka-kavaca, and the result of wearing it after writing it in a mandala on a gold or silver plate, or on a piece of cloth, or on a 'bhurja-patra'. Chapter 71. Power of raksa headed by the bijas of particular gods and written on planks of banian trees or on copper-plates or worshipped with flowers of particular colours. (The final verse of this chapter mentions Madhusudana with four Vyuhas-catur-vyuha-samayuktah . . madhusudanah). Chapter 72. Construction of forts (durga): first construction of a fort by Visvakarman at the instance of Brahma; selection of a suitable site for the construction of a fort; mutual relation between the different units of measurement of distance, viz., trasarenu, liksa, rajasarsapa, gaura-sarsapa, yava, anguli, pradesa, tala, gokarna, vitasti, aratni, hasta, kisku, dhanus, danda, nalika, yuga, krosa, gavyuti and yojana; protection of forts by means of ramparts, trenches, shrubs

56 STUDIES IN THE UPAPURANAS etc.; different kinds of forts, viz., parvata (or giri-durga), audaka (or jala-durga), dhanvana (or maru-durga) and vanaja (or vana-durga); further classification of each of these four kinds of forts into two, viz., (1) guha-durga and prantara-durga, (2) nadi-durga (also called antardvipa-durga) and sthaladurga, (3) nirudaka-durga and irina-durga, and (4) khanjanadurga and stamba-gahana-durga; comparative importance of these forts to kings; arrangements for the maintenance of these forts; furnishing the forts with sufficient food, arms and ammunitions, machines etc.; arrangements for the protection of forts; restrictions on the admittance of persons into the forts; maintenance of physicians, carpenters, astrologers, visa-vaidyas, bhuta-vaidyas, garudikas and others 155 in the forts; directions regarding the construction of gates (gopura), at the sides of which there should be images of Devi Mahisamardini, Ganesa, Kuvera or Brahma and which must be named after the different names of Devi. Construction of cities (pura): shapes of the cities (which should be square, circular, triangular, or oblong); different names of cities according to the difference in their dimensions, viz., Aindra, Vaisnava, Samkara, Brahma and Sarvabhaumika; position of the residence of the king and the houses. of the members of the four castes in the city; residences of people of lower castes to be located outside the city for preventing intermixture of castes; characteristics of different kinds of nagaras (viz., Sarvatobhadra, Rucaka, Svastika, Chinnakarna, Vinasa, Duhsthita, Krsa and Durbala). Measurement and position of kheta and grama. Names and measurements of different kinds of roads, viz., simamarga, raja-patha, sakha-rathya uparathya, alpa-rathya and jangha-patha. Chapter 73.-Directions for the construction of adhodurga and krtrima-durga which constitute the vijayakhya pura. Location of residences of different classes of people in these types of forts. Persons (viz., Sudras, actors, sons of prostitutes, and others) who are not to be entrusted with the charge 135 visa bhutopaharams ca garudika dikams tatha.--Devi-purana 72.110, ✓ * }

THE SAKTA UPAPURANAS 57 of forts. Determination of auspicious time for the construction of these. Worship of Siva, Durga, Matrs, planets, Vinayaka, and others on this occasion. Chapter 74. Names of rivers, 136 places and forests which are deemed as highly sacred during eclipses and on other occasions. Chapter 75.-Names of holy places, mountains and rivers where the performance of dhara-homa becomes beneficial. Chapter 76.-Names of some holy places. Praise of Kapota-tirtha, and the benefits of bathing Siva-lingas, Devi, Matrs, Surya, Narayana and others with the water of the pool (kunda) at Kapota-tirtha. Story of a pigeon (kapota) which was reborn as sage Suka as a result of its death in this pool. Chapter 77. Praise of the Kapota-kunda; and the results of worshipping Siva and Devi at Kapota-tirtha with the use of Tantric mantras and the performance of sixteen kinds of mudras, viz., yoni-mudra, linga-mudra, vyapini mudra, chatra-mudra, etc. Chapter 78. Praise of observance of Krsnastami-vrata in the different months of the year and of observing Naktavrata and worshipping Siva under different names on this occasion. (In this chapter, as also in a few others, there is mention of gomedha and purusamedha sacrifices). Chapter 79. Results of the worship of Devi's twelve forms (viz., Gauri, Kali, Uma, Bhadra, Durga, Kanti, Sarasvati, Mangala, Vaisnavi, Laksmi, Siva and Narayani), Ardhanarisvara, Uma-Samkara, Hari-Hara, and Narayana. Method and results of performing Uma-Mahesvara-vrata, Visnu-Samkara-vrata, Laksmi-Parna-vrata, BrahmaSavitri-vrata, and Candra-Rohini-vrata. (In all these Vratas the mediums of worship are the images or altars). Praise of sweeping temples of Devi by narrating the story of Kunkuma, chief queen of Candramindra, who was a kite in her previous birth but got rid of her kitehood by 186 Ganga, Narmada, Yamuna, Vetravati, Kausiki, Vindhya, Sarasvati, Candrabhaga, Tapi, Devika, Bhavika, Sindhu etc.

58 STUDIES IN THE UPAPURANAS removing the dusts from a Devi-temple in her attempt to take the food offered to Devi. tion Chapter 80.-Devi's presence in all things. Renunciacompulsory for Devi-worship. Powers of Mahamaya. Chapter 81. Description of Kalagnirudra (who is not different from Kalika) as well as of his residence (called Kalagnirudrapura) which is surrounded by flames. Names of the Rudras who surround Kalagnirudra and bring about the destruction of the world. Chapter 82.-Names of some of the numerous hells which are situated above the flames surrounding Kalapura. Nature of sins which cause one to visit these hells. Names and description of the seven Patalas; the residents of these Patalas. Description of the eighth Patala which is not found mentioned in any Agama or any other work and in which there is Ardhan-arisvara. Story of the rebirths of Taskara-vallabha, son of a Brahmin named Kuladeva of Maharastra who, in his attempt to commit theft in a Devitemple, kept a lamp burning there for some time. Chaps. 83-87.-Story of the origin of the demon Ruru from the mouth of Karttikeya's peacock and his war with the gods and death at the hands of the Saktis (also called Matrs) issuing from Brahma, Siva, Visnu, Karttikeya, Yama, Indra and Agni. [In these chapters, there are eulogies of Brahma, Visnu, Siva-and-Devi, and the Matrs, Siva has been praised, the connexion between Siva and Devi has been set forth, the forms of the different Matrs have been described, and the names of different weapons used in war have been given. Chapter 86 ends with a song (in four verses) sung by Visnu, which seems to be a prototype of those in Jayadeva's Gitagovinda. This song runs as follows: vikasita-karnikara-kamalotpala-laulajam mukuta-nighrstangam sasi-pannaga-vicitra-tanum / tridasa-vilasini-vadana-pankaja-gita-ravam dhruvam iha tanu namami candesa-sivam sirasa dhruvakam | |

THE SAKTA UPAPURANAS pranata-jana-hitam asura-bala-haram tridasadhipate candesvara namo 'stu sada/ 59 == giri-duhitr-pate vara-vrsa-gate namadhvam pasupate/ and so on (the fourth verse ending with the words 'jaya jaya devam canda-sivam')]. Chapter 88. Praise of the worship of Matrs even by Antyajas, Pasandas, Bauddhas and Garudavadins according to the Sivagamas, Garuda-tantra, Bhuta-tantra or Kalatantra. Sakti as the source of the universe. Chapter 89.-Method of worshipping Sarva-mangala in a mandala from Krsnastami to Sukla-navami in the month of Asvina with the mention of her different names (viz., Mangala, Bhairavi, Durga, Kanya, Kapali, Kaitabhesvari, Kali, Carca etc.), sacrifice of animals, performance of Ratha-yatra, and feeding of virgin girls. (This worship is open to the members of all the four castes as well as to women and children). Chapter 90.-The different places (viz., pura, grama, nagara, khetaka, hatta, Gamga-tira, Vindhya-parvata, Vedaparvata, Srisaila, Kiskindhya-parvata etc.) where the Matrs should be worshipped. Directions for the construction of their images and temples. Furnishing these temples with wells, tanks, gardens, harlots (vesya), bells, mirrors, ghatikayantra (for measurement of time), etc. Method of consecration of images (with the citation of Vedic verses and songs of females) and their worship (which is to be followed by honour shown to females, virgin girls and Brahmins). Chapter 91.-Devi to be worshipped even by Sudras and women. Results of Devi-worship in an earthen image and of giving Vidya (i.e. books), umbrellas etc. to Devi. Praise of Vidya. Method of Vidya-dana [which consists of the following operations: books, viz., Siddhanta-sastra, Moksa-sastra, Veda, Vedanga, Itihasa, Garuda-tantra, Bala-tantra, Bhutatantra, Bhairava-tantra, Jyotih-sastra, Vaidya-sastra, Kala- sastra, Kavya, Agama etc., that should be given to Devi; preparation of palm-leaves, placing them in a leathern jacket (? carmanam samkutikrte-verse 57), and furnishing

60 STUDIES IN THE UPAPURANAS them with two painted pieces of planks to serve as covers as well as with a piece of black or red thread having a coloured end for tying the leaves and the planks with; drawing of a mandala with a yantra in it; placing of the palm-leaves as well as of a picture of the presiding deity in the yantra; worship of these with flowers and other things; honour to be shown to the scribe, who is to be 'sastra-paraga', 'chandolaksana-tattvajna', 'sat-kavi' and 'madhura-svara'; method of copying books with the use of Nandi-nagara script in the case of sacred books; worship of the copy thus made, after placing it on a specially made stool furnished with bells, mirrors, painted pieces of cloth, and so on; worship of Matrs and other gods and goddesses on this occasion; taking of the manuscript in a car to a temple of Siva or Matrs, and giving it to a learned teacher; merrymakings on this occasion]. Praise of giving clothes, umbrellas, oil, lamps, food, pens, ink, ink-pots, sharp knives, covers of books, stools for books, etc. to students and scholars. Chapter 92. Devi's constant presence in the Vindhyas (where she manifested herself for killing the demon Ghora) and as Nanda in the Himalayas. Praise of worshipping Devi with mantra, dravya, kriya and dhyana in these places. Chapter 93. Siva's praise of the Nanda-tirtha (in the Himalaya) as the best of all holy places, the other three sacred places on the Himalaya being Bhairava, Kedara and Rudra-mahalaya. Glorification of Nanda who remains always present at Nanda-tirtha with virgin girls. Months (Asadha, Sravana and Bhadra) best suited for pilgrimage to Nanda-tirtha. Description of Nanda-puri (residence or temple of Nanda) and its environments. Mediums of worship of Nanda, viz., Siva-linga, sthandila, pustaka, paduka, pratima, citra, trisikha, khadga, jala, agni, and hrdaya. Characteristics of 'anivartitadhikara' and 'savikara' Lingas. Praise of Lingas established by Sukra and others, and denouncement of those established by Kaca and others with Vedic Mantras. Characteristics of the other mediums of worship (viz., pratima, paduka, citra etc.). Method of worshipping Nanda in these mediums. Rules to be observed

THE SAKTA UPAPURANAS 61 by the worshippers with respect to women. Names of places most suited for the worship of Nanda (viz., Mandara, Trikuta, Vindhya-parvata, Samkaresvara, Ramesvara, Amaresvara-tirtha, Hariscandra-tirtha, banks of the Ganga, Yamuna, Revati, Vetravati and Sarasvati, and so on). Method of mantroddhara. Performance of mudras. (In this chapter, there is mention of wall-painting and of education, pastimes, decoration and sports of girls). Chapter 94. Description of pilgrimage to the residence (i.e. temple) of Sunanda, with the mention of the rivers, holy places, and images of deities lying on the way, viz., Ganga with Marukesvara Siva on her bank, the river Siloccaya on the left of Urdhvayana, Kalakuta (tirtha) with its 'great temple' of Kalahamsesvara Siva and the river Kausika (flowing by it), Sulabheda (tirtha), Vasanta-vana, Karttikapura, Vaisravana-pura, and the river Vaitarani. Chapter 95.-Description and praise of Kanyakapura (residence of Kanyaka). Identification of Nanda with Siva. Chapter 96. Hereditary study 137 of the Vedas-one of the principal duties of the twice-born people (dvija). Contact with, and food offered by, Sudras-to be avoided by them. Fit places of Vedic study (viz., temples of Visnu and Surya, bank of a river, and so on). Method of reading the Vedas. Persons who are not entitled to read the Vedas. Days on which Vedic study should be postponed. Chapter 97.-Killing of animals in sacrifices-not a crime. Directions and prohibitions to be observed by the people.138 Chapter 98.-Unity of Devi, who is known under different names (viz., Narayani, Tara, Sveta, Mahasveta etc.). Method of pavitraropana after worshipping Devi (in any of her forms, viz., Carcika, Camunda, Nanda etc.) on a sword, dagger, foot-wear, picture, book, flag, bow, image or altar. Music, dancing etc. to be performed by bands of actors, 137 According to verse 4, those twice-born people who have Vedic Samskara from their fifth ancestors on their mothers' sides and from the seventh on their fathers', become pure (suddha) and fit for Vedic study. 138 People are advised to read medical treatises daily and to avoid quarrel with physicians and astrologers.-Devi-purana 97.17-19.

62 STUDIES IN THE UPAPURANAS dancers and prostitutes. General holy day to be observed by all for ten, five, three or one day on this occasion. Chaps. 99-101.-Methods and results of observing the Nanda-vrata (in which Nanda is to be worshipped in the twelve months from Sravana with the mention of her twelve names, viz., Nanda, Sunanda, Kanaka, Uma, Durga, Ksamavati, Gauri, Yogesvari, Sveta, Narayani, Sutaraka and Ambika), Vijaya-vrata (on the vijaya purnima in the month of Sravana), and Naksatra-vrata (in which the different limbs of Devi are to be worshipped under different naksatras). Feeding of Brahmins, women and virgin girls in all these Vratas. Chapter 102. Results of giving land, gold, cows, paddy etc. to Brahmins, women, virgin girls, and devotees of Devi at Nanda-tirtha, Gaya, Puskara, Kasi, Kuruksetra, Prayaga, Jambukesvara, Kedara, Dandakaranya, Somesvara, Amarakantaka, Vindhyacala, banks of the Ganga and the Narmada, and so on. Method of performing the Pada-vrata (in which Devi is to be worshipped on an altar, an image, or Devi's foot-wears). Chapter 103.-Method of giving cows (go-dana) and golden cows (hema-go-dana) to the devotees of Siva, after worshipping Devi in these cows. Chapter 104.-Donation of different articles in the different months of the year. Method of giving a cow to a qualified Brahmin in the month of Sravana, after worshipping Devi, Brahma, Siva, Surya and Visnu. Method of tiladhenu-dana (on the occasion of which Devi is to be worshipped). Chaps. 105-106.-Methods of ghrta-dhenu-dana and toyadhenu-dana (with the worship of Devi on these occasions). Chapter 107.-Devi as the mother of the Vedas. Division of the original single Veda into four, viz., Rk, Yajus, Sama and Atharva. Names of the ten different texts (bheda) of the Rgveda, viz., Aslesa, Samkhya, Carca, Yavaka, Carcaka, Sravaniya, Krama, Putakrama, Vatakrama, and Danda; names of its three schools (sakha), viz., Sakala, Brahma and Manduka; and the numbers of its Mandalas, Vargas, Rks ས? ད- ,-, ཀཏ

THE SAKTA UPAPURANAS 63 and Padas. 139 Names of the schools of the Yajurveda, and the number of its verses. 140 Names of the Angas, Upangas and Parisistas. 141 Names of the schools and divisions of the Sukla-yajurveda, Sama-veda and Atharva-veda, and the numbers of their verses. 142 The Upavedas, viz., Ayurveda, Dhanurveda, Gandharva-sastra and Arthasastra, which are attached to the four Vedas Rk, Yajus, Sama and Atharva respectively. Names of gotras (Atreya, Kasyapa, Bharadvaja and Vaitala) and adhidevatas (Soma, Rudra, Indra and Brahma) of the four Vedas. Description of the forms of the 139 atra bhedas tu rgveda dasa caiva prakirtitah // aslesah samkhyas carcas ca yavakas carcakas tatha / sravaniya ca krama ca putakrama-vatakramah // dandas ceti samasena punar ekaiva paraga / sakhas ca trividha bhuyah sakala brahma-mandukah // tesam adhyayanam proktam mandalas catuh-sastikah (v.I. catuh-saptatih) | varganam parisamkhyatam caturvirsa-satani ca // rcam dasa-sahasrani rcam panca-satani ca / manam asiti-padas ca tatra paranam ucyate // Devi-purana 107. 13 b-17. 140 The text of the verses containing these names is full of mistakes. It runs as follows:- 141 . samkhya yajurvedasya sruyatam / sadasiti-vibhedena maya bhinnam sivajnaya // dasadha caraka tatra karaka vidradhisaya (v.l. karakaharitradhiyaya) | kathah pracya-kathas caiva kapisthala-kathas tatha // caraniyah svetas ca sveta-tara maitrayaniti / punah saptatir bhedena maitrayanyah prakirtitah // manava-dundubha-varahas chageya haridraviya / samaya mayaniyas ca tesam adhyayanam ucyate // astadasa-sahasrani pathan sakhavido bhavet / dvigunam padapathi yas trigunam krama-paragah // Devi-purana 107. 18-23. siksa kalpo vyakaranam niruktam chando jyotisam // sadangani bhavanty ete tany upangani srnu kathyatam / pratipadam anupadam chando bhasa mimamsa ca // nyaya-tarka-samayukta upamgah parikirtitah / parisistas ca samkhyata astadasa srnusva tat // yupa-laksana-pratistha tu vakyam samkhyas carana-vyuhah / sraddhakalpas ca suklani parisadam rg-yajus ca // astakapuranam caiva pravaradhyayo'nga-sastram / kratu-samkhya nigama yajnaparsvantahautrikam // vratam ca pasavo homam kurma-laksana-samyutah / kathitah parisistas tu unavimsa mahamune // It is clear that the text of these verses is highly defective. Devi-purana 107. 24 b-29. 14 For the names of the different schools of these Vedas see Devi-purana 107. 30-44.

64 STUDIES IN THE UPAPURANAS four Vedas personified. The different objects in which these Vedas are to be worshipped. Chapter 108.-Praise of Ayurveda. Debate among the sages about the cause of disease in human body, and their conclusion that bad diet is its root cause. Chapter 109.-Classification of food with regard to the source of its origin, its effect on human body, the modes of taking it, its taste, and so on. Chapter 110. Names of different kinds of food, 143 as well as of roots, fruits, herbs etc. which are most beneficial or harmful to the human body. Praise of Durga-worship and of study of Tantras. Praise of Ayurveda-sastra, which is said to have been disclosed by Siva in order to kill Khatvasura. [Chaps. 108-110 are the reproduction of Caraka-Samhita I (Sutrasthana), 25 with minor additions and alterations.] Chapter 111.-Story of the origin of Khatvasura from the rage of Visnu and Siva; his successful war with the gods; and his penance. Chapter 112.-Visnu's rubbing (cf. sammathayitva) of his palms under the influence of rajasa-bhava, and the consequent appearance of Gajanana. Gods' occupation of the different parts of Gajanana's body. Chapter 113. Visnu's eulogy of Gajanana by describing his characteristics. 144 Gajanana's statement that he has been sent there by Mahadeva, and his consent to remove the impediments of gods. Chapter 114.-Worship of Gajanana by Brahma, Visnu, Siva and other gods. Offer of different articles (viz., crescent moon, white girdle, coral, conch-shell, bow, mace, goad etc.) by different gods, and Siva's appointment of Gajanana as the leader of gods under the name of Vinayaka. 143 In this chapter beef is denounced as the worst kind of meat; meat of cocks is praised as the most strengthening; sight of Nagnas and Kasayins is said to do the greatest harm to life; the Sun (Aditya) is called the best healer of leprosy; and so on. 144 In this eulogy Gajanana is said to have serpents as his head-ornament, carry danda, ankusa, parasu, mekhala and sutra in his hands, hold a 'ladduka' in one of his hands and sandal etc. in another, and have his neck adorned by a necklace and his cheeks drenched by ichor. ,,

THE SAKTA UPAPURANAS 65 Chapter 115.-Vinayaka's arrival at Udayacala and killing of Vighnasura. Chapter 116. Story of the origin of Vighnasura from the fire issuing from Brahma's mouth when the latter, being engaged in penance, thought himself to be the only god. Visnu's eulogy of Siva and Devi for killing Vighnasura, and Siva's boon to Visnu that Vighnasura, Jambhasura and Sulomasura would be killed by Vinayaka surrounded by the Matrs. Story of Khatvasura's penance for Devi's favour. Mandavya's saving of Hariscandra's life and kingdom by worshipping the Matrs at Vindhyacala according to the advice of the five Matrs (viz., Ambika, Rudrani, Camunda, Brahmi and Vaisnavi) whom Mandavya first worshipped at Somesa on the bank of the Sarasvati. Chapter 117.-Praise of the worship of Matrs, especially at a time when the sun remains at the Kanya rasi (kanyasamsthe divakare). Chapter 118.-Results of constructing new temples and of removing the images of Durga and the Matrs to these temples from the old ones. Suitable time (viz., daksinayana) for removing the images of Brahma, Indra, Siva, Visnu, Surya, Matrs and Bhairava. The different mantras to be used in removing the images of Matrs. Offer of wine, meat etc. to these female deities. Results of repairing old temples and of replacing old images with new ones. Chapter 119. After killing Khatvasura Siva assumed a dreadful form (of Bhairava ?), held a 'khatvanga' in his left hand and a 'kapala' in his right, and wore a garland of skulls on his head. Going to speak on the origin of the 'kapala' and 'khatvanga', Siva describes himself as the Eternal Being and Devi as the Mula-prakrti, and narrates how he created water and thence the universe with the gods, men and others. He also says that the skulls and limbs worn by him are respectively those of the innumerable Brahmas and Visnus whom he created in the past. Regarding the origin of the 'khatvanga' Siva says that the impediment to his yoga- 5

* py 66 STUDIES IN THE UPAPURANAS practice, being hurled up by him in rage, assumed the form of the demon Khatva and was destined to be killed by Vighnesa (i.e. Vinayaka). Chapter 120. Purification of the body by muttering the Padamala-mantra or the Gayatri, and by practising 'yama' (consisting of brahmacarya, daya, ksanti, dhyana, satya, ahimsa, asteya, madhurya and dama), 'niyama' (consisting of snana, svadhyaya, upastha-nigraha, guru-seva, sauca, akrodha and apramada), and various kinds of penance (viz., santapana-vrata, mahasantapana-vrata, parnakrcchra-vrata, taptakrcchra-vrata, padakrcchra-vrata, prajapatya-vrata, atikrcchra-vrata, krcchratikrcchra-vrata, paraka-vrata, saumyakrcchra-vrata, tulapurusa-vrata, candrayana-vrata, and krcchra-candrayana-vrata), which have been described. Chapter 121. Maintenance of the sacred fire-a duty of those who are versed in the literatures of the Vamas and the Daksinas as well as in Vedanta, Grhya etc. Method of performance of homa by the members of the higher three castes as well as by Sudras, women and children. Names (Agni, Hutasana and Anala) of the three kinds of fire and of the gods who reside in it. Chapter 122.-Origin of Ahavaniya Agni from Garhapatya Agni, and of the latter from Daksinagni. Names of the fifty-one descendants of Ahavaniya Agni, and the occasions for their use. Chapter 123. Different kinds of flowers, incense, scents etc. to be used in Devi-worship. Chapter 124.-Methods of worshipping Devi and the spiritual preceptor (guru).-Proper place for worship; decoration of the place; performance of sadanga-nyasa; bathing of Devi and the spiritual preceptor; offer of different articles including meat; performance of homa; and so on. Chapter 125.-Praise of the worship of the spiritual preceptor who is identified with Siva. Chapter 126.-Method of Agni-samskara and the performance of homa for removing the impediments to 'japa'. and 'adhyayana'. (In this chapter there is mention of different kinds of

THE SAKTA UPAPURANAS 67 mudras and mantras such as kavaca-mantra, astra-mantra, etc.). Chapter 127.-Necessity of the performance of Deviworship by kings, who are to keep their subjects in the path of virtue. Influence of the conduct of kings on their subjects; hence the necessity of imparting good education to a king even with the help of medicine, mantras etc. Necessity of sraddha on the part of a learner. Story of the revelation of Devi's Stava-raja. (In this Stava-raja there is mention of the different names¹45 and exploits of Devi as well as of the names of the different places¹46 of Devi-worship; Devi is said to be worshipped as Vindhyavasini by Savaras, Varvaras and Pulindas 147 and to move about in the sky in a form having eighteen hands; and the relation between Siva and Devi has been explained with various examples). Chapter 128.-Method of worshipping the Devi-purana and of reading it before a gathering. Honour to be shown to the reader. Fit places (viz., one's own house, Kasi, Kamapura etc.) for the study of the Devi-purana Results of such reading. The above analysis of contents of the present Devi-purana will show how important this work is from various points of view, especially for the history of Saktism. But, unfortunately, it does not contain the whole of the original Devi-purana, which must have been a much bigger work. We have already mentioned that in chap. 1 of the present Devi-purana Vasistha speaks of four Padas of the Devi-purana and gives the contents of all these Padas. This statement of Vasistha is corroborated by Bhagiratha who, in his commentary on Kumara-sambhava I. 1, says that Kalidasa wrote his Kumara-sambhava on the basis of the story of the birth of 145 Kanyadevi, Sakambhari, Mahasveta, Subhadra, Dhumra, Bhutamata, Matangi, Candika, Ghantarava, Virupaksi, Salankayani etc. 146 Himalaya, Kailasa, Meru, Mandara, Vindhya, Malaya, Gandhamadana, Naimisa, Kuruksetra, and the tract of land lying between the Ganga and the Yamuna. 147 The text of Devi-purana 127. 110 b is as follows:- - saravarsavarais capi pulindais capi pujyase. The reading 'saravarsavarais capi' must be a mistake for 'savarair varvarais capi'. Cf. Harivansa II. 3. 7 a- savarair varvarais caiva pulindais ca supujita.

• 68 STUDIES IN THE UPAPURANAS Karttikeya as given in the 'Devi-purana', and thus proves that the 'Devi-purana', with the contents of at least its fourth Pada, was known to him.148 But in the present text of the Devi-purana we find only those topics which have been mentioned (in chap. 1) as the subject-matters of the first two Padas or rather the second one; 149 and at least three of the complete Manuscripts of the Devi-purana end with the verse 'etac cabhyudayam padam satais tribhir narottamaih etc. '150 Hence it is evident that the present Devi-purana does not represent the bigger and entire work that was known under the same title in earlier days but contains only a portion ending with the Trailokyabhyudaya Pada and dealing with the first incarnation of Devi. 151 This fragmentary condition of our present Devi-purana is 148 Dacca University Manuscript No. 108 H (Bhagiratha's com. on Kalidasa's Kumarasambhava), fol. 2 a-'kavih kalidasah devi-puraniyam karttikeya-janma-katham asritya kumara-sambhavam nama mahakavyam idam niravatma (? niravartayat)'. This statement, however, is not correct. Bhagiratha calls himself 'pitamundi-ku.a-bhusana' and was a rahmin. It is to be noted that according to the present Devi-purana (1. 45-46) the story of Karttikeya's birth and exploits was dealt with in the fourth Pada. 149 The facts that the contents of the second Pada begin practically from chap. 2 of the present Devi-purana and that in chap. 1 the Purana is found to begin abruptly and the interlocution between Nrpavahana and Citramgada is not preceded by any introduction of the interlocutors, tend to show that even the Trailokya-vijaya Pada has sustained serious loss. Whatever the meaning of the line 'utpatti-kirtanam srsteh prathamam samudahrtam' (given in footnote 110 above) may be, it must be admitted that the Trailokya-vijaya Pada is almost totally lost. It is only Devi-purana, chap. 1 which seems to have belonged to the Trailokya-vijaya Pada. Cf. Devi-purana 1.28- adyadhyayena samksepat puranam samudahrtam / pade trailokya-vijaye sarva-kama-prasadhanam // With the present state of our knowledge we cannot say definitely whether Devi-purana 127. 24 ff., which deal with the first origin of Devi at the beginning of creation, originally belonged to the first Pada, or whether these verses were added only to deal very briefly with an important topic which was going to be omitted with the omission of practically the entire Trailokya-vijaya Pada. 150 See footnote 105 above. The Manuscript of the Devi-purana described by Eggeling in his Julius Eggeling, A Descriptive Catalogue, VI, pp. 1205-7 has the final colophon- 'ity adye devyavatare sri-devipurane 'bhyudaya-padah samaptah'. 151 That the present Devi-purana deals only with the 'first incarnation' of Devi is shown by the reading 'ity adye devyavatare etc.' occurring in the final colophon as well as in some of the chapter-colophons of the Manuscript of the Devi-purana described by Eggeling (see Julius Eggeling, A Descriptive Catalogue, VI, p. 1207) and in the final colophons of the two Manuscripts described in Mitra's Bikaner Cat. and Shastri and Gui's Calcutta Sanskrit College Cat. (see footnote 76). Cf. also Devi-purana 128.4 wherein the 'adya devyavatara' has been mentioned indi rectly, in connection with 'phala-sruti', as the subject-matter of the present Devi-purana

THE SAKTA UPAPURANAS 69 due most probably to the repeated revisions and abridgments to which this Purana, like many other works of its class, was subjected from early days. The Devi-purana itself records a tradition that this 'Sastra' was revealed by Siva to Brahma in ten lacs (of slokas ?) and by Brahma to Indra in one lac (of skokas ?). It then reached Agastya (most probably in a more abridged form), and Agastya narrated it to king Nrpavahana (perhaps far more concisely). 152 That, in spite of all its exaggerations, this tradition is not totally false but is based on some amount of truth, is shown by (i) the absence of the third and fourth Padas and most probably of the major portion of the first also 153 from the present Devi-purana, (ii) the abrupt beginning of the Purana, the sudden mention of Nrpavahana and Citrangada in Devi-purana 1. 63-66 without any previous introduction, and the abruptness of some of the interlocutions, 154 (iii) a comparison of some of the extracts quoted in the Nibandhas from the 'Devi-purana' with their parallels in the present Devi, 155 189 Devi-purana 128. 1-2. 153 See footnote 149 above. In his Krtya-ratnakara, pp. 357-8 Candesvara quotes twenty-nine metrical lines, ascribing them to the 'Devi-purana' at the outset and adding at the end: "bhagavatapurane devy-avatare navami-kriya-sucanam asta-catvarimso 'dhyayah." But, curiously enough, these twenty-nine lines constitute chap. 21 (and not chap. 48, as said in the Krtya-ratnakara) of the present Devi-purana Moreover, some of the long extracts and many of the verses quoted by Candesvara from the 'Devi-purana' are not found in our printed text. So, the text of the Devi-purana used by Candevsara was much more extensive than that of the printed work. 184 For instance, the interlocution between Saunaka and Manu begins from chap. 39, although no mention of the former has been made anywhere in the preceding chapters and nothing is said as to how these two interlocutors came to meet each other. 155 A long extract dealing with Krsnastami-vrata and consisting of 71 metrical lines has been quoted from the 'Devi-purana' in Hemadri's Caturvarga-cintamani, II. i, pp. 823-6. A comparison between this extract and chap. 78 of the extant Devi-purana shows that though the present Devi-purana has retained many of the lines of the quoted extract, it has summarised many of the remaining ones by retaining the language of the original as far as possible, that the Devi-purana (chap. 78) condenses the whole topic in 28 lines, whereas the quoted extract has 71 lines, and that in the Devi-purana Manu is the speaker, whereas in the quoted passage the interlocutors are Devi and Isvara. A comparison between a long extract (on Durga-varta) quoted in Caturvargacintamani, II. 1, pp. 856-862 and Devi-purana 33. 50-109 shows that in the quoted passage Brahma speaks to Sukra, whereas in the Devi-purana it is Isvara (i.c. Siva) who speaks to Sukra, and it is due to this change in the speaker that instead of the line 'etad vratavaram sukra maya rudrena visnuna" ('this best of vows was practised by myself, Rudra and

70 STUDIES IN THE UPAPURANAS and (iv) the loss of some of the chapters from the extant work.156 The above tradition as well as the probable loss sustained by our extant Devi-purana, shows that the changes made in the present Devi-purana are far more of the character of condensation than of expansion by additions. 157 So, in spite of these repeated changes the major portion of the present contents of this Purana can generally lay claim to an early date, if not to that of its first composition. The close interrelation among the chapters 158 also points to the general integrity of the Purana. Visnu') occurring towards the end of the quoted extract, the Devi-purana has 'etad vratavaram sukra maya brahmana visnuna". Again, a summary of contents of a few lines of the Devi-purana, together with the line "srnu sakra pravaksyami' and the verse 'subham vapy asubham kimcit' which precede and follow this summary respectively, has been given in Narasimha Vajapepin's Nityacara-pradipa, pp. 640-1. A comparison between this summary and Devi-purana 123. 11 ff. shows how the Devi-purana has been reshaped. In the quoted portion Sakra is addressed by somebody, whereas in the extant Devi-purana it is king Nrpavahana who is instructed by Agastya on the method of Devi-worship. Though the three quoted lines are not found in Devi-purana, chap. 123, and the summary of other lines agrees with the contents of Devi-purana 123. 11 ff. only imperfectly, it is Devi-purana, chap. 123 which must have been the new form of that chapter from which the above-mentioned lines and summary were derived, because, in the expression 'mukuladi-nisedhah purvavat/puspabhave patrani etc.' (Nityacara-pradipa, p. 641) there is unmistakable reference to Devi-purana 123. 18-20. Compare also the verses 'vama-daksina-vetta yo matr-vedartha-paragah' etc. quoted in Apararka's com. on the Yajnavalkya-smrti (p. 16) with Devi-purana 32. 42-43 or 51. 4-6; a long extract of 130 lines quoted in Raghunandana's Durga-puja-tattva pp. 25-29 may be compared with Devi-purana 50. 88 ff.; the verse 'pancaratrartha-kusalo' quoted in Nityacara-pradipa, p. 21 may be compared with Devi-purana 32. 42-43 or 51. 4-6; the description of Mangala, Rddhi, Rati, Kanti, Yasa, Sakri, Jayavati, Ajita, Manasi, Sveta and several other goddesses, as given in Vidyapati's Durga-bhakti-tarangini, pp. 77 ff., may be compared with that contained in Devi-purana 50.i. 30, 113-4, 45-7, 125-6, and ii. 11-12, 9-10, 16, 24-25 a, 31, etc; and so on. Such comparison will show definitely that the present Devi-purana is the result of a recast to which the earlier Devi-purana was subjected. 156 The mention of 'Yudhisthira' and 'Kuru-sardula' as interlocutors in two extracts quoted from the 'Devi-purana' in Kalasara, p. 103 and Varsa-kaumudi, pp. 30-31 respectively but not found in the present Devi-purana, and the occurrence of the word 'partha' in the Vocative Case in Devi-purana 27. 6 show that our Devi-purana once contained chapters in which Yudhisthira was one of the interlocutors. But in the present Devi-purana there is no such chapter. The Manuscripts of the Devi-purana say that this Purana, ending with the Trailokyabhyudaya Pada, consisted of 12300 slokas, whereas these Manuscripts as well as the printed edition actually contain 7000-8000 slokas. (See footnote 105 above). 187 That the present Devi-purana is not totally free from comparatively late additions, we shall see later on. 158 For instance, Devi-purana 21. 14 refers to Devi-purana 2. 33-35; 36.1 refers to 33.2; 44.4 refers to 42.10(?); 46.9 refers to chapters 47 ff.; 57.3 refers to chap. 50; 65.90 refers

THE SAKTA UPAPURANAS 71 So, at least the great majority of the chapters must be taken as contemporaneous. The Devi-purana calls itself a 'Purana'159 or a 'Sastra'100 and never an 'Upapurana'; but it has not been included in any of the comparatively early lists of eighteen Puranas or Upapuranas contained in the different Puranic and non-Puranic works, 161 most probably because of its abounding in Tantric elements.162 Yet it is certainly not a very late work. It has been profusely drawn upon by a large number of Nibanto chapters 50-54; 74.2 refers to chapters 24 ff.; 92.1-4 refer to chapters 80-81; 93.22-23 refer to chap. 63; 93.59 refers to 93.26 ff.; 121.4-5 refer to chap. 26(?); and so on. 159 See Devi-purana 1. 1; 1.78; 1.57; 128.49; also the colophons of chapters. The reading 'ityadye mahapurane' after Devi-purana 7.90 must be a wrong one for 'ityadye devipurane' which is found in many of the chapter-colophons. It is noteworthy that in the chapter-colophons the Purana calls itself 'adya'. In spite of its claim to the status of a 'Purana' it contains very little of the five topics characteristic of a 'Purana', 'vamsa', 'manvantara' and 'vamsanucarita' being safely omitted. 160 See Devi-purana 128, verses 1, 13, 15, 25, 34 and so on. 161 It is only in the Ekamra-purana (a work of Orissa) that the 'Devi' has been included in the list of eighteen Upapuranas (see Volume I, p. 13). In his Smrti-tattva Raghunandana also includes the 'Daiva' in the list of eighteen Upapuranas contained in the verses quoted by him from a 'Kaurma' (i.e. Kurma-purana). Raghunandana explains 'Daiva' as 'Devi-purana' (daivam devipuranam).-Smrti-tattva, I, pp. 792-3. The text of the verses of the Kurma-purana, as given by Raghunandana, is found to agree neither with that of the extant Kurma nor with those of the verses on the list of Upapuranas quoted from the 'Kurma-purana' in the different Smrti Nibandhas. (For the texts of these verses of the 'Kurma-purana' see Volume I, Chapter I, footnotes 6-9). By his reference to the list of eighteen Upapuranas as occurring in the 'Kurma-purana' and by his express mention that the Devi-purana was not included in any of the lists of eighteen Puranas and Upapuranas, Vallalasena also disapproves definitely the text of the abovementioned verses as given by Raghunandana. See Danasagara, p. 3 (verse 12 b-13 a)- kurmapuranadipuranayoh / uktany upapuranani vyakta-dana-vidhini ca || and p. 7 (verse 67)- tat-tat-puranopapurana-samkhya-bahiskrtam kasmala-karma-yogat / pasanda-sastranumatam nirupya devipuranam na nibaddham atra // Hence the age and authenticity of Raghunandana's text are extremely doubtful. The Devi-purana is also included in the list of Upapuranas given in the Sabda-kalpadruma, but this list has been derived most probably from Raghunandana's Smrti-tattva. 162 Vallalasena ascribes the exclusion of the Devi-purana from the lists of Puranas and Upapuranas to its Tantric character. See Danasagara, p. 7 (verse 67)- kasmala-karma-yogat / tat-tat-puranopapurana-samkhya-bahiskrtam pasanda-sastranumatam nirupya devipuranam na nibaddham atra //. It is to be noted that at the time the present Devi-purana was first composed, the Puranas imbibed very little Tantric influence and looked down upon the Tantras as 'Pasanda- sastras'. So, it is natural that the Devi-purana, with its prominent Tantric elements, should be kept outside the group of the 18 Upapuranas. See R. C. Hazra, Studies, pp. 260 ff.

72 STUDIES IN THE UPAPURANAS dhakaras, both early and late. For example, Gopalabhatta quotes numerous verses from Devi-purana, chapters 51, 61 and 118 in his Hari-bhakti-vilasa; Gadadhara from chapters 21, 22, 24, 59, 61 and 96 in his Kalasara; Anantabhatta from chap. 54 in his Vidhana-parijata; Narasimha Vajapeyin from chapters 32 (cf.), 34, 50, 51 and 123 in his Nityacara-pradipa; Raghunandana from chapters 21-24, 50, 51, 54, 59, 61 and 78 in his Smrti-tattva and from chapters 21-23, 50, 54, 59 and 123 in his Durga-puja-tattva; Govindananda from chapters 21 (cf.), 24, 59, 61 and 78 in his Varsa-kaumudi, from chap. 59 in his Sraddha-kaumudi, and from chap. 60 in his Danakaumudi; Srinatha Acarya-cudamani from chapters 22, 24, 59 and 61 in his Krtya-tattvarnava; Vacaspatimisra from chap. 54 in his Tirtha-cintamani, from chap. 59 in his Sraddha-cintamani, and from chapters 48 and 59 in his Krtya-cintamani; Rudradhara from chap. 24 in his Varsakrtya and from chap. 59 in his Sraddha-viveka; Suulapani from chapters 21-23 and 61 in his Durgotsava-viveka, from chap. 97 in his Prayascitta-viveka, and from chap. 59 in his Sraddha-viveka; Candesvara from chapters 12, 21, 22, 24, 33, 50, 59, 61, 62, 74, 79, 89, 91, 99, 102, 104, 117, 123 and 127 in his Krtya-ratnakara, from chap. 97 in his Grhastharatnakara, and from chapters 59, 103, 105 and 106 in his Danavakyavali;Madanapala from chapters 51, 59 and 97 in his Madana-parijata; Madhavacarya from chapters 24, 74 and 104 (cf.) in his Kala-nirnaya; Vidyapati from chapters 21-23, 50, 51, 53, 59, 89, 91, 117, 123 and 124 in his Durga-bhaktitarangini and from chap. 50 in his Ganga-vakyavali; Sridatta Upadhyaya from chapters 50 and 51 in his Acaradarsa (alias Krtyacara) and from chapters 24, 45, 59, 61 and 79 in his Samaya-pradipa; Hemadri from chapters 11, 12, 21-24, 27, 31, 32, 45, 50, 56, 58-62, 64-67, 74, 78, 79, 89, 91, 98, 99-107 and 121 in his Caturvarga-cintamani; Vallalasena from chap. 100 in his Adbhuta-sagara;163 Apararka from chapters 32 (cf.), 13 The authenticity of the quotation made from the Devi-purana in the Adbhutasagara is more or less debatable. Towards the beginning of his Adbhuta-sagara the author gives a list of authorities utilised by him in his work. But an actual examination of the Adbhuta-sagara shows that this work contains verses from such Puranas and

THE SAKTA UPAPURANAS 73 51 (cf.), 59, 74, 97, 102 and 104 in his commentary on the Yajnavalkya-smrti; Laksmidhara from chapters 12, 23-27, 33, 58-59, 64, 74, 79, 89, 91, 96-99, 101-106 and 127 in his Krtya-kalpataru, Volumes I-VI, VIII and XI; and Jimutavahana from chapters 22, 48, 61, 74 and 93 in his Kalaviveka. Of the 21 metrical lines ascribed to the 'Devi-purana' in the Samvatsara-pradipa, six are found in chapters 48, 59 and 61 of the present Devi-purana 164 Some of the verses of the Devi-purana, which were derived by Raghunandana from 'Vacaspatimisra', 'Krtya-maharnava' (of Vacaspati-misra), 'Krtyacintamani', 'Navya-vardhamana', 'Durga-bhakti-tarangini' (of Vidyapati), 'Naiyata-kalika-kalpataru' and 'Kamadhenu' (1000-1100 A.D.), are found in Devi-purana, chapters 21, 24 and 59.165 Hence the date of the present Devi-purana, which was known far and wide in the eleventh century, can never be placed later than 850 A.D. Though the authenticity of the quotation made in the Adbhuta-sagara from Devi-purana, chap. 100 is more or less debatable, Vallalasena's knowledge of the present Devi-purana can by no means be questioned. By his remark that the Devi-purana, which was exluded from the lists of eighteen Puranas and Upapuranas due to its treatment of 'sinful acts', was rejected by him because of its Tantric character, 166 Vallalasena undoubtedly means the present Devi-purana which had Tantric elements even before the time of Apararka. 167 Now, Upapuranas (viz., Devi-purana, Visnu-purana, Nandi-purana, Vayu-purana, Aditya-purana, Linga-purana and Brahmanda-purana) as have not been mentioned in the list. The exclusion of the names of these works from the list may be due either to the inadvertance of the author or to the insertion of verses of these works by people other than the author. Though such incomplete lists are to be found in other Sanskrit works also, a consideration of Vallalasena's critical outlook as well as of the fact that Laksmanasena completed this work left unfinished by its author, tempts us to suppose that most probably it was Laksmanasena who revised and improved the Adbhuta-sagara with further addition of materials. 14-15 Complete lists of all the traced and untraceable verses will be given at the end of the final Volume of the present work. 1** Cf. Danasagara, p. 7 (verse 67)- 'tat-tat-puranopapurana-samkhya-bahiskrtam etc.' 167 Cf. Apararka's com. (on the Yajnavalkya-smrti), p. 16 yad api devi-purane- vama-daksina-vetta yo matr-vedartha-paragah / sa bhavet sthapakah srestho devinam matara(-trka)su ca //

74 STUDIES IN THE UPAPURANAS by way of explaining why he rejected some of the Puranas and Upapuranas, partly or wholly, in his Danasagara, Vallalasena makes valuable remarks on their contents and authenticity as a result of a highly critical study of the entire Puranic literature and names a set of Puranas which he regarded as spurious, viz., Tarksya (i.e. Garuda), Brahma, Agneya, Vaisnava consisting of 23000 slokas, and Lainga of 6000 slokas, 188 But he expresses no such opinion about the Devi-purana On the other hand, he says that the lists of Puranas and Upapuranas excluded it because of its Tantric character, thus implying that the Devi-purana is as old as any of the Puranas and Upapuranas mentioned in the lists. All this shows that in Vallalasena's days the reputation of the Devi-purana as a genuine Puranic work was firmly established and no question was raised against its authenticity and early origin. So, the Devi-purana must be much older than the extant Garuda and Agni which are identical with the spurious Tarksya and Agneya mentioned by Vallalasena. 169 The extant Garuda and Agni being works most probably of the tenth and ninth centuries respectively, 170 the Devi-purana must be placed earlier still. Hence the Devi-purana cannot possibly be later than 700 A.D.171 pancaratrartha-kusalo matr-tantra-visaradah / .... .... siva-sasana-vetta yo graha-matr-ganarthavit / .... .... .... 168 Though it is not yet possible to identify definitely the spurious Brahma, Vaisnava and Lainga mentioned by Vallalasena, the contents of the spurious Puranas, as given by Vallalasena in his Danasagara, p. 7 (verses 63-66), show definitely that the spurious Tarksya and Agneya of Vallalasena are the same as the present Garuda and Agni-purana respectively. For more detailed information on this point see R. C. Hazra, Studies, pp. 137-8 and 143. 169 See the immediately preceding footnote. 170 R. C. Hazra, Studies, pp. 144 and 138. 171 In his Durgotsava-viveka, p. 18 Sulapani says that the line 'astamyam balidanena putra-naso bhaved dhruvam' of a 'Devi-purana' was explained by Srikara as prohibitive of samdhi-puja and bali-dana during the Astami Tithi (yat tu 'astamyam balidanena iti devipuraniyam tad astami-ksane samdhipuja-balidana-nisedhakam iti sridatta-srikara-misrah). So, it is sure that a 'Devi-purana' was known to Srikara, who flourished between 800 and 1050 A.D. and probably in the ninth century. But as the above-mentioned line is not found in the present Devi-purana, we are not sure whether the present Devi-purana is the same as that known to Srikara. The want of any tradition or evidence as to the existence of a second Devi-purana and the fact that the present Devi-purana ཏཏ པ , ཏ ཙཏ ཨ ཨ

THE SAKTA UPAPURANAS 75 This lower limit of the date of the Devi-purana seems to be pushed further up by a good number of references, contained in it, 17 to the independent worship of Brahma. The opinion of scholars that the sect of Brahma became prominent during the period ranging from 200 to 600 A.D. and that the five gods of the Smartas threw Brahma into the background towards the beginning of the seventh century, 173 tends to show that at the time the present Devi-purana was first composed, the worship of Brahma was more or less popular. Hence the Devi-purana is to be placed not later than about 600 A.D. Again, Devi-purana, chap. 69 (except verses 1 and 22 b), which is the same as Yajnavalkya-smrti I. 271-289 a and 290-293, must have been taken from the latter work; Devi-purana, chapters 108-110 are mainly the reproduction of Carakasamhita I (Sutra-sthana), chap. 25; and there are references to the decadent state of Jainism and Buddhism in several places of this work. 174 Moreover, there is mention of hora and drekkana, 175 of the names of the zodiacal signs (rasi) 176 and week-days, 177 of the Saivas and Pasupatas as two distinct Sivaite sects, of the Vamas and Daksinas as two divisions of the Tantriks, of the Buddha as one of the ten incarnations of Visnu, 178 of the Tantrik Baudcontains the majority of the verses ascribed to the 'Devi-purana' in Sulapani's Durgotsavaviveka, however, tend to show that the two Devi-puranas, mentioned above, are identical, and that the text of the Devi-purana, used by Srikara and Sulapani, had a few more verses (on Devi-worship) which do not occur in the present text of this Purana. 172 See Devi-purana 2. 13; 35. 18-19; 50. iv. 57; 61. 1; 82. 15; 104.9; 118. 2; and so on. 173 J. N. Farquhar of the Religious Literature of India, pp. 148 and 179-180. Cf. also Brhat-samhita, chapters 58 and 60, wherein there are rules for the construction and erection of the image of Brahma, thus proving the wide spread of the worship of the god in Varahamihira's time. 174 See Devi-purana 9.32; 13.24; 27.18; 33.61; 65.52; 69.3; 88.2; 110. 3; and so on. These verses show that the Jains and Buddhists were regarded as wholly untrustworthy, and their sight was deemed inauspicious, unholy and most detrimental to longevity. Their sight even in dreams was considered as foreboding evil. 175 Devi-purana 73.49. Though the occurrence of the term 'hora' in two verses quoted by Varahamihira from Garga in Brhat-samhita, pp. 7 and 9 (chap. 2) shows that it came to be used by the Indians much earlier than the sixth century A.D., there is nothing to prove that it was used by them earlier than the end of the first century A.D. 176 Devi-purana 22.7; 32.8-10; 39.167; 43.64; 44.8; 46.81-82; 47.2; 50. iv. 1 ff. 177 Devi-purana 43.64; 45.3 ff.; 76.41. 178 The Buddha began to be regarded as an incarnation of Visnu most probably from about 550 A.D. See R. C. Hazra, Studies, pp. 41-42.

76 STUDIES IN THE UPAPURANAS dhas, 179 and of the Hunas as practising the different mystic Vidyas. 180 So, the Devi-purana cannot be dated earlier than 500 A.D. This upper limit of the date of the Devi-purana seems to be supported by its silence about the Devi-mahatmya of the Markandaya-purana as well as by the distribution of the different limbs of the kurma (kurmamga-samsthana), as described in Devi-purana, chap. 46 (verses 63 ff.), which prominently betrays the influence of Varahamihira's Brhat-samhita, chap. 14 (naksatra-kurmadhyayah) not only by one of its opening verses 181 but also by metrical similarity in a few others. 182 Besides these there are a few more verses in which the influence of the Brhat-samhita may reasonably be suspected.183 In the description of Ghora's nature and administration, as given in Devi-purana 16. 3-11, there are some statements 184 which have striking similarity with those made by Banabhatta in his Kadambari in course of describing the condition of king Sudraka's subjects, 185 and from these it is very natural 179 Devi-purana 88.2-3- pasandibhir bhavisyais tu bauddha-garuda-vadibhih / svadharma-niratair vatsa svena nyayena pujitah // This verse refers to the Pasanda (i.e. Tantrik) Bauddhas who worshipped the divine Mothers in their own way. See also Devi-purana 13. 10-12, which, by their mention that Silamati, wife of the demon Ghora, became devoted to the Digambara sect and observed vows but did not worship the Matrs, seem to betray their knowledge of the Tantrik Bauddhas. 180 Devi-purana 39.143. 181 Cf. Devi-purana 46.64 with Brhat-samhita 14.1. 182 Devi-purana, chap. 46, verses 68, 72-74, 80 and 82 are composed in the same metre as that of Brhat-samhita, chap. 14. See also Devi-purana 12. 34-38, 44 and 54-57. 183 For instance, Devi-purana, chap. 12, which deals with the hoisting of Indra's banner (indra-dhvajocchraya), has verses (viz., 22, 24 and 28-29) which may be compared with Brhat-samhita, chap. 43, verses 21, 22 a and 39-40; Devi-purana, chap. 47, dealing with the movements of the planets (graha-gati), opens with six lines written in the same metre as that of Brhat-samhita, chap. 14; and so on. 184 For these see verses 8-11 yasya kari-mahagandha mada-matta na rastrajah | yasya hataka-dandani chatresu na jane kvacit // yasya ghata asvostresu na pure na ca ghotake / yasya dutah priya-kope karmukanam na vigrahe // yasya cadhvara-yajnesu hy asru-pato na sokajah / yasya sasi-krpanesu kalanko na ca bhi-krtah // yasya svapna-prabha mithya na ca vaktavya-yojane / yasya bale mukhabhango na ca krodha-bhayat kvacit // 185 See Kadambari, pp. 10-11-yasmims ca rajani jita-jagati paripalayati mahim བར་

THE SAKTA UPAPURANAS 77 for one to take the Devi-purana as the borrower. But the grammatical and other errors occurring in some of these verses 186 go definitely against their spurious character. So, these verses cannot be utilised to bring down the date of the Devi-purana Thus, it is highly probable that the Devi-purana, as we have it now, comes down from the sixth century A.D. and most probably from its latter half. The mention of the earlier arrangement of the naksatras from Krttika to Bharani in Devi-purana 46.83 and 127.87-92 should not be taken to go against the above general date of the Devi-purana in its present form; because in the Brhat-samhita itself there is mention of the naksatras from Krttika to Bharani (see Brhat-samhita, chap. 15). A large number of verses is found common to Devi-purana, chapters 46-49 and Samba-purana, chapters 18 and 22-23.187 Although we have seen in course of our analysis of the Samba-purana in Volume I (p. 66) that chapters 22-23 of this work must have been added later, we cannot say definitely whether the Samba-purana borrowed the common verses from the Devi-purana, or both these Puranas utilised a common source. The grammatical correctness of the great majority of the verses of the said four chapters of the Devi-purana, especially the last three (viz., chapters 47-49), as contrasted very remarkably with that of the verses preceding and following them, shows that most of them were derived from a different source. But we are in complete darkness about the period of their inclusion in the Devi-purana The above general date, to which we have assigned the citra-karmasu varna-samkarah dandah karisu mada-vikarah svapnesu vipralambhas chatresu kanakaetc. 188 See, for instance, Devi-purana 16. 3 ('-madotsekat' for '-madotsekena'), 5 ('badhanti' for 'badhante', '-krodhadir na gano visate tanum' for '-krodhadinam na gano visati tanum'), 7 ('-mandala-m-tattva-vedita' for '-mandala-tattva-vedita', and 'karastham api dhatrijam' for 'karastham iva dhatrijam'), 8 ('kari-mahagandhah' for 'mahagandha karinah'), and so on. 187 Samba-purana 18. lb-38. Devi-purana 46.2-37 47.6 b-36 18. 39-68. 48.1-22 22. 3-24 a. 49.2-30 23. 9 b-39.

78 STUDIES IN THE UPAPURANAS present Devi-purana, should not be taken to be the date of composition of all the chapters without exception. It has already been said that this work, as we have it at present, is the result of repeated revisions and abridgments to which its original text was subjected with the progress of time; and we shall see presently that its peculiar language points to its earlier beginning. So, it is very likely that in the present text of the Devi-purana there are chapters (or parts thereof) and verses which were composed much earlier than the sixth century A.D. Being a great authority on Sakti-worship and consequently a work of extensive use, the Devi-purana could not escape interpolations. In connection with vidya-dana mention has been made, in Devi-purana chap. 91, of the Nandi-nagara script, which, as Buhler says, originated from the Southern Nagari of the 8 th-11 th centuries A.D.188 Hence chap. 91 (or rather verses 12-82, which deal with vidya-dana) must have been interpolated later. The comparatively late date of this chapter is also shown by its mention of yantra (diagram) as a medium of worship. It is to be noted that in the whole of the present Devi-purana there is no second instance in which the yantra has been prescribed to be used as a medium of worship. Chapter 91 is, however, not to be dated later than 1050 A.D., because Hemadri and Laksmidhara quote almost the entire chapter in their Caturvarga-cintamani and Krtya-kalpataru respectively. 189 Besides these, there may be other interpolations, but it is very difficult to find them out. Though the Devi-purana calls itself a 'Purana' or a 'Sastra' but never an 'Upapurana' and is a fairly early work, it has been included among the Upapuranas by Raghunandana¹00, Narasimha Vajapeyin¹¹, Mitra-misra 192, and Saiva NilaBuhler, Indian Palaeography, p. 51. 189 A complete list of the traced verses will be given at the end of the final Volume of the present work. 190 See footnote 161 above. 191 Nityacara-pradipa, pp. 18-19- astadasabhyas tu prthak puranam yat tu drsyate / vijanidhvam muni-sresthas tad etebhyo vinirgatam // vinirgatam samudbhutam / yatha kalikapuranadiniti laksmidharah / yac ca vayupurana-devipuranadi tad apy etesv antargatam / 1ºª Viramitrodaya, Paribhasa-prakasa, p. 15- ཨ, ཏ མ ཎ ཨ བཝ ཨ ཙ ད དཏ ཏ ཨ ཨན

THE SAKTA UPAPURANAS 79 kantha 193. But there was a section of people including Candesvara who claimed that this Devi-purana was the real Bhagavata. This opinion, being wholly unjustified 19, failed to find any wide acceptance among the people and was consequently discarded. As regards the provenance of the Devi-purana, there is little scope for doubt that it was written in Bengal. Besides the mention of the names of countries, rivers, holy places etc., mostly belonging to Northern India¹96 and showing the relation of this Purana with that part of the country, there are other evidences which point to Bengal as the place of its origin. These evidences are as follows: (1) As we shall see presently, the Devi-purana, unlike all other works of the Puranic literature, is written in highly incorrect Sanskrit which compares very favourably with Buddhist hybrid Sanskrit found in the Mahavastu, Lalitavistara and other works of East Indian origin. So, there can be little doubt about the fact that the Devi-purana also had its origin in Eastern India. (2) Of the numerous lists of Upapuranas contained in different works it is only those given in the Ekamra-purana (a work of Orissa) and Raghunandana's Malamasa-tattva which include the name of the Devi-purana197 (3) Although the Devi-purana mentions 'Kamarupa' and astadasabhyas tu prthak puranam yat tu drsyate / vijanidhvam dvija-sresthas tad etebhyo vinirgatam // vinirgatam samudbhutam / yatha mahajana-parigrhita-nandikesvara-puran-adipurana-devipuranaditi samksepah / 13 See Nilakantha's com. Tilaka (on the Devi-bhagavata), p. 3 a- devipurana-kalikapuranayor upapuranatvasya niscitatvat 194 Cf. the verse- 'bhagavatyas ca durgayas caritam yatra vartate / tat tu bhagavatam proktam na tu devipuranakam //' occurring in chap. 23 of the Uttara-khanda of the Bengal Siva-purana (for which see Julius Eggeling, VI, No. 3665, p. 1357). This verse is the same as verse 129 of chap. 44 of the Uma-samhita of the Siva-purana (Venkatesvara Press (Bombay) ed., which reads 'vidyate' for 'vartate'). In his Krtya-ratnakara (p. 358) Candesvara once names the Devi-purana both as 'Devipurana' and 'Bhagavata-purana'.-See footnote 153 above. 194 See R. C. Hazra, Studies, pp. 52 ff. 1** See Devi-purana, chapters 38, 39, 42, 46 (verses 63 ff.), 63, 74-76, and so on. 187 For these two lists see Volume I, pp. 4-5 and 13.

80 STUDIES IN THE UPAPURANAS 'Kamakhya' as places of Devi-worship, it does not add any importance to these, nor does it betray any influence of the method of Devi-worship followed in these places. On the other hand, it concerns itself with the praise and worship of Vindhyavasini (of whom, as the Dasa-kumara-carita tells us, there was a famous temple at 'Damalipta' or Tamluk) 198, and the importance of Kamarupa as a place of Sakti-worship and its influence on Bengal become more and more prominent in the later Puranic works such as the Kalika-purana, Mahabhagavata and Brhaddharma-purana (4) Most of the extant Manuscripts of the Devi-purana are found in Bengal and are written in Bengali script, and those which are now available at or near about Benares, might have been copied from their originals taken there from Bengal. In other parts of India, Manuscripts of this Purana are very rare. (5) As we have already seen, 199 the Devi-purana mentions satru-bali in Durga-puja in the following verse (22.16): tasyagrato nrpah snayac chatrum krtva tu pistajam / khadgena ghatayitva tu dadyat skanda-visakhayoh // Curiously enough, this custom survives in Bengal even to the present day. Satru-bali as a malevolent rite is very ancient. In early days it was performed after a pacificatory rite (santi) by a king's priest with the use of abhicara-mantras for the good of his royal patron. Varahamihira mentions it in the following verse of his Brhat-samhita (44. 21): "santim rastra-vivrddhyai krtva bhuyo,'bhicarakair mantraih/ mrnmayam arim vibhindyac chulen-orahsthale viprah //" "After performing the propitiatory rite for the growth [and prosperity] of the kingdom the [learned] Brahmin [priest] should, by citing Abhicara-mantras again, thoroughly pierce with a lance at its chest [the effigy of] an enemy made of clay." But the association of 'satru-bali' with Durga-puja is of 198 See Dasa-kumara-carita, Ucchvasa VI, p. 149- ahvayasya nagarasya bahyodyane vasantya vindhyavasinyah 199 See above p. 45, with footnote 123. ayatane suhmesu damalipt vismrta-vindhya-raga-vasarp

THE SAKTA UPAPURANAS 81 comparatively late date. So far as we have been able to find, it is our Devi-purana which is the earliest extant work to mention this rite in connection with Durga-puja. The other comparatively early Puranic works to prescribe it are the present Agni 200 and Garuda-purana, the Mahabhagavata and the Kalikap. Of these, the first, viz., Agni-purana, says in connection with the worship of sixteen- or eighteen-handed Durga annually on the Mahanavami Tithi: "tasyagrato nrpah snayac chatrum pistamayam haret / dadyat skanda-visakhabhyam .... .... (185.13-14). yyn "In her (i.e. Devi's) presence the king should take his bath, destroy [the figure of] an enemy made of powdered rice, and give [it] to Skanda and Visakha.. " In much the same words (especially as those of Devi-purana 22. 16) the Garuda-purana also says in the same connection: "tasyagrato nrpah snayac chatrum krtva ca paistajam / khadgena ghatayitva tu dadyat skanda-visakhayoh //" (I. 135.3=Jivananda Vidyasagara (Calcutta)'s ed.I.134.2).201 It is to be noted that these two verses of the Agni and the Garuda-purana (and more particularly that of the latter) agree very much with Devi-purana 22. 16 quoted above, and have even the incorrect Samdhi in 'tasyagratah' (for 'tasyah agratah') like the Devi 202. Moreover, chap. 134 of Garuda-purana I, which also deals with Durga-puja on the Mahanavami Tithi, has one line (6 b) tallying completely with Devi-purana 22. 10 a and two more (6 a and 7 a) agreeing remarkably with Devi-purana 22. 9 a and 10 b respectively. So, there is little doubt about the fact that both the Agni and the Garuda-purana had their 200 This is a spurious work available in more printed editions than one and quite different from the genuine Agneya-purana still surviving in Manuscripts under the title "Vahnip.' For information about this genuine Agneya-purana (alias Vahni-purana) see Hazra in Our Heritage, I, 1953, pp. 209-245 and II, 1954, pp. 77-110, and Journal of the Oriental Institute, Baroda, V, 1955-56, pp. 411-416. 201 Jivananda Vidyasagara (Calcutta)'s ed. wrongly reads 'sakram krtva ca paistakam' for the second half of the first line. 202 We shall see hereinafter that the language of the Devi-purana, unlike that of the other works of its class, is characterised by grammatical incorrectness at every step very much like that of the Mahavastu, Lalitavistara, and several other Sanskrit works of the Buddhist literature. 6

82 STUDIES IN THE UPAPURANAS verses on satru-bali composed on the basis of that of the Devi-purana Now, the Agni-purana was compiled either in the eastern part of Orissa or in the western part of West Bengal (and most probably in the latter) 203, and the Garuda-purana originated 203 The earliest reference to the (spurious) Agni-purana occurs in the Danasagara of Vallalasena, a ruler of Bengal, who, by way of explaining why he did not utilise the contents of some Puranas, or particular parts thereof, in his great digest on gifts, says about four spurious Puranas including the present Agni and Garuda: tarksyam puranam aparam brahmam agneyam eva ca / trayovimsati-sahasram puranam api vaisnavam // sat-sahasra-mitam laingam puranam aparam tatha / diksa-pratistha-pasandayukti-ratna-pariksanaih // mrsa-vamsanucaritaih kosa-vyakaranadibhih / asamgata-kathabandha-paraspara-virodhatah // tan minaketanadinam bhanda-pasanda-linginam / loka-vancanam alokya sarvam evavadhiritam || (Danasagara, p. 7, verses 63-66). From the contents of these works (viz., initiation, consecration of images etc., Tantric practices, testing of gems, false genealogies and accounts of kings and sages, lexicography, grammar, etc.) as mentioned in the above verses and from the fact that these works were utilised by Minaketanas and other Tantriks for deceiving people, it is sure that the Agni and the Garuda-purana, rejected by Vallalasena, are the same as our present Agni and Garuda-purana published so often as genuine Puranas from different parts of India. Being compiled some time during the ninth century A.D. this Agni-purana began its fight against the genuine Agneya-purana for occupying the latter's position and was soon able to attain much popularity especially among those who had been influenced by Tantricism. But the apathy, or rather hatred, which the comparatively early Smrtiwriters bore towards the Tantras, prevented this Purana from attaining a position of general acceptance as the genuine Mahapurana spoken out by the Fire-god. Thus, in the twelfth century, as Vallalasena's evidence shows, this Purana was trying in Bengal to hold the ground against the genuine Agneya-purana and took a few centuries more to attain a position of authority among the Smrti-writers. In the meantime it was gaining ground in Orissa, where Visvanatha Kaviraja (1300-1384 A.D.), the first known scholar to recognise its authority, quoted from it two verses (viz., 337.3 and 338.7) in his Sahityadarpana, Pariccheda I, p. 2. The next writer to utilise this spurious Agni-purana also hailed from Orissa. He was Narasimha Vajapeyin, a Smrti-writer who flourished about the middle of the fifteenth century A. D. Even in the first quarter of the sixteenth century this work was able to attain only partial recognition in Bengal, where Govind- ananda Kavikamkanacarya (1520-1560 A.D.) was the first writer to look upon it as an authority and to utilise its contents profusely in his works. (For the above date of Govindananda see Hazra in Journal of Oriental Research, Madras, XVIII, pp. 108, and for the list of the verses quoted by him from the 'Agni-purana' or 'Agneya' and traced in the present Agni-purana, see R. C. Hazra, Studies, p. 318, and Our Heritage, I, 1953, p. 244). It is interesting to note that outside Bengal and Orissa this work attained recognition of scholars much later and took a few decades more to be looked upon as an authority in Western and Southern India, and in Mithila it was not even referred to by any writer down to the beginning of the seventeenth century A.D., if not later. So, there is little scope for doubt that the spurious Agni-purana originated somewhere in Eastern Orissa or Western Bengal. The complete silence of Maithila writers about 97 M

THE SAKTA UPAPURANAS 83 in Mithila, but most likely in that part of Bengal which was adjacent to Mithila 204. As to the Mahabhagavata and the Kalika-purana, we shall see hereinafter that the former, which mentions the rite of satru-bali in the verse "tatah satrum balim dadyat krtva pistamayam mama / navamyam pujitaham tu balibhir vividhair api //" (45.33), was written in Bengal (and most probably in its eastern part) some time during the tenth or eleventh century A.D., and the latter, which gives a description of the rite in chap. 71, verses 177 ff., was compiled in the tenth or the first half of the eleventh century A.D. either in Kamarupa or in that part of Bengal which was very near to it. this Agni-purana down to the beginning of the seventeenth century A.D. (although some of them, such as Candesvara, Vidyapati and others, recognised the authority of Tantras much earlier) and the fact that in spite of the mention of Kamarupa as a sacred place on two occasions (in Agni-purana 143.7 and 144.19) the Brahmins of that place (and not those of Bengal or Orissa) have been included in Agni-purana 39. 6-7 among those who were considered unfit for consecrating the images of Visnu according to the Pancaratra method, tend to show that the Agni-purana was compiled neither in Mithila nor in Kamarupa. On the other hand, the facts that Purusottama (modern Puri in Orissa), mentioned more than once in the Agni-purana (viz., in 117. 63, 174. 18, and 219. 68), has been located in 'Audra' (in Agni-purana 305. 13), that no Smrti-writer of Orissa makes any mention of satrubali in connection with Durga-puja, and that the name 'bhelakhi' for a particular type of mystic Vidya (widely known in Bengal under the popular name 'bhelki') occurs in Agni-purana 133. 40-41 and 143. 3, make it very likely that this Purana was compiled in the western part of West Bengal. *04 Like the spurious Agni-purana, the present Garudda-purana also was mentioned, for the first time, by Vallalasena of Bengal but rejected by him as unauthoritative for its spurious character and the Tantric elements in its contents. (For the relevant verses of Vallalasena's Danasagara see the immediately preceding footnote). Its authority was recognised, first of all, by Vidyapati (1395-1440 A.D.) and Vacaspati-misra (1425- 1480 A.D.), both of whom hailed from Mithila and drew upon it in their Gamga-vakyavali and Tirtha-cintamani respectively. In Bengal, it was Govindananda Kavikamkanacarya (1520-1560 A.D.) who was the first Smrti-writer to recognise its authority and to draw upon it profusely in his Sraddha-kaumudi and Varsa-kaumudi and once in his Suddhikaumudi. Following him, Raghunandana, Gopalabhatta (author of the Hari-bhaktivilasa), and others also drew upon this work. In Orissa, it was not recognised till the beginning of the seventeenth century A.D., after which Gadadhara utilised it in his Kalasara. On the other hand, the Garuda-purana names, among the great holy places of India, not only Ekamra, Viraja and Sri-purusottama of Orissa but also Kokamukha, Kamarupa and Pundravardhana with the particular mention of the deities Kamakhya and Karttikeya of the last two places respectively. (See Garuda-purana I. 81. 11-17). So, the Garuda-purana, undoubtedly a work of Eastern India, was compiled either in Mithila or in the western part of Bengal (and most probably in the latter).

84 STUDIES IN THE UPAPURANAS Again, the earliest Smrti work to mention the said rite of satru-bali in Durga-puja is Laksmidhara's Krtya-kalpataru, but its only authority is the Devi-purana, from which it quotes, in its Raja-dharma-kanda, pp. 192-5, verses 3-24 of chap. 22 (including the verse on satru-bali). In his Durgotsavapaddhati Udayasimha Rupanarayana also quotes many verses from the Devi-purana, and says (on fol. 15 b): "anantaram satru-pratimam pistamayim krtva khadgena tac-chiras chedayitva 'om skandaya namah' iti skandaya siro dattva 'om visakhaya namah' iti visakhaya kabandham dadyat". Although Udayasimha Rupanarayana does not quote Devip. 22.16 (on satru-bali), his wordings show definitely that his authority for this rite was the Devi-purana In his Vidhanaparijata (II, p. 651) Anantabhatta also quotes the said verse of the Devi-purana (and no other authority) in connection with satru-bali in Durga-puja. In Eastern India it is the Samvatsara-pradipa which is the earliest Smrti work to include this rite in the annual worship of Durga. It was written in the twelfth century A.D. by Halayudha, a high state officer in charge of religious affairs under king Laksmanasena of Bengal; and its relevant lines (given without quoting any authority) are as follows: ... "tatah sali-pistaka-satru-pratikrtim nirmaya mandapa-dvara-parsve uttara-sirasam samsthapya lohitapuspena sampujya tiksna-khaddgam adaya amukamukam marayamiti chittva skanda-visakhabhyam nama iti nivedayet." (See fol. 25 b). There is also another Smrti work of Bengal, viz., Durga-pujapaddhati of Vidyabhusana Bhattacarya, in which there is mention of satru-bali (see Durga-puja-paddhati, folios 46 b- 47 b). Of the Smrti-writers of Mithila it is Candesvara and Vidyapati who mention satru-bali in their Krtya-ratnakara (pp. 353, 360) and Durga-bhakti-tarangini (pp. 35, 197) respectively, the authority cited by them on this rite being the Devi-purana only. The influence of the Bengal Nibandhas on the section on Durga-puja in Candesvara's Krtya-ratna kara and Vidyapati's Durga-bhakti-tarangini is obvious. It

THE SAKTA UPAPURANAS 85 is specially remarkable that in all cases the non-Bengal Nibandhas use the Devi-purana as their only authority on satrubali, that all the Puranas, except the Devi, which mention this rite, connecting it with Durga-puja, belong either to Bengal or to places very near to it, and that this rite is found to be followed in some form or other in Bengal even at the present day. From all this it appears that both the rite of satru-bali in Durga-puja and the Devi-purana recording it originated in Eastern India and very likely in Bengal. (6) The Devi-purana names, in some cases more than once, Kamarupa, 205 Kamakhya 206, and the different parts of Bengal, viz., Vanga 207, Radha 208, Varendra 209, Samatata 210 and Vardhamana 211, and thus shows its familiarity with and partiality for this part of Eastern India. (7) Among the few chief holy places sacred to Devi this Purana recognises and names 'Ujjayani' (a Sakta holy place in the district of Burdwan in West Bengal) and its presiding deity Ujjani (popularly known as Ujani) 212, which, being of local importance, are found mentioned only in the works of Bengal 213. (8) As in Bengali versification, long vowels are often to be read as short for the sake of metre, and vice versa. See, for instance, Devi-purana 8.3 (brhaspatina cakhyatam, in which the final 'a' of 'brhaspatina' is to be read as short), 14.25 (tathapi kalo gada tam mumoca, in which the 'a' of 'gada' is to be read as short), 85.48 b (kunta-karpura-yastibhir bhusundi-hala-mudgaraih, in which the 'i' of '-yasti-' is to be read as long), and 85.49 b (salakaih silimukhaih sulaih pattisair musalair halaih). See also Devi-purana 2.10 a, 3.24 a, 3.25 a, 4.1 a, and so on. 206 Devi-purana 42. 8; 46. 71. *** Ibid., 39. 6 and 144. 107 Ibid., 46. 69. *** Ibid., 39. 144. *** Ibid., 39. 144; 42. 9. *10 Ibid., 46. 70. *11 Ibid., 46. 70. *1* Ibid., 38.8-ujjayanyam tu ujjani jambu-marge tatha sthita. *1* For the exact location of this 'Ujjayani' and the names of the works mentioning it see footnote 125 above.

86 99 STUDIES IN THE UPAPURANAS (9) In the Devi-purana there are many words and expressions which are clearly based on those in popular use in Bengal. As instances, only a few are noted below:- (i) The inflected form 'devya' of 'devi' has been used in a large number of places as a basic word and declined like 'lata' and other feminine words ending in 'a'. [It should be mentioned here that in rural Bengal the word 'devya' (for ('devi') is popularly used as a title of Brahmin widows even at the present day.] Similarly, 'mata', 'duhita', 'trata', 'data', etc. have been used on several occasions as basic words respectively for 'matr", 'duhitr', 'tratr', 'datr' etc.-See Devi-purana 6. ii. 44; 17.17; 43.18; 7.38; 31.36; and so on. (i) As in Bengali, the word 'sammata' has been used to mean 'the person who has agreed or given his consent' (and not 'the matter agreed upon or consented to') in Devi-purana 7. 96-evam tah sammatikrtva . ../ niyuktah sambhuna • • . || (Cf. Bengali-'sammata kariya'). (iii) Like the word 'ki' (for 'kimva' meaning 'or') in Bengali, the Devi-purana uses 'kim' in 8. 12- ... • dvare | naradas tisthate deva sthapyatam kim pravesyatam // Cf. Bengali a, dvare ... narada avasthana karitechena [tamhake sekhanei ] rakhite haive, ki, pravesa karaite haive ? ( Note the word 'sthapyatam' for Bengali f'). (iv) The verbal forms 'uttha' (for 'uttistha') in Devi-purana 8. 17 (uttha rajan mahabaho ...) and 'karanti' (for 'kurvanti') in Devi-purana 35. 27 (... badham karanti ca), are very similar to their Bengali equivalents '35' (rise) and '' (present tense, third person, plural number of the Bengali root for Sanskrit 'kr'). [See also the forms 'utthan' (for 'udatisthat') in Devi-p 4.31, 'samutthanti' (for 'samuttisthanti') in 85. 33, 59, 'utthantam' (for 'uttisthantam') in 85. 43, and 'uttha' (for 'utthaya') in 85. 58.] (v) As in Bengali, the Sanskrit root 'bhuj' (meaning 'to enjoy') has been used in the form 'bhunj' in many places of the Devi-purana See, for instance, Devi-purana 2.42 (svarga-bhusapta-patalan bhunja tvam...), 2.49, 5.16 (bhunja svargam),

THE SAKTA UPAPURANAS 87 4.10 (bhu-patalani bhunjatha), 9.12 (bhunjate prthvim), 9.36 (bala bhunjaniya), 9.39 (bhunjan prthvim), 9.45 (bhunjayitva), and so on. (vi) On numerous occasions, nouns derived from verbal roots by means of Krt affixes denoting action have been allowed to govern objects having the second case-ending. For instances of such use we may refer to Devi-purana 1. 57 (sravanat puranam siva-bhasitam, in which the word 'sravana' governs the Accusative Case in 'purana'), 8.57 (kanyam udvahanopari -about marying the girl), 9. 42 (tam udvaha-samutsukah -eager to marry her), 9. 50 (vidya mohana-sila ya sa-surasura-manavan), 11. 22 (divam utsahate jaye), 11. 57 (evam sakrasya brahmena kathitam ketumucchrayam?), 12. 19 (pankam uddharanam kupe), 13. 22 (asthini darsanam), 15.2 (-bhitan pariraksanaya), and so on. Similar use of Krd-anta nouns denoting action is very common in Bengali and has its root in the Vedic literature, in which there are a few instances, viz., Rg-veda VIII. 11. 7-agne tvam-kamaya gira (which Sayana explains as 'tvam-kamaya tvam abhilasantya gira stutya'), Atharva-veda VI. 139. 2, 4-mam kamena-by loving me (Whitney), Satapatha-brahmana II. 3. 1. 16-yady enam ksiram kevalam pane 'bhyabhavet-if it should happen to him to have to drink pure milk (Eggeling), and so on. In the Ramayana and the Mahabharata also there are several instances of this type (for which see Whitney, Sanskrit Grammar, p. 273). (vii) The forms 'jvalanta' (for 'jvalat'), 'bhramanta' (for 'bhramat'), 'vahanta' (for 'vahat'), etc. have been used as basic words in Devi-purana 14. 27 (jvalanta-kopah), 84. 14 (jvalantagni-), 85.38 (bhramantavrtta-kutilam), 85.39 (vahantaih), etc. (Cf. Bengali '-blazing fire, Bengalidad 'calanta gada़ी ' - moving car, 'ghumanta sisu ' sleeping child, and so on). The form 'tejin' (for 'tejasvin'-spirited) occurs in Devi-purana 53.1 (srastaram tejinam. ...). (Cf. colloquial Bengali a spirited man). 'teji loka ' (viii) As in Bengali, the word 'purve' (with the seventh

88 STUDI IN THE UPAPURANAS • case-ending) has been used in Devi-purana 6. ii. 1 (maya purve ca tvam deva uktam asit ..) in the sense of 'formerly', 'previously', and the word 'kapala' appears in Devi-purana 33.31 in the sense of 'forehead'. (ix) In several places of the Devi-purana the seventh caseending has been used in place of the fourth in the Dative case. See, for instance, Devi-purana 11. 7 (vasisthenapi sa datta sarasvate . . . .; bahvrcas tarune dadau, 'tarune' being used for 'tam+arune'; and so on), 106.5 (devya bhakte pradatavyam). ... • Cf. Bengali 42 greno 919 fae q" (Bharata-candra's Annada-mangala)- ... gave [me] to such a bride- ...; 'satpatre dana kara ucita ' - [one] should make gifts to a worthy recipient; and so on. groom • •• On one occasion the seventh case-ending has been used in place of the third, viz., in 14.4-mahamohe 'timohite (for '-mohenatimohite'). Cf. afflicted with grief, cokhe dekha - seen with the eyes, hate kara --- handmade, kale tairi machine-made, and so on. (x) As in Bengali, words have sometimes been used in different cases without any case-endings at all. See, for instance, Devi-purana 3.12, 21 (mahasangrama cakrire), 14. 20 (vamam bhuja purayitva dhavati kruddha kopat), " • ... 14. 25 (tathapi kalo gada tam mumoca), 14. 27 (evam sa kalo hata bhairavas ca ..), and so on. ... (xi) There are a number of cases in which adjectives, qualifying nouns in the feminine gender, have been used, as in Bengali, without any case-ending or change of gender required by the nouns they qualify. See, for instance, Devi-purana 2. 40-param astanga devasya bhaktim 9. 29-evam tasya matir bhuta narada-pathaga.. (xii) The words 'astasi' (for 'astasiti') in Devi-purana 12.31 (astasim tu ...) and 'panca-pancasa' (for 'pancapancasat') in Devi-purana 17. 8 have their parallels in colloquial Bengali (viz. 'astasi ' or 'atasi ', and 'pamcapancasa ' ). ... " and so on. (xiii) As in Bengali, the a of a final or intermediate syllable of a word, though existing in writing, is on many occasions not taken into account in pronunciation. See, 17

THE SAKTA UPAPURANAS 89 for instance, Devi-purana 13. 57-ingud-odumbaram kharjuramatulungaih sadadimaih (in which the final a of 'kharjura' is not to be pronounced); similarly, 32.34-parasu-cakradharau canyau damaru-darpanna caparau, 32.36- kurvanti kalakalaravaih, 39.159-sarabha-sankhau hatau drstva ... ... and so on. (xiv) As in colloquial Bengali, the final Visarga of a word is sometimes not to be pronounced at all (for the sake of metre). See, for instance, Devi-purana 27.21 . . . amrtam havih kamikam (in which the Visarga of 'havih' is to be ignored in pronunciation), 31.39-sa sukham yasah saubhagyam... . pulahadibhih sattamaih, and so on. 36.34- • (xv) In a number of cases the a of a final or intermediate syllable of a word has been eliminated in writing, and for such elimination popular Bengali pronunciation must be held responsible. For instances see Devi-purana 39.15-parsuramena (for 'parasuramena') tosita, 43.8-tada tvam parsuna (for 'parasuna') hatah, 43.19-parsuh (for 'parasuh³) pibati sonitam, 98.27-balagbhir (for 'balakair') na ca atmiyam and so on. (xvi) As in Bengali, the word 'vina' has been allowed to enter into an Avyayibhava compound with a substantive and the seventh case-ending has been added to the final member. See, for instance, Devi-purana 51.21-bali-homakriyadini vina-patre na sidhyati (vina-patre vina patram). Cf. Bengali 'lokatira vina-dosa sasti haila ' - the man was punished for no fault [of his ] ; 'se vina-kharace sara desa ghuriyache ' - he has travelled without expense throughout the whole country; and so on. === (xvii) The irregular compounds 'nrpa-saha', 'rocana-saha' etc. (in Devi-purana 12.56, 35.7, and so on), in which the indeclinable word 'saha' (meaning 'with' or 'together with') stands as the final member, have their parallels in Bengali. (xviii) In Devi-purana 86.30 (nrtyante paramo devo asmakam saha vasava) the sixth case-ending has been used in connection with the word 'saha'. In Bengali also, the sixth case-ending is employed in connection with the words 'sange' and 'sathe',

90 STUDIES IN THE UPAPURANAS which are the indeclinable synonyms of 'saha', e.g. va< ad (or, ) musi ganfod'-I had a quarrel with Ram. (xix) As in Bengali verse (especially of the early and the mediaeval period), the word 'darsana' has been used with svara-bhakti as 'darasana' in Devi-purana 10.ix.11-darasanam api punyam yoga-marga-sthitanam. It is needless to cite more instances. From the evidences adduced above it is clear that the Devi-purana is an East Indian work written in Bengal. The use of the simile of ships (pota) in more places than one in this Purana, 214 tends to show that it was written somewhere about Tamluk, where, as we have already seen from the Dasa-kumara-carita, there was a famous temple of Devi Vindhyavasini and whence ships were sent to distant countries outside India. Although in the Devi-purana, Devi appears mainly as a wargoddess 215, the method of Devi-worship, as given in this Purana, is in many respects very different from that found at present in Bengal. The Devi-purana mentions satru-bali but is silent about the customs of nava-patrika-puja (worship of the nine plants), Devi's bodhana and worship in a Bilva tree, and the Savarotsava. This disagreement between the method of Devi-worship given in the Devi-purana and that found in present-day Bengal, can be explained away by saying that the above-mentioned customs (except satru-bali), which are not found mentioned in any comparatively early work, can scarcely claim to have grown as early as the Devi-purana and that they were most probably due to the influence of Kamarupa. The direction for the worship of many gods and goddesses, especially of Brahma, Visnu, Siva, Surya and Ganesa during Devi-worship, shows that the attempt at creating a sense of 214 Devi-purana 39.35 (pankodadhi-nimagnanam atipota bhava tvaran); 72.73 ('poti potasya va yatha' used in connection with the maintenance of a fort by its owner); 77.3 (.. .. siva-potam tu tena param bhavarnavat); 77.15 (potam navaplavam khyatam); 83.36 (isu-cakr-asi-makare tvam poto bhava acyuta); 83.112 (daityaugha-majjamananam tvam pota bhava sulini). 215 It is for this reason that the Devi-purana is said to have been introduced on earth by kings and that it deals, among other topics, with warfare and the construction of towns and forts and the method of protecting these.

THE SAKTA UPAPURANAS 91 brotherhood among the mutually conflicting sects among the Hindus began at an early date, and that such directions of the early Puranic works must have been incentives to the introduction of the 'pancayatana-puja' by the elimination of Brahma, whose worship grew unpopular in later times. The Devi-purana supplies us with important information regarding the different branches of Sanskrit literature known to it. Besides Itihasa, Kavya, Nataka, Akhyayika, Jyotih-sastra, Vaidya-sastra, Vina-sastra (93.65), etc. it frequently mentions a Devi-sastra (literature on Devi) 216 which grew up before the composition of this Purana; it refers to 'Puranas' dealing with Devi¹7, thus showing that Devi had become the subject-matter of Puranic works even before the present Devi-purana was composed; and it speaks of the four Vedas as well as the Upavedas, Angas etc.,218 of Usanas as an authority on warfare and construction of 'nadidurga' (river-fort) 219, of Mathara as an authority on the method of Sun-worship 220, of the 'Siva-siddhantas' on Yoga and worship 221, of the 'Sivagamas' as authorities on the method of worship of the divine Mothers 222, and of the political works spoken out by Sambhu, Usanas, Visnu and It not only mentions 'Tantra' and 'Agama' very frequently 224 but names a good number of Tantric works, viz., Kala-tantra 225, Garuda-tantra²², Matr-tantra²²7, Brahma 223 216 See Devi-purana 34.3 and 8; 89.14; 91.7 (devy-avatara-sastrani); 99.17 (nanda- sastra); 101.20; 105.9 (mangala-sastra). 17 Devi-purana 37.74; 39.22 218 Devi-purana, chap. 107. 219 Devi-purana 3.33; 72.55. 220 Devi-purana 51.8. The meanings of the word 'mathara' are given by Vikrama Samvat Apte as (1) name of Vyasa, (2) a Brahmana, (3) a distiller (saundika), and (4) one of the attendants on the Sun. It is most probably in the last sense that the word 'mathara' has been used in the Devi-purana This word seems to have nothing to do with the name of the author of the Mathara-vrtti on the Samkhya-karika. 221 Devi-purana 10.4; 51.8. 112 Devi-purana 88.1. 233 Devi-purana 5.2. *24 Devi-purana 6.30; 10.4; 33.49; 39.24-25; 56.22; 83.1 and 17; 110.3. 225 Devi-purana 6.31; 88.1. 22 Devi-purana 6.31; 32.43; 88.1; 91.14. 227 Devi-purana 39. 25.

92 STUDIES IN THE UPAPURANAS , , Mula-tantra 228, Bhuta-tantra 229, Bhairava-tantra 230, Balatantra 231 and Nitya-tantra 232. In Devi-purana 76.39 there is mention of a 'Markandeya-munisrestha-purana' dealing with topics about the Narmada and the Sarasvati; but we have not been able to identify this work. It is also not known which Purana is meant by Devi-purana 63.18 (asta-sastis tu namani deva-devasya dhimatah purane copagitani brahmana ca svayambhuva //). An extract, quoted by Raghunandana in his Smrti-tattva, I, p. 668 from the 'Devi-purana' but not found in the present Devi, contains the name of a 'Brahma-purana' in connexion with a mantra to be pronounced during the worship of Visvakarman, the divine architect. In this connexion we should like to say a few words on the Tantric elements in the Devi-purana The Vedic influence on this work is quite evident from the facts that the contents of the Devi-purana are said to be 'vedartha-tattva-sahita', that the four Vedas are called Devi's door-keepers, that the muttering of the Gayatri is prescribed during Saiva-yoga, that the Vedas are called the highest authorities on dharma, that those who violate the rules of castes and orders of life are doomed to hell, that the worship often requires Vedic mantras, kunda, homa, etc., that the Brahmins are engaged as priests, and so on 233. Yet this Purana is infused with Tantric influence from beginning to end. It not only recognises the Tantra as an authority 234 but mentions 'Tantra' and 'Agama' on several occasions and names a good number of Tantric works. The study of a large number of Tantras is even called the best of all purifiers 235. The form of worship, which is often mainly Tantric, *** Devi-purana 9.65. 29 Devi-purana 32.43; 88.1; 91.14. The Bhuta-tantra has been mentioned by Varahamihira in his Brhat-samhita 16.19. *** Devi-purana 39.25; 91.14. 331 Devi-purana 39.146; 91.14. Devi-purana 55.8. 233 See Devi-purana 1.55; 7.22; 8.6; 10.v.9; 12.26; chapters 25-26; 35.23; 72.26; 73.8; 82.11; 93.45; and so on. 234 Devi-purana 55.8; 56.22. 235 Devi-purana 110.3-bahu-tantravalokanam vimalikaranam.

THE SAKTA UPAPURANAS 93 requires the use of Tantric mantras and the performance of nyasa, mudra etc. 236; the use of wine and meat is prescribed in some cases 237%; the feeding of virgin girls is regarded as a very important part of all worships and festivals connected with Devi; the image (pratima), altar (sthandila), sword (khadga), trident (sula), circle (mandala, with the figure of a lotus padma - in it), book (pustaka), phallus (linga), foot-wear (paduka), cloth (pata), knife (churika), arrow (bana), water, fire, heart (hrdaya), picture (citra), and bow (dhanus) are to be used as mediums of worship in different cases 238; and so on. The Devi-purana mentions many mystic lores (vidya), viz., Kamika, Padamala, Aparajita, Mohini, Mrtyunjaya etc.,239 which are intended not only for the attainment of magic powers such as akarsana etc. but also for final emancipation (mukti) 240. It elevates the position of the spiritual preceptors (guru) by prescribing their worship which is said to be as important as, or more important than, that of the gods, Fire and the Vidyas²¹¹. As to the position to be allowed to women and Sudras, the Devi-purana holds a liberal view. It allows them as well as the Pukvasas, Candalas and others to perform the vows and worships connected with Devi 242 and even prefers a qualified Sudra to a worthless member of any of the higher castes 243 Women and Sudras are directed to perform homa daily in a sanctuary (matha) or a kitchen (mahanasa) but never in a kunda 244. Virgin girls are to be worshipped and fed in all the worships and festivals connected with Devi, because Devi, being herself a maiden, resides in all virgin • 236 Devi-purana 7.65-66; 7.88; 9.56; 9.68; 26.21-22; 29.14; 50.i.91; 50.iv.71; 52.8; 77.21-22; 91.40 f.; 93.267 f.; and so on. 237 Devi-purana 39.173; 50.iii.18, 28 and 30; 118.6; and so on. The Vira-vrata requires the use of 'mahamamsa' in homa.-Devi-purana 9. 68-69. 288 Devi-purana 22.10; 26.35; 31.18; 50.i.61; 50.iv.42 f. and 100; 54.10; 57.13; 58.19; 91.40 f.; 93.102-106; 98.7-9; 98.18-19; 100.5; 102.11-12; and so on. 65.56 f.; 239 Devi-purana 1.63-64; 2.1; 9.49 and 52; 11.4; 76.31; and so on. 240 Devi-purana 39.27; 39.122-4. 241 Devi-purana 67.5; 124.1; chap. 125. Also cf. 39.170; 128.24 and 26. 243 Devi-purana 22.5-6; 22.23-24; 24.17; 88.4; 89.19; 91.1; and so on. 243 Devi-purana 51.4-5. 244 Devi-purana 121.5.

94 STUDIES IN THE UPAPURANAS girls 245 Other women also are held in high esteem 246, so much so that their worship is prescribed on several occasions 247. As has already been said, the highly incorrect language of the Devi-purana shows characteristics which are unique in the Puranic literature but have very often their parallels in the Mahavastu, Lalitavistara and other Buddhist works written in hybrid Sanskrit 248. Although these linguistic peculiarities are too numerous to be noted here exhaustively, some striking cases are referred to below for giving a general idea of the highly varied nature of these, with the citation of a limited number of instances of each type taken mainly from the first few chapters of the Devi-purana I. Unlawful 249 changes made not infrequently in the stems: (i) Substitution of the final i of 'sakhi' (friend) with aya, as in sakhaya (for 'sakhi') in 39.146250 - ... . graha-rajasya sakhayatve vyavasthitah, 95.46-susakhayaih sahaiva tu, 95.44-sakhayanam tu laksanam, and so on²51. (ii) Substitution of the final ? (of masculine words) with a, ara, ara or u, as in bharta (for 'bhartr') in 9.10-bhartanupalakah; hota (for 'hotr') in 11.33-yajna-hotaya namah; karta (for 'kartr') in 26.34-bhuti-kartaya namah; *45 Devi-purana 35.17-18- kanya devya svayam prokta kanya-rupa tu sulini / yavad aksata-yonih syat tavad devya surariha // *** Devi-purana 93.165 f. 247 Devi-purana 22.19; 90.21; 91.61. 248 For information about Buddhist hybrid Sanskrit and its literature see F. Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary (Volume I--Grammar, and Volume II-Dictionary). Although the name 'hybrid Sanskrit', given by Edgerton to the mixed language (Prakrit-Sanskrit) of the Mahavastu, Lalitavistara, etc., is not very satisfactory, we have used it here from various considerations. 249 So far as Sanskrit grammar is concerned. 250 It is needless to say that this and similar other references below are to the Devi-purana 251 For the use of 'sakhaya' for 'sakhi' in Buddhist works written in hybrid Sanskrit sec Edgerton, op. cit., I, p. 69 (10.8).

THE SAKTA UPAPURANAS 95 pitara (for 'pitr') in 50.i.72-pitara-matarah, 91.50 matarah pitaras tatha; hotara (for 'hot') in 121.27-avedakas tu hotaro naiva prinati vai suran; vettara (for 'vettr') in 32.42- tacchastra-vettaraih (Instrumental, plural), 121.6-mahavidyartha-vettaraih; vidhatu (for 'vidhatr') in 7.9-jaya dhindriya-vidhatave (Dative for Vocative); and so on 252 -bud- (iii) Substitution of the final ? (of masculine words) with a (the resultant form being used in the feminine gender), as in data (for 'datr') in 31.36-sivayah siva-datayah (for siva-datryah', the word 'data' having been used in the feminine gender, for 'datri', on account of its final 'a'); stota (for 'stotr') in 6.ii.32-33- ... devi ... stota tvam ca stutis tvam ca (for '. . . devi . . . stotri tvam. ..'); trata (for 'tratr') in 7. 38-tratas tah sarva-devanam (for 'tratryas tah .. .'); and so on. It may be that of the words 'datr', 'stotr', 'tratr' etc. the final r was first replaced by a (as in 'hota', 'karta' etc. mentioned above) and then the feminine suffix a was added for turning them into feminine bases. So far as we have been able to find, the Buddhist works written in hybrid Sanskrit do not contain a single instance of the above use of masculine words ending in 'r'. (iv) Substitution of the final i of certain monosyllabic feminine words (such as 'dhi', 'sri' and 'stri') with iya or iya, as in dhiya (for 'dhi') in 93.111-martyair jada-dhiyair vatha . . . .; sriya (for 'sri') in 7.51-sriya rupe sthita tasya, 50.i.10-pustih sriya uma, 50.i.84-tesam ayuh sriya, 50.i.115 karya padmasana-vyavasthita, sriya 252 For similar use of 'napta' (for 'naptr'), 'mandhata' (for 'mandhatr'), 'yasta' (for 'yastr'), 'karta' (for 'kartr'), 'bhratara' (for 'bhratr'), 'pitara" and 'pitu' (for 'pitr'), 'bhartu' (for 'bhartr'), 'sastu' (for 'sastr)', and so on in Mahavastu, Lalitavistara, etc. see Edgerton, op. cit., I, pp. 90 (13.3-6) and 91 (13.16-17, 21-28).

96 STUDIES IN THE UPAPURANAS 66.29-sriya caiva tathoma ca; striya and striya (for 'stri') in 101.27-striyaih karyam, 9.68-striyakarsanam, 67.74-striya va laksanopeta, 91.75-striya vanena vidhina vidya-dana-phalam labhet; and so on 253 • (v) Substitution of the final ? (of feminine words) with a, ara, or ara, as in duhita (for 'duhitr') in 43.18-agastya-duhitam devim; mata, matara or matara (for 'matr') in 6.ii.44-suhrdbandhu-mata-balabala-vadham, 17.17, 106.6-vedamate (Vocative), 119.24-vira-mate (Vocative), 7.84 matara-samnidhau, 59.11-ravi-matara-rupa ca, 86.20 matarah, 86.4, 6, 8, 10, 14- -mataraih, 88.11-devamatarotpattim, 26.13-matara bhujagambika, 45.6 matarasu ca (also 7.96-nayika matarartihah, 22.21, 65.90-mataranam, 26.17-matara loka-palanam, 53.9- mataranam sada-cakram, 65.81-matara graha-nagas ca, etc.); and so on 254. (vi) Shortening of the final a or i of certain feminine words (very often for the sake of metre), as in dirghika (for 'dirghika') in 82.64, 93.75-dirghikaih; jyotsna (for 'jyotsna') in 32.25-mayukha-danta-jyotsnena cakasanti tadid iva; mekhala (for 'mekhala') in 26.3 mekhalais tribhih, 50.ii.1 - munja-mekhala-bhusita; mudra (for 'mudra') in 52.8-varadabhaya-mudrau tu ...; narmada (for 'narmada') in 32.4-ganga-narmadavindhyadri-; paduka (for 'paduka') in 93.105-paduke pratimasu ca, 98.7--paduke pate, 98.19-paduke vatha khadge ca; pataka (for 'pataka') in 51.27-patakaih; sobha (for 'sobha') in 50.ii.44-sobha-krt; devi (for 'devi') in 14.23-sa devi-ghato hata-bhunisannah, 14.28-te devi-banasani-bhinna-vaksah, 50.i.59 -bhusayed 50.i.61-darsayed devidevi-vedikam, 259 For the use of 'sriya' and 'striya' (for 'sri' and 'stri' respectively) in Buddhist works see Edgerton, op. cit., I, p. 69 (10.4-6). 254 For instances of similar use in Buddhist works see Edgerton, op. cit., I, pp. 90 (13. 7-13) and 91 (16-18).

THE SAKTA UPAPURANAS mandale, 93.269- 97 devi uddharet ...; parvati (for 'parvati') in 43.37-parvati-nandanah; patni (for 'patni') in 33.36-pratyan-mukhah patnisu; and so on 255. (vii) Addition of a after the final i of certain feminine words (such as 'devi', 'koti', 'kumari', 'vedi', etc.), as in devya (for 'devi') in 35.17-kanya devya svayam prokta, 15.17-devya kopanala-sudipitah, 7.37-ta devya rsi-manavah, 9.71-devyayai vinivedaya, 32.14-devyaya bhakti-karanam, 37.100-devyaya esa siddhantah, 92.13 -devyayam; kotya (for 'koti') in 2.37-kotyayuta-gajanvitah; kumarya (for 'kumari') in 12.28-kumaryah prathamamsagah (Nominative, plural); vedya (for 'vedi') in 65.99-pascimayam tu vedyayam ...; and so on 256. (viii) Addition of a to the final consonants mostly of masculine words and rarely also of feminine and neuter ones, as in bhisaja (for 'bhisaj') in 58.5-purodha-bhisajanvitah, 58.7-bhisajah ... laghu-hasto jitendriyah; sraja (for 'sraj') in 50.ii.45-asoka-sraja-homena, 65.2, 99.26, 29- srajaih, 99.34, 101.7-sahakara-srajaih, 101.2-jatimayaih srajaih; anicchata (for 'anicchat'-unwilling) in 9.36-na hi anicchata 257 bala bhunjaniya ('anicchata' being the feminine form of 'anicchata' and qualifying 'bala'); bhavata (for 'bhavat') in 9.31-yathapi bhavato nasman ..... bhasate (bhavatah bhavan); janata (for ''janat' jna + satr) in 4.13-desam caiva na janata (the visarga of 'janatah' being dropped); sampada (for 'sampad') in 50.iv.6-ayuh-sampadaFor instances of similar shortening of long vowels in Buddhist works see Edgerton, op. cit., I, p. 69 (10.10). 256 The occurrence of the inflected forms 'devyah' and 'kotyah' (Accusative plurals of 'devi' and 'koti') in the Saddharma-pundarika and Lalitavistara, and of 'hiranyavatyayam (Locative singular of 'hiranyavati') in Arya-manjusri-mulakalpa (580.9), shows that the use of stems formed by adding a to feminine words ending in i was not unknown to the Buddhists.-See Edgerton, op. cit., I, p. 79 (10.162, 164), and II, p. 620. 257 Some Manuscripts of the Devi-purana read 'anicchati' for 'anicchata". 7

98 STUDIES IN THE UPAPURANAS dayakam; sastravida (for 'sastravid') in 16.12. te deva sastravidasya .; tadvida (for 'tadvid') in 27.17 -deya ca tadvidaih; samidha (for 'samidh') in 50.i.63-samidhani, 55.16- -samidhany etani; atmana (for 'atman') in 8.19-atmanam paramam devam tosaniyam sada budhaih; brahmana (for 'brahman') in 31.39-ity evam brahmano 'bravit, 85.78-brahma- nasya; arthina (for 'arthin') in 32.11-sthapayed devadevesam sarva-kamarthino yadi; asvina (for 'asvin') in 17.28-sasvinan, 26.29-vasu-asvina-rupaya; -dharina (for '-dharin') in 7.81-maharudrah khadga-khetaka-dhari- nah; -harina (for '-harin') in 4.56-. . . deva-devesa umadehardha-harinah / agatah. ||; karina (for 'karin'- elephant) in 50.i.116-karinaih kalasanvitaih; sulina (for 'sulin') in 9.66--vaktum arhasi sulina; diva (neuter, for feminine 'div') in 4.6-divam duhkhena sadhyati, 4.10-na jetum sakyate kala divam dharmena raksitam; disa (for 'dis-direction) in 50.i.52-purva-disaih; nabhahsprsa (for 'nabhahsprs') in 13.56-sala-talair nabhahsprsaih; visa (for 'vis'-a Vaisya) in 51.5- visah (singular), 22.5-visaih; ayusa (for 'ayus') in 27.25-ayusa-kamena; dhanusa (for 'dhanus') in 50.i.42-dhanusanam, 53.2-dhanusavadhi; vapusa (for 'vapus') in 43.50-divya-vapuse (Locative singular); ... angirasa (for 'angiras') in 2.8-bhrgv-angirasa-vasisthamandavyah; chandasa (for 'chandas') in 17.29 -namabhis chandasaih kirtitam ...; divaukasa (for 'divaukas') in 33.17- modanti divaukasa iva ...; jatavedasa (for 'jatavedas') in 54.17-kusmandi-jatavedasaih; pumsa (for 'pums') in 8.44-pumsah saubhagyam apnuyat, 33.109 ... labhate pumsah 44.18 bhavate dhanavan pumsah ..., 108.32-tatha pradhana-pumsabhyam anyah; sumedhasa (for 'sumedhas') in 9.38-raja namna sumedhasah, 9.47-sumedhasas tato yatah; tejasa (for • "

THE SAKTA UPAPURANAS 99 'tejas') in 85.75-sva-tejasaih; vasasa (for 'vasas') in rudrani krsna-pinga ca nila-kauseya-vasasa; 127.85 and so on 258. (ix) Addition of a to the final consonants of certain feminine words (such as 'apad', 'mrd', 'sampad' and 'sruc'), as in apada (for 'apad') in 83.117-smaranad eva stotrasya sramsate mahad-apada; mrda (for 'mrd') in 33.57 mrdaya. angam upalipya. 33.89-mrdambubhih; ... • " sampada (for 'sampad') in 8.25-sampada dharmabhoga hi ; sruca (for 'sruc') in 50.iv.78, 80, 84; and so on 259. (x) Substitution of the final at (of words formed with the Taddhita suffixes 'matup' and 'vatup' and the Krt suffix 'satr³) with anta, as in himavanta (for 'himavat') mahagirau; in 93.10-himavante tavanta (for 'tavat') in 50.i.27-tavanta-bhedena; bhramanta (for 'bhramat') in 85.38-bhramant-avrttakutilam; jvalanta (for 'jvalat') in 14.27-jvalanta-kopah, 84.14 jvalant-agni-latakara-; vahanta (for 'vahat') in 85.39-silimukhair vahantais ca ...; and so on 260. (xi) Substitution of the final an or at with ana, as in svana (for 'svan') in 13.19-svano vanti-mukho bhavet; yuvana (for 'yuvan' or 'yuvat') in 12.22--yuvanair balasampannaih, 50.iii.48 bala yuvana-madhya ca, 91.64 yuvanair api tam neyam; and so on 261. - (xii) Substitution of the final as of some masculine words with a, as in purodha (for 'purodhas') in 58.5-purodha- 258 For similar forms in the Mahavastu, Lalitavistara etc. see Edgerton, op. cit., I, pp. 95 (15. 7-8, 10-13) and 96-97 (16. Iff.). 259 For similar modification of stems in Buddhist works see Edgerton, op. cit., I, pp. 94 (15.1) and 95 (15.9). 260 For instances of the use of similar stems in Buddhist works see Edgerton, op. cit., I, p. 102 (18.5 ff.). 261 For similar use in Buddhist works see Edgerton, op. cit., I, p. 100 (17.37-39, 44).

100 STUDIES IN THE UPAPURANAS bhisajanvitah, 58.10 so on 262. -purodhanumate sthitah; and (xiii) Loss of the final single consonants (following a) of masculine and neuter words, as in atharva (for 'atharvan') in 7.22-rg-yajuh-sam- atharvanam; bhasma (for 'bhasman') in 4.64-sitabhasmena, bhasmat; brahma (for 'brahman') in 7.44 brahma (Vocative, singular), 11.57, 40.11, 83.46 brahmena, 80.10-brahmaya, 6.ii.14, 12.59, 37.83, etc.-brahmasya; carma (for 'carman') in 14.20 carmena; karma (for 'karman') in 34.7-karmena, 122.27 -nama-karmaih, 9.64 siddhanta-veda-karmanam; nama (for 'naman') in 33.96-svasti-namena, 46.83 namena, 16.16-mantra-namaih, 16.20-namaih, 93.287, 289-sva-namaih, 16.19-sarva-namebhyah, 50.ii.11- -name; (for adha kancanamayah, - 'adhas') in - 46.3 adhordhvam 70.7 adha-pattagam, 84.10 bhutadham (for 'bhutadhah'), 87.5-adhordhva-; ambha (for 'ambhas') in 50.iv.94 -kumbhambhaih; angira (for 'angiras') in 46.54-angiradya-; apsara (for 'apsaras') in 14.29-sahapsarah kimnara-caranas ca, 93.244 apsarah, 12.40--apsaradi, 17.20- -yaksapsaraih, 93.217 apsaraih; ayu (for 'ayus') in 79.iii.7- -putrayu-samyutah, 82.50--kalpayu-jivitah; jataveda (for 'jatavedas') in 12.25 jatavedamayena; mana (for 'manas') in 37.51 -manadinam, 39.69-manepsitaphala-pradam, 50.i.40 dig-bhage mana-tustide, 116.iii.10-mane krtva; oja (for 'ojas') in 8.51-sukrauja-balam eva ca, 84.25-tarjayanti mahaujena; raja (for 'rajas') in 37.33-rajadi-triguna mata, 65.59, 61-rajah; raksa (for 'raksas') in 3.21-naga-raksendradaiteyanam, 4.20-nirjita asura raksah patala-talavasinah, 15.6-jalesa-raksaih, 17.28-saraksan; sara (for 'saras') in 65.28-sare, 82.68-dirghika-sara-madhyagah; sira (for 'siras') in 20.30-devyah sire mudgara-pasaFor similar modification of stems in Buddhist works see Edgerton, op. cit., I, p. 97 (16.29).

THE SAKTA UPAPURANAS . . . 101 ghatan, 50.i.53-sarvesam sira-pattastham .. 50.ii.53 -tac-chirottha-, 85.58 -kantha-siroru-; sumana (for 'sumanas'-flower) in 31.25, 65.93 -sumanaih; tama (for 'tamas') in 32.22-tama-vaktrena; tapa (for 'tapas') in 2.37 vidyavams tapavams caiva; teja (for ''tejas') in 2.59-satejas ca sada grahah, 4.32-pasuna tu satejena, 7.22-mahatejah (feminine, plural), 20.24 dipta-tejam tam, 39.102-tam sutejam; usana (for 'usanas') in 5.2-nayah ... ausanah 3.33-ausanoddista-vidhina, 67.70-usanac ca guruh praptah; vaksa (for 'vaksas') in 14.28-te devi-banasani-bhinnavaksa gatasavah, 32.26-pinau vaksau, 43.19 ... te vakse parsuh pibati sonitam; sapta (for 'saptan') in 25.18-pancatha sapta va, 96.4-panca saptas tatha dvijah; astadasa (for 'astadasan') in 2.76-mukto hy astadasair dosaih; dvatrimsa (for 'dvatrimsat') in 17.8 --laksa dvatrirsa; dvipancasa (for 'dvipancasat') in 17.31-dvipancasasamkhyan; pancapancasa (for 'pancapancasat') in 17.8 -asvanam pancapancasa; and so on. 263 (xiv) Loss of the final consonant (following a vowel other than a), as in nandi (for 'nandin') in 85.66-nandim caiva mahabalam; pathi (for 'pathin') in 95.9-tato gaccheta pathim; and so on. 264 • (xv) Loss of the final syllable (of certain words, mostly those denoting cardinal numbers), as in astasi (for 'astasiti') in 12.31 -astasim tu svayam sakro nila-varnam pradapayet; astavimsa (for 'astavimsati') in 54.16-astavimsasta-homam tu....; caturvimsa (for 'caturvimsati') in 107.16-varganam caturvirsa-satani ca; unavimsa (for 'unavimsati') in 107.29 -kathitah parisistas tu unavimsah; kata (for 'kataka') in 93.207-kundalaih kata-keyuraih; 263 For similar forms in the Mahavastu and other Buddhist works see Edgerton, op. cit., I, pp. 99-101 (17.1-64). 264 For similar loss in Buddhist works see Edgerton, op. cit., I, p. 77 (10. 42-43). •

102 STUDIES IN THE UPAPURANAS ulu (for 'uluka') in 65.21--grdhr-olu-kakani, 65.52- -kak-olu-varah-ahi-; and so on. 265 (xvi) Loss of the initial vowels of api and atha, as in pi va (for 'api va') in 12.9-sasi-sakamayam pi va, 26.9 dvi-gunam tri-gunam pi va, 35.6-. . . . . dhatujam pi va, 50. iv. 4-kanakam pi va, 50. iv. 78, 54.15, 65.61, etc.-pi va; thava (for 'athava') in 50. iii. 40-varadam danda-matsyam va vamardhe yuvati thava, and so on. 266 (xvii) Use of pantha (for 'pathin', or for 'panthan'?) in 19.8-yama-panthagah, and panin (for 'pani' placed at the end of a Bahuvrihi compound) in 7.61-krpana-varapanine, 14.12-krpana-panini, 16.40-varad-odyata-panini. And so on. II. Shortening of long vowels and lengthening of short ones (evidently for the sake of metre), as in 4.11-yuddham dhvamks-olukam yatha (in which the 'u' of 'uluka' has been shortened for the sake of metre), 4.92--akhu-kak-olukam yatha, 15.9-yam samkhyayogaih sa-patanjal-akhyaih (in which the 'a' of 'patanjala' has been shortened), 32.4-ganga-narmadavindhyadri- (for '-narmada-'), 65.95-sa-yajnopavitani ca (for '-vitani ca'); 4.13-asmad-ripu-bala-saktim (in which the final 'a' of '-bala-' has been lengthened), 7.82-vasuki-krta-bhusanah (for 'vasuki-'), 25.19-panca-sapta-navasyam va (for '-sapta-'), 27.2 --sita-gandhi ca pujitah (for '-sita-'), 32.37-simha-padmasana-samstha (for '-asana-samstha'), 45.11-santy-abhiseka-yagesu (for '-abhiseka-'); and so on. 267 III. Samdhi (euphonic combination) often made against the rules of Sanskrit grammar: 265 The Buddhist works contain forms like 'astavimsa', 'caturvimsa' and 'unavimsa' mentioned above, but none like 'astasi', 'kata' and 'ulu'.-See Edgerton, op. cit., I, p. 107 (19. 30-31), but pp. 107 (19.35) and 31 (3.118). 266 For instances of similar changes in Buddhist works see Edgerton, op. cit., I, p. 32 (4.3 and 12). 267 For instances of similar changes in Buddhist works see Edgerton, op. cit., I, pp. 23-26 (3.5-46).

THE SAKTA UPAPURANAS (i) Double Samdhi (?), as in 103 ayatavadan (for 'ayatah+avadan') in 39.38-kim ayatavadan surah; danavapi (for 'danavah+api') in 13.65; ekamanadhuna (for 'ekamanah+adhuna') in 29.3- .. tvam... srnusv-aikamanadhuna; ghoranikavatasthire (for 'ghoranikah+avatasthire') in 13.59; graha-nagapare (for 'graha-nagah+apare') in 51.9; jambhakadyatha (for 'jambhakadyah+atha') in 7.69; matarartikah (for 'matarah+artihah', or 'matarah+artihah') in 7.96; tasyagratah (for 'tasyah+agratah') in 22.16; yasyarthe (for 'yasyah+arthe') in 9.57; chinna-pakseva (for 'chinna-paksah+iva') in 85.62- chinna-pakseva parvatah; catur-odbhasya (for 'caturah+udbhasya') in 87.1%; pranmukh-odanmukhah (for 'pranmukhah+udanmukhah') in 12.10; samksepat-oddhrtya (for 'samksepatah+uddhrtya') in 1.53; soktah (for 'sah+uktah') in 11.19%; yamadandogradandah (for 'yamadandah+ugradandah') in 3.19; and so on. 268 As a matter of fact, the above instances are not those of double Samdhi at all, the preceding words having been used without their final Visarga like some of their parallels in Prakrit and Apabhramsa. (ii) Irregular Samdhi, as in mumocanena (for 'mumoca+anena') in 20.35; patantyadhah (for 'patanti+adhah') in 85.55; sarvamangalayamsagah (from 'sarvamangala+amsaga') in 50. i. 29; markanda-rsayavasam (from 'markanda-rsi+avasa') in 13.64; sravanesadhe (for 'sravane+asadhe') in 93.48; vidhvamsayisyeti (for 'vidhvamsayisye+iti') in 84.8; sabdamatraiva (for sabda-matre+eva') in 9.22; tasyonkaram (for 'tasya+omkaram') and yasyomkaradhivasitam (for 'yasya+omkara-') in 10. vii. 6; vedonkara- (for 'veda+ omkara-') in 57.20; sravanarabhya (for 'sravanat+arabhya') in 98.9; brhodarau (from 'brhat+udara') in 50. iii. 39; sarvaInstances similar to many of those cited above occur in the Mahavastu, Lalitavistara, etc.-See Edgerton, op. cit., I, pp. 33 (4.29) and 34 (4.34-35).

104 STUDIES IN THE UPAPURANAS gatanubhuya (for 'sarva-gatan+anubhuya') in 33.25%; vividhanubhuya (for 'vividhan+anubhuya') in 33.22; tvaham (for 'tvam+aham') in 6. ii. 36; devyante (for 'devyam+ante') in 6. ii. 46-mucyate sarva-papebhyo devyante liyate narah; sakti-dvitiyaham (for 'saktidvitiyam+aham') in 84.26-asya sakti-dvitiyaham srjami aparajitam; tarune (for 'tam+arune') in 11.8-bahvrcas tarune dadau; nandikarabhya (for 'nandikam +arabhya') in 23.1-ksirasi nandikarabhya devya bhakti-rato nrpah; devistva (for 'devim+istva') in 60.30; satyedam (for 'satyam+idam') in 39.181; stavedam (for 'stavam+ idam') in 83.11; satyevam (for 'satyam+evam') in 3.4; hiranyakhya andhakaras ca (for 'hiranyakhyah+andhakaras ca') in 3.17; devya apane (for 'devyah+apane') in 7.51; bhayebhya muncati (for 'bhayebhyah+muncati') in 15.4; tatrastha tatrastha citrakarmara- (for 'tatrasthah+ citrakarmara-') in 12.27; purno apraptah (for 'purnah+ apraptah') in 10. iii. 7; vanaro rksah (for 'vanarah+ rksah') in 13.22; so gatasuh (for 'sah+gatasuh') in 14.26; so vai (for 'sah+vai') in 11.18; raho'tha (for 'rahoh+atha') in 26.33; alipanktiva (for 'alipanktih+ iva') in 32.21; and so on.269 Besides the different types of irregular Samdhi cited above there are also many others including those having d (rarely), m or as an intervening Samdhi-consonant (or hiatus-bridger), viz., krtva-d-upagatah (for 'krtvopagatah') in 8.6; comadya-m-antau (for 'adyantau') in 46.60; avilambita-mutthanam (for 'avilambitotthanam' -Bahuvrihi pound) in 12.46; ayuta-m-ekam (for 'ayutaikam') in 95.7; dandasi-pasa-m-udyata (for '-pasodyata') in 85.21; eka-mekam (for 'ekaikam') in 85.39, 90.24; eka-m-ekena (for 'ekaikena') in 31.20; isa-rakso'nila-toya-m-analatma-namaskrte (for '-toyanala-') in 6. ii. 22; kala-garuda-m-adisu (for '-garudadisu') in 6. ii. 31-jyotir-vaidyadi-sastresu *** For instances of Samdhi similar to many of those given above see Mahavastu and other Buddhist works cited by Edgerton, op. cit., I, pp. 34 (4.36), 43 (7.14), and so on.

THE SAKTA UPAPURANAS 105 kala-garuda-m-adisu; ketu-m-ucchrayam (for 'ketucchrayam') in 11.57; koti-m-ekam (for 'kotyekam') in 93.255; krsnayasa-m-alamkarah (for '-yasalam-') in 13.74; mayuram-asanarudha (for 'mayurasana-') in 83.89; -patala-m-akara- (for '-patalakara-') in 84.12; subha-torana-m-akulam (for '-toranakulam') in 12.46; surya-m-ujjvalah (for 'suryojjvalah') in 85.59; tatha-m-ayuta-kotidha (for 'tathayuta-') in 19.5; vayu-m-astram (for 'vayv-astram') in 15.21; vidrumam-anvitaih (for 'vidrumanvitaih') in 93.75; ghanta-kinkini-r-avrte (for '-kinkiny-avrte') in 31.21; puskarini-r-iva (for 'puskariniiva') in 93.79; raudrayasapuri-r-iva (for '-puriva') in 93.91; rju-r-avrana-nisnataih (for 'rjv-avrana-') in 50. i. 51; sambhu-r-ajnaya (for 'sambhv-ajnaya') in 16. 15; and so on.2 270 IV. Hiatus created by absence of Samdhi in Samasa. As in Pali and Prakrit, the adjoining vowels in compounds have been left unchanged, with hiatus, on numerous occasions. See, for instance, rasa-anya-kriya-vada- (6. ii. 31), vasuki-samkhabjaananta-krta-mekhala (7.14); asva-mahisa-marjara-akhu-kakolukam odgirita-ananah (2.58); karnapurita-isave (7.14); (4.92), veddeva-udyanajan (12.5), krmi-jalaka-utthane (12.53), -mahacchadma-uragair iva (7.26), -mahakalpa-utpatti- (6. ii. 28), maha-ulka-nibhaya (7.64), -skanda-uma- (2.13), yuga-manvantara-kalpa-utpatti- (7.43), yuganta-yugautpatti- (6. i. 7); paksa-masa-rtu-dvi-tri-ayanesu (6. ii. 27), sam-ayana- rtu-masa- (26.10), -sattva-rsi-siddha-nisevite (7.93), vahnisa- rsi-mantraih (12.39); vasuki-upavitine (11.33); vasu-asvina-rupaya (26.29); and so on. 271 270 For similar instances in Buddhist works see Edgerton, op. cit., I, pp. 35-37 (4. 59-60, 64). 271 Similar instances of absence of Samdhi in Samasa are to be found in Buddhist works also.-See Edgerton, op. cit.. I, p. 35 (4.51-54).

106 STUDIES IN THE UPAPURANAS V. Wrong formation of feminine bases, viz., (i) anukirtana 272 (from 'anukirtana', a neuter word meaning 'narration', 'statement') in 1.19-cakracaranukirtanam (an Inflectional Tatpurusa compound); kirtana (from the neuter word 'kirtana') in 6. ii. 33 sarva krta-krtyasya kirtana; raja (from the neuter word 'rajas'>'raja') in 50. iv. 52 -puspa-raga-krta raja; sahasra (from the neuter word 'sahasra') in 49.7- trayas ca tri-sahasrasu devah . . . .; vahana (from 'vahana', a neuter word meaning 'a mount') in 7.45-vahana hy ari-darpaha; pota (from 'pota' meaning 'a ship') in 83.112-tvam pota bhava sulini; -karaka (for '-karika') in 16.17-utpatti-sthiti-karake (Vocative); -dayaka (for '-dayika') in 90.11-tah phala-dayakah; • • anicchata (or anicchati according to some Manuscripts, for 'anicchanti') in 9.36-na hi anicchata bala bhunjaniya; dhatumaya (for 'dhatumayi') in 32.19-varksam va sailajam vapi ratna-dhatumayam api; haima (for 'haimi') in 2.54, 22.10; maharaja (for 'maharaji') in 39.16 maharajeti ya devi. ..; mohana (for 'mahani') and bhairava (for 'bhairavi') in 9.49-mohana japyate vidya padamaleti bhairava; mrnmaya (for 'mrnmayi') in 57.6; pauranika (for 'pauraniki') in 39.22-etah pauranika devyah .; raudra (for 'raudri') in 50. i. 15-kali raudra kapali ca; varksa (for 'varksi') in 32.19 cited above in connection with 'dhatumaya'; ... data (for datri') in 31.36-sivayah siva-datayah . . . .; trata (for 'tratri') in 7.38-tratas tah sarva-devanam; (ii) daityantaki (for 'daityantika') in 20.25; -karaki (for '-karika') in 6. ii. 23--bahu-karana-karaki (Vocative), 46.84-utpata-mrtyu-karaki (Vocative); mayatmaka-rupi 272 The final a of 'anukirtana', 'kirtana', 'vahana' and 'pota' is really due to the lengthening of the short vowel a, as often made in Apabhramsa. See Siddha-hema- sabdanusasana 4. 330-sy-adau dirgha-hrasvau // apabhramse namno 'ntya-svarasya dirgha-hrasvau sy-adau prayo bhavatah /-In Apabhramsa the final vowel of a word is lengthened or shortened, in case the case-terminations 'si' etc. follow.

THE SAKTA UPAPURANAS 107 (for '-rupa') in 36.9 --mayatmaka-rupi (Vocative); -mocaki (for '-mocika') in 6. ii. 24-vadha-bandhana-mocaki (Vocative); -pravartaki (for '-pravartika') in 6. ii. 24- -nana-bhava-pravartaki (Vocative), 9.65-sarva-karmapravartaki; virodhaki (for '-virodhika') in 6. ii. 22- -candra-surya-virodhaki (Vocative); sadhaki (for 'sadhika') in 24.26; samkampayanti (for 'samkampamana') in 39.137; sathi (for 'satha') in 8.49; vagesi (for 'vagisa') in 16.18; adanti (for 'adati'-eating) in 85.18; rudanti (for 'rudati') in 9.44; vacanti (derived from the root vac with the suffix 'satr') in 16.41; krpana-panini (for 'krpana-pani' meaning 'a female carrying a sword in her hand') in 14.12; varadabhayapanini in 93.288; varadodyata-panini in 16.40, 50. i. 97, 101, 119, etc.; kartari (for 'kartri'), pradattari (for 'pradatri'), vettari (for 'vettri'), etc. in 6. ii. 21-srsti-samharakartari rudra-murti-prabhavati (Vocative), 83.49 jaye bhuvana-kartari (Vocative), 83.58-sarva-vidyapradattari vidyesvari namo 'stu te, 6. ii. 29-vidyavedana-vettari (Vocative), and so on; ghanta-mudgara-dhari (for '-dharini') in 50. ii. 24; mukt-aksa-sutra-dhari (for '-dharini') in 50. ii. 41; padma-kundala-dhari (for '-dharini') in 50. i. 120; sruva-mekhala-dhari (for '-dharini') in 50. ii. 50; ugrasamcari (for '-samcarini') in 83.66; kama-rudra-samakrni (for 'kama-rudra-samakrti', a Bahuvrihi compound) in 59.11; and so on. 273 VI. Wrong use of Number on many occasions (i) in Dvandva compounds (irrespective of their nature or the number of their component parts), as in astami-navamisu -27.29, 15.6, brahma-suryanam 4.59, devi-brahmesa-suryam-104.9, brahma-janardanaihbrahma-visnoh-15.5, eka-linga-druma-sailaFor some of the above-mentioned feminine forms as occurring in Buddhist works see Edgerton, op. cit., I, pp. 84 (11.1-3) and 81 (10.183-185).

108 STUDIES IN THE UPAPURANAS grha-gostha-trikantake (Locative, singular)-32.7, indr-agni-devasya-5.16, irsy-asuya-krodha-paisunyasahasam-8.36, madya-maithuna-mamsasya-8.41, magha-caitre (Locative, singular)-27.29, nadi-tiradrum-odyana-vivikta-jana-samsadi 1.59, naga- 3.17, ravi-candramah 4.17, raksasa-daiteyah sadhu-linga-brahma-hareh-12.15, siva-visnoh-2.33, vibhav-antahpurasya-8.54, visnu-brahmana-16.13, visnu-brahmasya-7.18, visnu-sakranam-18.1, yamendubhih 15.6, and so on; 274 - (ii) in verbs (irrespective of the Number of their subjects), as in 4.1- • tau bhargavam | |, 4.28 • • • vajra-kalau • ... | prstavan • te |aruroha purim " . yamyam.... //, 5.10-sukham icchan divaukasah (for icchantah divau-'), 5.16-krtva brhaspatir vissnur jagmuh, 6. ii. 18 paraya bhaktya • vayam • | tutosa 11, 9.38-tasya patnisahasrani astav astau bhavet kila //, 9.57-yasy-arthe udaram jagmur madiyam bhargavah pura, 14.28-deva mumoc-opari puspa-vrstim, 15.13-pranasta-rajyadravyani punar eva bhavisyati, 20.24-ghorah jagmuh 39.115- saranam jagmuh rat tada, 55.6-rudhiram sravanti, 82.32 panani - -sura- " piban | asur-anena bhavena devim pujayate sada //, 86.30-nrtyante paramo devah, 92.2-sa ca •• on;275 . | yatra yatra ca pujyante (iii) in adjectives, as in | |, and so 5.16-deva-devi tavantakau, 13.7-vanaspatih samastas ca phala-puspaih susobhitah (for 'vanaspatayah . .'), 26.4-catvari trini dve kuryad angulam kunda-manatah, - ... eka-vimsati kulam ... • 43.41-na 34.4. samkhya vidyate tata ghatamanasya danavan (for • hanyamananam danavanam³), 79. ii. 15-bhogan 174 For similar instances in Buddhist works see Edgerton, op. cit., I, p. 125 (23.1-4). 875 For similar use in Buddhist works see Edgerton, op. cit., I, p. 129 (25.1-6) and p. 38 (5.4).

THE SAKTA UPAPURANAS 109 mahan 276 (for ... mahatah'), and so on; 277 (iv) in pronouns (irrespective of the Number of the nouns for which they are used), as in 7.18-varam vara hare brahman yat te manasi vanchitam; (v) in words denoting cardinal numbers, as in 50. i. 66- vidya-mantrais ca sastibhih (for '... sastya'); and so on. VII. Frequent use of words in Genders other than those of their own (in very much the same way as that in Apabhramsa, in which, as Hema-candra, Trivikramadeva and others say, Gender has no rule -lingam atantram 278): (i) Masculine words used as neuter, viz., • adhivasa in 50. i. 64-adhivasani • " _ • " • .. krtva ...; • anta in 15.10 -na cadir asya na ca madhyam antam ..; atmana in 8.19-atmanam paramam devam tosaniyam sada budhaih; drstanta in 37.91-drstantani bruvanti ca; ghata in 4.18-yuddhe ghatam bhaved vatsa...; grama in 9.7-pura-gramani; homa in 27.28- bhaveddhomam 54.16, 55.13-laksahomam 54.18, 55.13 ... koti-homam 54.20-homam karyam, 57.14, 15-homam; hutasana in 15.7-tvam . . . . hutasanam ...; jiva in 8.51 tac ca jivam matam budhaih; ksaya in 12.24 nemya nase ksayam tatha (in which 'ksayam' is the subject); mantra in 10. vi. 3-yoga-mantram visisyate; pata in 13.30-sakrn-mutr-asru-patani (Tatpurusa compound); putra in 35.29-danam istam ca putram ca devya dipas tathaiva ca; rudra in 54.3-sivam rudram sada vatsa karnikayam nivesitam; sara (arrow) in 43.43-agneyam cintitam saram, 43.45. muktam narayanam saram; sesa in 12.30-sesani; soma in 51.20'-vedah somam kratur yajnah patrany • 276 The Accusative plural (masculine) form of 'maha' derived from 'mahat' by dropping the final consonant 't'. 277 For similar use in Buddhist works see Edgerton, op. cit., I, p. 38 (5.5). 278 See Hema-candra's Siddha-hema-sabdanusasana 4.445, and Trivikramadeva's Prakrta-sabdanusasana III. 4.68. etc. For instances of want of restriction as regards Gender in Mahavastu,Lalitavistara see Edgerton, op. cit., I, pp. 39-41 (6.1-20).

110 STUDIES IN THE UPAPURANAS evam vidur budhah; stava in 26.26-tad-ante ca stavam karyam ...; svapna in 8.35-diva-svapnam (Nominative); svarga in 4.83-tvaya dattam mama svargam...; ucchraya in 11.57-... kathitam ketu-mucchrayam; uddesa in 26.10 -uddesam kimcid atrapi . . .; vara in 1.38-tapas taptva varam lebhe visnuna 39.131 varam varaya yat te hrdi vyavasthitam; vigraha in 3.8-patala-vigraham / . . " prarabdham. " • • • .. . | |; yoga in 10. i. 12-kenopayena tad yogam prapyate • 10. i. 8-yena yogena • sa vidya. . . sidhyati / tac ca deva samakhyahi . //, 10. vi. 4-alpa-viryam hi tad yogam 13.2 kathitam yogam narada-prcchitam; and so on. ... (ii) Neuter words used as masculine, viz., • • • " astra in 4.90-narayanastra-brahmastra-saivas canye 'tha varunah; bhuta in 6.ii.40--na bhuta na ca raksasah; daru in 90.13-darum; indriya in 8.27 sevyaman-endriya brahman pravrddhim upayanti hi (in which the word 'sevyaman-endriyah' is a Karmadharaya compound and subject to the verb 'upayanti'), 8.32- 8.32 -gandha-vag-indriyas tatha; naksatra in 39.92-naksatra bahu-rupas ca; patala in 3. 18-patalan, 3.30-tam jitva tu patalam gatva anyam rasatalam, 4.3-nirjitah sapta-patalah, 4.6 patalah, 8.1-martya-patalah, etc.; puspa in 54.6 gandha-puspas ca datavyah, 69.18-puspams citran, 98.17 puspan; sata in 49.7-trayas ca tri-satas caiva ..; sthana in 68.9, 90.9-ete sthanah • - - .; vasu (wealth) in 2.92 vasum prati; vrata in 1.18-vratas ca niyamadayah; and so on. (iii) Feminine words used as neuter, viz., diva (for 'div') in 4.6-... divam duhkhena sadhyati (passive voice), 4.10-divam dharmena raksitam, 15.7-tvam bhumi-vayu kham jalam huta- sanam diso divam sagara-rksa-cakram; koti in 4.34 laksa-kotini; pataka in 12.42-patakani; pramada in 22.19- pujyeta kanyakah pramadani ca; sobha

THE SAKTA UPAPURANAS 111 in 50.i.61-anekani ca sobhani . . .; and so on. (iv) Difference in Gender between nouns and their adjectives, as in 1.34- 3.2-aham • • ... artha-sampraptih bhavyam ...; praptam ; 7.38 sumahatsu ...; 8.21-brahma • • .. ... apatsu ../ stri-rupadharini bhutva . . . //; 8.28-daha-jvara-mahatapo vahni-pitta-sumudbhavam; 9.9-amatya-sahito vagmi candrabuddhir amantrayat (in which Candrabuddhi is the name of a female); 9.29-evam tasya matir bhuta narada-pathaga ...; 12.42-tatah sakadaliksu-dandan patakani samucchrayet; 16.18 - • samhara-karaya rudra-deha-bhavaya ... (qualifying 'devi' in verse 17); 16.36-jayam ca samare nityam vidya-labham ca durlabham / dirgham ayur athotsaham parthivanam ca istada //; 17.11-yam- antakah ghatitam devya. .; 20.18 samdhyayam samupasthite; 27.29-maha asvina-mase tu astami-navamisu ca (in which 'maha' qualifies 'astami-navamisu'); 31.29- ... stribhir mangalavadibhih; 33.12-nilotpala-patra-gandha-surabhim pitva. . . madhu; 33.25-suracitaih sragbhih; 40.3 mahati pujayam; 42.11-devyah (for 'devyah') sarvartha-datarah . . . ...; 43.50-tada bhayam mahan asit ; 43.58-yenaham smaritam (for 'smrtam') tvaya (said by Pasupata Sara to Gajanana); 50.iv.84 tabhyam homam sukhavaha; 51.24-sankha-nilotpalakaran patrani parikalpayet; 55.16-phalavati apamargam; 58.2-tustam umapatih; 82.74-evamvidhaih sada stribhir nityam smara-nipidibhih; and .. so on. (v) Difference in Gender between nouns and their pronouns, as in 10.vi.9-sabda-brahma param brahma tasmin ksine yad aksaram / sada tam (for 'tat') manasa dhyayet .. //; 16.12-alpa-dosa visantimah (for ... visantime'); 26.20-chede bhayam vijaniyat tad-artham tan na (for 'tam na') karayet

112 STUDIES IN THE UPAPURANAS ... (the word 'tat' having been used for 'cheda'); 33.21- tam bhittva daity-angan-antahpuram (for 'tad bhittva ...'); 34.14-tasya jalam... tam (for 'tat') avalambayet; 43.40-pasodyata-karam ghoram amayopari pata sah (i.e. sarah) / ratha-nagasvapadatam hanyamanam (for 'ghnan') sahasradha //; 65.40-ratrau tasmin ...; and so on. (vi) Adverbs used in the Masculine Gender, as in 3.17 ksubhito mahan (for 'mahat'). VIII. Wrong declension, as in (i) -mogha-viryayah (for '-mogha-viryah') in 84.17, su-hrstatmah (for 'su-hrstatmanah') in 72.5, surarinam (for 'surarim') in 5.12, pathim (for 'panthanam') in 95.9, urdhva-retam (for '-retasam') in 63.9, purodhan (for 'purodhasah') in 67.63, visvakarmena (for '-karmana') in 93.201, murdhana' (for 'murdhna') in 91.38, balagbhih (for 'balakaih') in 98.27, vettaraih (for 'vettrbhih') in 32.42, hotaya (for 'hotre') in 11.33, -kartaya (for '-kartre') in 26.34, -vidhatave (for 'vidhatre') in 7.9, vighnahantaya (for '-hantre') in 83.34, maharetaya (fo '-retase') in 100.21, antyajam (for 'antyajanam') in 88.4, -mahoragam (for '-mahoraganam³) in 98.10, -go-vrsam (for '-govrsanam') in 98.11, matr-ganam (for '-gananam') in 90.20, rajanam (for 'rajnam') in 55.11, 56.1, pratyusi (for 'pratyuse' or 'pratyusasi') in 90.20, himavante (for 'himavati') in 93.10-himavante mahagirau, maha-kapala-malaya (masculine, Vocative) in 11.32, and so on; • saptah or sapta (for 'sapta') in 96.4-panca-saptas tatha dvijah, 25.18- pancatha sapta va 25.19 panca-sapta-navasyam va, astam (for 'asta') in 82.59-upapatalam astam tu . . ., anyah (for 'anye') in 12.43-anyas ca vividha bhogah.. imaih (for 'ebhih') in 65.96-dravyair imaih, 93.188-imair mantraih; . . "

THE SAKTA UPAPURANAS 113 (ii) patakani (for 'patakah') in 12.42, patakan (for 'patakah') in 65.94, pramadani (for 'pramadah³) in 22.19, sobhani (for 'sobhah') in 50.i.61, laksa-kotini (for 'laksa-kotyah' or 'laksa-kotayah') in 4.34, mahakotyah (for 'mahakotyah' or 'mahakotayah') in 17.11, vidyutani (for 'vidyutah') in 87.9, mrdani (for 'mrdah') in 50.iv.90-kastha-saila-mrdani ca, srajani (for 'srajah') in 61.7, matrm (for 'mataram') in 91.6, gadgada (for 'gadgadaya') in 36.29, -jyotsnena (for '-jyotsnaya') in 32.25-mayukha-danta-jyotsnena, mekhalaih (for 'mekhalabhih') in 26.3, vidyaih (for 'vidyabhih') in 86.4, bhaktina (for 'bhaktya') in 22.23, 39.69, 89.9, buddhina (for 'buddhya') in 27.25, sva-saktina (for 'sva-saktya') in 90.21, -panktibhih (for '-panktibhih') in 94.6, -prakridaya (for '-prakridayai') in 8.20, yuga-pidat (for '-pidayah') in 50.iv.2, nandaya (for 'nandayah'-Genitive case) in 93.273, navamim (for 'navamyam') in 76.40, pratimau (for 'pratimayam') in 79.iii.4, trtiyam (for 'trtiyayam') and dvadasim (for 'dvadasyam') in 99.12, samharakaraya Vocative) in 16.18, rudra-deha-bhavaya (feminine, asya (for 'asyah'-Genitive case) in 84.26, tisranam (for 'tisrrnam') in 80.24; (iii) puspan (for 'puspani") in 98.17, darum (for 'daru' -Accusative case) in 90.13, dadhina (for 'dadhna') in 33.64, danasu (for 'danesu') in 91.17, -kasthasu (for '-kasthesu') in 6.ii.26, -madhyasu (for '-madhyesu') in 6.ii.3 -sarit-sagara-madhyasu, anyam (for 'anyat') in 3.30; and so on. For wrong declension of many other words ending in vowels or consonants see above (in the sections on changes made in the stems and on Genders). [Similar instances of wrong declension are to be found 8

114 STUDIES IN THE UPAPURANAS in the Buddhist works also.-See Edgerton, op. cit., I, pp. 48 ff.]. IX. Incorrect use of case-endings (vibhakti): (i) The first case-ending used (a) for the second, as in 1.3-kathyatam yatra ghoradya bhutah samprata-danavah / bhavisyas ca vinasisye devi deva-namaskrta // (for ... ghoradyan bhutan samprata-danavan / bhavisyams ca ... · · Il'), 2.75-kosa-vrddhih praja-raksa ... (for '-vrddhim praja-raksam ..'), 7.69- te te khadaya ... ... • (for tan tan khadaya ...'), 7.72-kala-pida kriya-pida papa-pidatha dhatuja vata-pitta-kaphodbhutam samaye bhairavah sada // (for 'kalapidam kriya-pidam papa-pidam atha dhatujam / . samayed bhairavah ... //'), and so on; (b) for the third (in the anukta-karta in the passive voice), as in 1.38-mantradyah sadhita yatra nrpa naga rasatale (for '... nrpair nagaih ...'), 7.27 ardhendur iva lalata nimn-onnata-virajitaih (for 'ardhendubhir iva lalatair nimn-onnata-'), 7.76-ye canye vairinah krtah (for . vairibhih krtah'), and so on; 6 • (c) for the fourth, as in 26.35-... tvam pitre . . . (for ... tubhyam pitre . . .'); .. (d) for the sixth, as in 4.92-na yuddham sreyase deva uraga nakulaih saha (for '. uraganam ...'), 7.37-santida vita-raganam ta devya rsi-manavah (for . -manavanam')-those goddesses are the givers of peace to those sages and men who are free from desire; (e) with the separated component parts of a compound, as in 6.ii.25-brahma-prajesa-somas ca yaksaraksah-pisacaki (for ... -soma-yaksa- ...'), 13.75- . . . pumbhih pasa-dandodyatair mahan (for 'pumbhih mahapasa-dandodyataih'). (ii) The second case-ending used (a) for the first, as in 3.5-kusa-dvipam pura tena sa-balenaiva arjitam / jambum sakam tatha krauncam

THE SAKTA UPAPURANAS 115 salmalim atha pauskaram / sapta dvipas tatas tena deva-raja vasikrtah //, 3.27-paundarikam ca duspreksyam sveta-bhadram tathoragah (for 'paundarikas ca duspreksyah sveta-bhadras tathoragah'); (b) for the third, as in 33.72-daksinam priyatam siva (for 'daksinaya. . .'), 33.84-daksinam priyatam jaya (for 'daksinaya . . .'); (c) for the fourth, as in 58.15-varam dattam dvijottamam (for '... dvijottamaya'), 69.21-brahma- nan bhojanam dadyat (for 'brahmanebhyah . . .'), 93.191-mahadevim namas krtva (for 'mahadevyai ..'), 94.25-guhyesvaram namas krtva (for 'guhyesvaraya . . .'); .. - (d) for the sixth (also in the anukta-karma of Krd-anta verbs in the active and the neuter voice), as in 1.57- sravanat ... idam puranam siva-bhasitam (for asya puranasya siva-bhasitasya'), 7.28-nilotpaladala-prakhyair harinir iva locanaih (for ... hari- ninam iva ...'), 8.22-sarve ca sukham arthinah (for '. . . sukhasya arthinah'), 21.1-devim upasakah (for 'devyah ...'),43.41-na samkhya vidyate tata ghatamanasya danavan (for '... hanyamananam danavanam³); • (e) for the seventh (in the Adhikarana-karaka), as in 4.69- satrur asmakam vadham udyatah (for ... vadhe udyatah'), 9.53-yadi mam varado devah . . . (for ... mayi varado ...'), 66.6 devatah sarva vestayanti caturdisam (for ... catur- dai diksu'), 93.67-pati-hinam na rajate (for 'pati-hine .'=pati-hinayam. .....); • (f) in connection with the words 'pravrttah', etc., as in 14.26-gadam samadaya jayam pravrttah (iii) The third case-ending used (a) for the first, as in 1.32-agastya gis tu nrpagaih (=nrpaga bhutva) loke khyatim gamisyati; (b) for the seventh (in the Adhikarana-karaka), as in 6.ii.35-varada ca abhud ubhau (for ... abhud ubhayoh') granted boons to both (of them), 8.38-

116 STUDIES IN THE UPAPURANAS visayaih samnivesyatam (for 'visayesu sam-'), 93.106--ebhih sthanair mahadevi pujita varada bhavet (for 'esu sthanesu . . .'); (c) in Avyayibhava and Bahuvrihi compounds having 'yatha' and 'yavat', and 'saha' (contracted to 'sa'), respectively as their first members. See, for instance, 2.49, 65.66, 65.70, 68.7, 95.35, etc. yathepsaya (for ... yathepsam'), 8.55, 32.43, 32.44, yathasaktya... (for '... yathasakti . . .'), 100.14 yavac-chaktya (for 'yavac-chakti'), 3.12 -samnahya sabalenaiva mahasamgrama cakrire (for . . sabalam eva ...'), and so on. (iv) The fourth case-ending used etc.- " • (a) for the second, as in 39.78 -indraya saranam gatah (for 'indram ..'), 83.16 veda-vedanta-garbhaya namami ... (for '-garbham namami ...'); (b) for the sixth, as in 3.22 -kala-vajrasya hetave (for ... hetoh'), 52.11-snanam sisyaya kartavyam (for 'snapanam sisyasya kartavyam . . .'); ... (c) for the seventh, as in 39.112-indray-abhimukho 'dhavat (for 'indrasy-abhi-'); tave, jaya (d) in the Vocative case, as in 7.7 ff. (jaya -karana-traya-hetave, jaya... -buddhindriya-vidha- -pradhana-purusatmane, jaya suline, jaya kala-maha-kuta-visa-kanthastha-jirnave, jaya sambhave, jaya smasana-vasine, jaya sardra-gajacarma-pravrtaya mahatmane, jaya trisula-hastaya karnapurita-isave, and so on), 7.68-chinda chinda mahacakra isu-hastaya samkara, 7.70-gada-trisulahastaya sarvam badham vinasaya, 16.18-jaya samharakaraya rudra-deha-bhavaya ca; and so on. In the compound 'kala-maha-kuta-visa-kanthasthajirnave' ( = kanthastha-kalakuta-mahavisa-jirnave) in Devip. 7.12, the part 'jirnave' has clearly been formed by adding the fourth case-ending (singular) to 'jirnu', a peculiar word having its parallel neither in the Devi-purana nor in any of the Buddhist works written in hybrid Sanskrit. The meaning of the word 'jirna' (which points to the thing digested and

THE SAKTA UPAPURANAS 117 not to the person who digests a thing) and the fourth caseending of 'jirnave' appear to indicate that the final u of 'jirnu' was not due to the influence of Apabhramsa, in which, as Hema-candra and others say 279, words ending in a have u before Nominative and Accusative singular terminations. So, one may feel tempted to assume that this word (jirnu) was derived from the word 'jirna' after turning it into a Nominal verb (Nama-dhatu) by the addition of nic and then adding the Krt suffix u (as in 'bhiksu', 'cikirsu' etc.); and thus, the said compound would come to mean 'one (i.e. Siva) who digested the great poison Kalakuta remaining in his throat'. But it is more probable that 'kala-mahakuta-visa-kanthastha-jirnave' is a Bahuvrihi compound formed in imitation of those of the 'ahitagni' class 280 and that its final u (before adding the fourth case-ending) was due to the influence of Apabhramsa. (v) The fifth case-ending used (a) for the third, as in 14.21-krpana-ghatad api tam jaghana (for '..... -ghatena ...'), 14.24-sarebhyas ciccheda (for 'sarais ciccheda'). (vi) The sixth case-ending used " • (a) for the second (in Karma-karaka), as in 3.2 aham. praptam tava janardanat (for '... praptas tvam janardanat'), 9.68 -ekaikasya padasyasta-sahasram japet... tilanam madhunaktanam asta-sahasram juhuyat (for ''ekaikam padam asta- tilan madhunaktan asta- ...'), 25.9 -vipranam koti-kotinam bhojayitva (for 'vipranam koti-kotir bhojayitva'), 28.1-indrac ca mama agatam (for . . . mam, or mayi, agatam'), 39.108-panya-striva yatha lobhat kamukanam varayate (for . kamukan variyati'), 93.15 -parvatyas ca prasamsante (for 'parvatim ca . . .'); . . 27 See, for instance, Hema-candra's Siddha-hema-sabdanusasana 4.331-syamor asy-ot. The final a of a word is replaced by u in case 'si' and 'am' (i.e., the Nominative and Accusative singular terminations) follow. 280 See Panini's Astadhyayi ii.2.37-vahitagnyadisu, according to which the word 'ahita' (having the kta-pratyaya) may be placed before or after the word 'agni' in a Bahuvrihi compound. Similarly, both 'pita-taila' and 'taila-pita' are corrent,

118 STUDIES IN THE UPAPURANAS (b) for the third, as in 4.72-mamayam vinipatitah (for 'mayayam vini-'), 86.30-nrtyante asmakam saha ... (for '... asmabhih saha'); * ... (c) for the fourth very often in Sampradana-karaka and rarely also on other occasions (as in connection with the word 'namah'), e.g. in 1.14-yatra samkirtayed brahma manv-adinam praprcchatam (for ' manvadibhyah praprcchadbhyah'), 1.36-utpattikirtanam srsteh prathamam samudahrtam / . . . rsi- nam pariprcchatam // (for '... rsibhyah pariprcchadbhyah'), 2.20... vidya visnor datta . . . / pitamahasya tenapi... // (for '. . . visnave . . . / pitamahaya //'), 2.32 tasya vara-danam (for ... tasmai vara- ..'), 2.41-pradadau tasya daityasya yathepsita-varam nrpa (for '. . . tasmai daityaya ...'), 2.50 evam dattva varam tasya .. • : • (for ... tasmai ..'), 2.100-evam tasya varam dattva .... (for tasmai ...'), 40.15-ugrasenasya samkruddha (for 'ugrasenaya . . .'), and so on, 26.33 (namah) brhaspataye sukraya sane raho 'tha 281 ketave (for ... sanaye rahave 'tha...'); (d) for the fifth (in Apadana-karaka), as in 3.4- srnusva gadato mama (for '. gadato mat³), 3.18- bhitas tesam (for 'bhitas tebhyah'), 93.143-prasadasyottare (for 'prasadad uttare'); (e) for the seventh (in Adhikarana-karaka), as in 1.32-tatha manv-atri-bhrgubhir asmakam avatarita (for ... asmasu. . .'), 2.39-tustas tasya khagasanah (for '... tasmin . . .'), 2.40-param astanga devasya bhaktim ekam tu yacate (for '. . . deve ...'), 2.49- artasya me sudinasya dayam tvam kuru kesava (for 'arte mayi sudine '), 7.18- ... tada devo visnubrahmasya tustavan (for '... visnu-brahmanoh . .'), 11.51-tutosa kesavas tesam. (for ... tesu ...'), 33.10-yesam harah prita-mana babhuva (for 'yesu • • . .. ... 281 We have already seen that 'raho 'tha' is the result of Samdhi between 'rahoh' and 'atha'.

THE SAKTA UPAPURANAS 119 harah ...'), 39.47- ... yacanam kasya karomi ... kasmin ...'). (for ... kasmin (vii) The seventh case-ending used (a) for the first (in the secondary object in the passive voice), as in 35.2- . . . sakrena brahmani prstah purvam (for brahma . . . '. . . • ..'); ... (b) for the second (and also for the third or fifth in some cases), as in 83.16-... namami tvayi (for tvam ...'), 51.21-vina patre na sidhyati (for 'vina patram, or patrena or patrat, ."); • (c) for the third (in Karana-karaka), as in 14.4 mahamohe 'timohite (for 'mahamohenatimohite'), 51.6 avidhau yah sivam pujyeta. (for "avidhina yah . . .'); * • (d) for the fourth in connection with the word 'namah' and in Sampradana-karaka, as in 6.i.2-- bhuta-bhavya-bhavisyanam karanakarane namah (for '. . . karanakaranaya ...'), 7.58-namah ... kotar- aksaya bhairave (for '... bhairavaya'), 7.59-namas te. . . surasura-bhayamkare (for '... -bhayamkaraya'), 26.27-... namas te.../ pita-vasaya pavane // (for pavanaya'), 26.28- ... brahmane dahane namah / . . . sarvasine sarva-gate. sarvasine sarva-gate . . . namo namah // (for dahanaya .../... sarva-gataya ∙//'), 26.34-sarvage graha-rupaya (namah) (for 'sarvagaya ...'), and so on, < • .. .. • ... • ... 9.60-tvayapi vatsale deya abhakte najitendriye (for ... vatsalaya abhaktaya najitendriyaya'), 11.7-11- sa datta . . . sarasvate / .. trivrsena bharadvaje / antariksena bahvrce / tasy- arunena balaje tenapi ca krtamjaye / krtamjayena rnaje ... vajasravas tatha some // (for sarasvataya / . bharadvajaya ... / bahvrcaya balajaya... krtamjayaya .../... rnajaya • • • | ../ somaya ...//'), and so on; " . .. (e) for the fifth, as in 2.78-... naparadho mamopari / kartavyo mama vakyesu... // (for ... vakyebhyah...'), 7.78-79- simha-khadgi-bhayesu ca/ ...

120 STUDIES IN THE UPAPURANAS traya mam taskaresu // (for ... bhaye- ... • • bhyah .../ taskarebhyah ||'); • ... (f) for the sixth, as in 33.32 vidyadharisu nivasanti kucantaresu (for ... vidyadharinam ni-'); (g) in Sambodhana, as in 7.7-jaya .../ saksi-bhuta-guna-traye // (for '... -guna-traya'); (h) in certain Avyayibhava compounds (such as 'prati-sthanam' etc.), as in 39.146-prati-sthane sthitah. (viii) Different case-endings used in nouns and their adjectives. See, for instance, 95.8-trayo laksan. (ix) For other instances of wrong use of case-endings see 4. 35 b-37 a, 7. 27 ff., 9. 29-32, and so on; also 6.ii.1 purve (for 'purvam' meaning 'formerly') 282, 6.ii.28-manavan deva-satrunam brahmady-asmakajantusu, 7.55-bhuta-raksasa-vetalady-arinam samka- tesu ca, 16.36-jayam ca samare nityam vidya-labham ca durlabham / dirgham ayur ath-otsaham parthivanam ca istada // (in which the root 'da' in 'istada' has been taken as governing the Accusatives in 'jayam', 'vidya-labham' etc.), 93.207-kundalaih kata-keyurair mukutadi-vibhusita, 94.7-nana-sayya-samakirnair nana-camara-sobhitaih | nana-vastra-vitanais ca nanavimana-samkula //, 20.23-tasya samcaramanasya kampate ca vasumdhara -and the earth quakes as he moves, 27.2-megha-dundubhi-sankhanam venu-vinasvanah subhah (for '-sankha-venu-'), vrsendra-nrpakakanam kokila-svana pujitah (for -'kaka-kokilasvanah pujitah'), 27.3-simha-barhina-syenanam camarakrtir istidah (for '-syena-camara-'), and so on. It is to be noted that instances of wrong use of caseendings are innumerable, and on many occasions words have been formed and case-endings added to agree with the sounds of the final words of the neighbouring compounds. For such word-formation and use of case-endings see Devi-purana 6.ii. 19 ff., 7.7 ff., 7.58 ff., and so on. *** For similar use of 'purve' see Lalitavistara 196.5 (=P.L. Vaidya's ed. 142.17).

THE SAKTA UPAPURANAS 121 [For instances of indifferent use of cases in Buddhist works see Edgerton, op. cit., I, pp. 42 ff.] X. Use of words (in the Nominative or the Objective case) without the relevant case-endings, as in • • .. " • " 2.40-param astanga (for 'astangim') devasya bhaktim ekam tu yacate, 2.92-mantri-vaky-anal-otthena udyotita vasum prati / matir danava-nathasya vrat- aradhanam ayayau // (for . . . udyotita . . . / mati-'), 3.12, 21-mahasamgrama cakrire (for '-samgramam .'), 4.12- yusmat (for 'yusmakam') suraih sardham yuddham. 7.73 carmena vamam bhuja (for 'bhujam') purayitva jaya-mukho dhavati kruddha (for 'kruddhah') kopat, 14.25-tathapi kalo gada (for 'gadam") tam mumoca, 14.26-kalam hatam bhairava (for 'bhairavah') samniriksya, 14.27-evam sa kalo hata (for 'hatah') bhairavas ca... ... devya (for 'devyam' =devim) samasadya jvalanta-kopah, 15.22-jighamsa vajram bala-vajra-virya (for '-viryam') sarena haimadala-patritena, 17.14-rakso 'nalam vayu (for 'vayum') tatha kuveram 17.31-phalam vapi suksma (for 'suksmam') tyajitva vrajante, 27.2- kokila-svana pujitah (for '-svanah pu-'), 32.34-35- sara-cap-aparau tabhyam rsa-mudgara caparau/ parasu-cakra-dharau canyau damaru-darpana caparau // sakti-kunta-dhrtau canyau hala-musala caparau pasa-tomara canyau tu dhakka-panava caparau //, 37.72-sravana syandanasyarthe (for 'sravanam. . .'), 37.73-aveti raksane dhatu adhiprakatane tatha (for .. dhatuh ...'), 50.ii.44-sobhakrd rati kartavya. . . (for '. . . ratih . . .'), and so on. 283 In putting down the characteristic features of Apabhramsa in his Siddha-hema-sabdanusasana 4.329 ff. Hemacandra gives two rules, viz., 'sy-am-jas-sasam luk' (4.344) and 'sasthyah' (4.345), which say that in Apabhramsa the Nominative, Accusative and Genitive terminations, both singular and plural, are often dropped. (For similar rules • 283 Instances of similar use of words without case-endings are to be found in the Buddhist works also.

122 STUDIES IN THE UPAPURANAS see also Trivikramadeva's Prakrta-sabdanusasana III. 4.16-17). XI. Krd-anta nouns governing cases. For instances see above (pp. 87,115). Also 43.48-brahm- astram prativarane-for preventing the Brahmamissile, 53.9-10-mataranam sada ، pujanam . // (for ... cakrasya .. cakram ... I . .'), 84.6- sambhum ghataya (for 'sambhoh...'), 91.14-sastrani pathanad danat. . . (for 'sastranam. . .'), and so on. 284 XII. Use of cases (Karaka) without verbs: (i) Nominatives without verbs, as in 4.35-durniriksyam ripur ghoras tasya drstva tu vahinim, 6. ii. 36-tada tvaham (=tvam+aham) ca samcintya sahaya bhava suvrate, 13.15-ta drsta munayah ksobham kim punar asuradhipah, and so on. (ii) Accusatives without verbs, as in 2.105-nirjitya sarva-nrpatims tatha dvipesu codyamam, 10.i.2-sanatkumaram varadam tapasa dhutakalmasam/mama putram mahaprajnam siva-bhavena bhavitam//, 12.23-vrtta va bhangam adhatte rajnah putram purohitan, 12.51-chatra-pato nrpam hanyat pataka mahisi-vadham, 41.9-evam tam krsna-dharma- nam mahabala-parakramam/ samgare nihatam vatsa brahmendra-pariraksitam// (iii) Datives without verbs, as in 12.58-59-evam purvam harih ketum praptavan vrsa-vahanat tatha brahmasya tenaiva brahmanah sakram agatam/ XIII. Irregularities in the formation of compounds (Samasa): (i) Wrong position of the component parts, as in 2.1-nrpavahana-mahatma, 39.115-brhaspati-mahamatim (=mahamati-brhaspatim), 2.16-visa-bhaujamga-(for 'bhaujamga-visa-'), 7.12-kala-maha-kuta- 'kanthastha-kalakutavisa-kanthastha-jirnave (for 284 Instances similar to these and most of the others noted in sections XII-XXI are to be found in the Buddhist works also, but it is no longer necessary to point out these parallelisms, except in special cases.

THE SAKTA UPAPURANAS 7.85-nadi-samgama-punye 123 mahavisa-jirnave'), 7.26-veni-bandha-mahacchanda 285 uragair iva prsthagaih (for 'mahacchanda-venibandhair uragair ...'-with highly graceful braids of hair moving like snakes on the back), 7.27-ardhendur iva lalata nimn-onnata-virajitaih (for 'ardhenubhir iva nimn-onnata-virajita-lalataih'-with foreheads having depressions and elevations like the half-moon), 7.32-guru-vistirna - nitamba - mams - opacita-sobhitaih (for 'upacita-mamsa-sobhita-guru-vistirna-nitambaih'), (for 'punya-nadi-samgame'), 7.88-patra-bhurjesu (for 'bhurja-patresu'), 8.15- raja bhumyam janu-gata-sirah (for 'raja bhumi-gata-janu-sirah'), 12.56-nrpa-saha rastram (for 'saha-nrpam, or sa-nrpam, rastram'), 13.15-sasankasampurna vilanchana ivananah (for 'vilanchana-sampurna-sasank-ananah'-having faces like the stainless full-moon), 15.16-laksyate bana-dharotthair vajradanda-mahasvanaih (for 'laksyate vajradanda-banadharotthair mahasvanaih'), 16.8-yasya kari-mahagandha mada-matta na rastrajah (for 'yasya mahagandhakarino mada-matta na rastrajah'), 66.33-sarit-sarahsakhatena (for 'sakhata-sarit-sarobhyam³), 93.82-jatiphala-sadadimaih (for 'sadadima-jati-phalaih'), and so on. (ii) Unlawful separation of the component parts (sometimes adding no case-endings to those which are not final), as in ... 7.4- anghrim ca yugasyadhah (for '... anghriyugasyadhah'), 7.78-nakra-vyaghra-varahesu simhakadgi-bhayesu ca (for '-varaha-simha-'), 8.15 (cited above), 12.14-sukl-ambara-dharam caiva samudrataranam nadi (for ... nadi-samudra-taranam'), 13.15 (cited above), 13.75- tatha pumbhih pasa-dandodyatair mahan (for ... tatha pumbhir maha-pasadand-odyataih'), 26.6-catvari katak-opetam (for 285 For '-mahacchanda-' the printed edition of the Devi-purana reads '-mahacchadma-', which, though occurring in many Manuscripts, seems to be a mistake for '-mahacchanda-'meaning 'highly graceful.'

124 STUDIES IN THE UPAPURANAS tu 'catus-katak-opetam'), 27.29-maha asvina-mase astami-navamisu ca (in which the word 'mahat' qualifies the words 'astami' and 'navami'), 53.3-subhodayam jayam bhagyam kalyanam aparajitam / mangalam asta-siddhim ca vibhavam subhadam subham // (for 'subhodaya-jaya-bhagya-kalyan-aparajita - mangal-astasiddhi-vibhava-subhadam subham'), 92.5-kalau ghore maha prapte (for 'maha-ghore kalau prapte'), 2.40-param astamga devasya bhaktim ekam tu yacate, and so on. (iii) Occasional insertion of m or r (or ra) either as a Samdhi-consonant or as a connecting link between the component parts. For instances see under Samdhi above. Also 7.89 sarva-tirtha-tapo-dana-m-sarva-vratapradayika (for '-dana-sarva-'), 13.57-ingud-odumbaram-kharjura-matulungaih (for '-odumbara-kharjura-'), 16.7-vyuha-kriya-bhava-mandala-m-tattva-vedita (for '-mandala-tattva-'), 33.5-tam buddhva tad-gata-mcittam (for '-tadgata-cittam'), 84.12-prabha-m-patala- (for 'prabha-patala-'); 85.38 cala-drsti-ra-kataram kataram'); and so on. (for 'cala-drsti- (iv) Occasional insertion, between the component parts, of any of the indeclinables 'atha', 'hi', 'ca', 'tatha' 286 etc., as in 6.ii.26-lava-syanda-truti-mesa-muhurta-atha-kastha- 13.22-trna-kesa-mukta-hy-asthini su, 13.43 darsanam, -makarakula-c-odaka (narmada), 79.iv.4 ayuta-traya-c-ottama (for '-trayottama'), 17.18- -caturthi-tath-aikadasi-krsnapaks-otsave (for 'caturthyaikadasi-'), and so on. It is to be noted that in 'anghri-m-ca-yugasyadhah' (for 'anghri-yugasyadhah' in 7.4) there are both the intervening Samdhi-consonant 'm' and the indeclinable 'ca'. 286 For an instance of insertion of 'tatha' between the component parts of a compound in a Buddhist work see Sadhana-mala, No. 171 (p. 343)-puspa-dhupa-tatha-dipagandha-naivedya-samcayaih.

THE SAKTA UPAPURANAS (v) Wrong use of the Samasantas, as in 4.51-kaustubhoraska-manditah, oraska-bhusanam, etc. 125 11.50-kaustubh- (vi) Various other irregularities, as in 2.58, 50.ii.2 vedodgirita-ananah, 4.33-yamabhangaya mahisasya (for 'yamasya mahisasya bhangaya' -for the defeat of Yama's buffalo), 6.ii.28-brahmadyasmaka-jantusu, 6.ii.34-evam-bhutartha-bhavyaih, 7.27-ardhendur iva lalata-nimnonnata-virajitaih (cited above), 7.40-anaupama (meaning 'incomparable'), 7.42-svami-bhuta, 7.55-atmanah para-raksasu-in protection of (one's own) self from others (i.e. enemies), 7.76-isuj-opala-varksah-(injuries) created by arrows, stone and trees, 83.117-mahad-apada (for 'mahapat'), 8.34 mahad-aisvaryat (for 'mahaisvaryat'), 54.7 ksayadi-mahadapadah (for '-mahapadah'), 97.8- ... brahma-visnu-mahesvarah striyah pana-mano- ... • have • 'nugah- ..... Brahma, Visnu and Mahesvara their minds running after women and drink, 17.12- ghorah yuddha-mano'nugah- Ghora had his mind bent on war, 7.61-pinaka-vara-panine (for '-panaye'-Dative), 14.12 krpana-panini-(a female) with a sword in (her) hand, 16.40, 50.i.97, 101, 119, etc.-varad-odyata-panini-(Devi) who has raised her hand for granting a boon, 93.288-varad-abhayapaninim, 17.23-bhutani-tari (feminine, Vocative)-(a female) who delivers beings, 25.14, 19-pakvesta 287- saila-darvam-made of burnt bricks, stone and wood, 25.18-vimsaikam (for 'eka-vimsatim'), 26.10-sam- ayana-rtu-masa-paks-aho-ratra-purvavat, 32.33-vims- asta-twenty-eight, 32.35-sakti-kunta-dhrtau-holding (the missiles) Sakti and Kunta, 33.6-vibhisana-karam -terrifying, 33.70-darbha-mulabhih (mrdbhih) (with mud) that sticks to the roots of Kusa grass, 35.8 dhvaja-palana-kartavya-protection of the banner should 287 For 'pakvesta-' Devi-purana 25.14 and 19 wrongly read 'pastanka-' and 'pastanga-' respectively. For the reading 'pakvesta-' see Devi-purana 118.2-pakvesta-daru-failam..........

126 STUDIES IN THE UPAPURANAS be made, 37.25-matarany-agraja-(a female deity) born before the mothers, 39.45-sa-vistari, 39.157 vinahava-nipatane, 43.14-vana-kananam (Bahuvrihi compound), 48.3-pasc-astamita-bhaskare, 50.i.27 tavanta-bhedena, 50.i.32, 124-sarvabharana-bhusangi (feminine), 50.iii.27, 85.20-bhayanana, 82.56-trayaraksasah, 83.35-raksa-palakah, 82.47-khadga-kar- 85.21-dandasi-pasa-m-udyata, 86.27-sa- (adverb), 93.63 sosnena, 93.285-vinagrhita-hastam (Bahuvrihi compound), 91.38-evamvidha-krtena, and so on. odyatah, ksipram XIV. Wrong formation and use of verbs: (i) Parasmaipadiya roots used in their Atmanepadiya forms, viz., (a) bhvadi-ganiya bhu in 8.53, 9.20, 23.17, 19, 24.7, 26.43, etc.--bhavate (for 'bhavati'), 86.27-bhavadhvam (for 'bhavata'), 33.21-bhaveta (for 'bhavet'), car in 47.16 carate (for 'carati'), drs in 7.5-pasyate (for 'pasyati'), 13.72-pasyeta (for 'pasyet'), 34.13-apasyata (for 'apasyat'), 1.70- dadrse (for 'dadarsa'), gam in 33.62, 34.5, 42.15, etc.-gacchate (for 'gacchati'), 86.16-gacchase (for 'gacchasi'), 84.4-gacchadhvam (for 'gacchata'), 95.9-gaccheta (for 'gacchet'), jap in 1.40-japate (for 'japati'), ji in 4.5-jayamahe (for 'jayamah'), jval in 8.30-jvalate (for 'jvalati'), krs krs (ava+) in 49.2-avakarsate (for 'avakarsati'), ksar in 49.12, 17, 19-ksarate (for 'ksarati'), pa (to drink) in 48.11-pibate (for 'pibati'), pat in 27.10-patate (for 'patati'), path in 7.84-pathate (for 'pathati"), raks in 7.80, 14.8-raksate (for 'raksati'), raks (pari+) in 41.2-pariraksase (for 'pariraksasi"), sad (pra+) in 1.26-prasidate (for 'prasidati'), smr in 91.52-smarate (for 'smarati'),

THE SAKTA UPAPURANAS stha (to remain) in 2.14, 3.34, 7.90, 8.12, etc. tisthate (for 'tisthati'), - 127 stha (ut+, meaning 'to rise') in 55.6-ketavas cottisthante (for ... cottisthanti'), tap in 13.25-tapate bhanuh (for 'tapati ...'), tyaj in 13.19-tyajate (for 'tyajati'), vad (to say) in 4.84, 6.ii.12, 14.6, etc.-vadate (for 'vadati'), 37.41-vadante (for 'vadanti'), vas (to reside) in 9.11-vasante (for 'vasanti'), vraj in 99.33-vrajate (for 'vrajati'); (b) adadi-ganiya i (to go) in 39.108-esyate (for 'esyati'), sna (+nic) in 68.2-snapayeta (for 'snapayet'); (c) divadi-ganiya nrt in 13.74-nrtyante (for 'nrtyanti'), tus in 51.11-tusyante (for 'tusyanti'); (d) svadi-ganiya ap (pra+) in 6.ii.40-prapayisyate (for 'prapsyati'), sadh in 1.24 sadhate (for 'sadhnoti'), sru in 3.4--srnusva (for 'srnu'); (e) tudadi-ganiya is in 4.17-iccheta (for 'icchet'), pracch in 2.18-prcchate (for 'prcchati'), 9.67, 13.1, 35.2-prcchase (for 'prcchasi'), 13.3-prcche (for 'prcchami'), 2.101-aprcchata (for 'aprcchat'), vis in 16.5-visate (for 'visati'). (ii) Atmanepadiya roots used in their Parasmaipadiya forms, viz., (a) bhvadi-ganiya badh in 16.5-badhanti (for 'badhante'), drs (used in the passive voice) in 85.60-drsyanti (for 'drsyante'), 95.16-drsyet (for 'drsyeta'), ji (vi) in 2.98-vijayami (for 'vijaye'), 98.9 vijayet (for 'vijayeta'), labh in 33.43-labhati (for 'labhate'), 11.11, 59.3, etc.-labhet (for 'labheta'), lamb (vi+) in 9.56-vilambasi (for 'vilambase'), rabh (sam+a+) in 58.10-samarabhat (for 'samara-

128 STUDIES IN THE UPAPURANAS bhata'), 59.20 samarabhet (for 'samarabheta'), ram in 82.75-ramanti (for 'ramante'), 7.92-rarama (for 'reme'), sah (sam+ut+) in 43.51-samutsahet (for 'samutsaheta'), sev in 93.20-sevanti (for 'sevante'), 10.iv.1, 8- sevet (for 'seveta"), stha (ava+) in 69.4-avatisthati (for 'avatisthate'), trai in 6.ii.12-sa traya (i.e. 'sa atrayat', for 'sa atrayata'), 7.79-traya (for 'trayasva'), yac 288 in 6.ii. 35, 16.41, 83.21-yaca (for 'yacasva'), 39.131-yayaca (for 'yayace'); (b) adadi-ganiya han (used in the passive voice) in 85.12-hanyanti (for 'hanyante'); (c) divadi-ganiya jan in 33.22, 28-jayanti (for 'jayante'), man in 93.290-manyet (for 'manyeta'), yudh in 4.13, 14-yudhyanti (for 'yudhyante'), vid in 33.30- ... na vidyati nrnam ....... ... (for '...... vidyate ...'), (d) kryadi-ganiya stambh in 7.75-stambha (for 'stabhana'-Lot, second person, singular). (iii) Irregular conjugation of verbs (sometimes with unauthorised change of Gana, sometimes without the augment a in Lan, sometimes with the loss of the reduplitive syllable, and so on): (a) bhvadi-ganiya bhas in 39.55-avabhasata (for 'avabhasata'), bhram (pari+) in 'paryabhramat'), 43.15-paribhramat (for bhu in 14.1-bhavat (without the augment 'a' in Lan, for 'abhavat'), bhu (sam+ut+) in 4.66-samudbhavat (without the augment 'a' in Lan, for 'samudabhavat'), dis (a†) in 39.133-adisaya (for 'adisa'), 288 It is rarely used as a Parasmaipadiya root.

THE SAKTA UPAPURANAS 129 dru (abhi+) in 11.23, 39.77, 43.15-abhidravat (without the augment 'a' in Lan, for 'abhyadravat'), gam in 32.1-jagma (for 'jagama'), gam (prati++) in 93.200-pratyagamami (for 'pratyagacchami'), ji in 4.74 jayisyati (for 'jesyati'), pat in 43.40-pata (for 'papata'), raks (conjugated as a curadi-ganiya root) in 12.47 raksayet (for 'rakset'), smr (anu+) in 14.3-anusmarat (in Lan, for 'anvasmarat'), sri (ut+) in 23.19-ucchraye (for 'ucchrayet'), stha (ava+, sam+ut+, ut+) in 13.45-avatisthata (without the augment 'a' in Lan, for 'avatisthata'), 85.33, 59-samutthanti (for 'samuttisthanti'), 4.31 utthan (for 'udatisthat'), subh in 2.53-sobhire (for 'susubhire'), vah in 13.25-vahan (in Lan, for 'avahan'), 14.29 vavaha (for 'uvaha'), vas in 10.iv.11-uset (for 'vaset'), vrdh in 43.3-vrdhe (for 'vavrdhe³), yac (conjugated as a curadi-ganiya root) in 9.52, 39.63 yacayami (for 'yace'), 39.59-yacaya (for 'yacasva'), 4.77-yacayisyati (for 'yacisyate), yac (prati+) in 43.17-pratyayaca (for 'pratyayacata'), yaj in 99.36-ijyet (for 'yajet'), yat (pra+) in 93.268-prayatneta (for 'prayateta'); (b) adadi-ganiya cakas in 119.38-cakasanti (for 'cakasati'), i (ut+) in 119.34-udayamasa (for 'udiyaya'), han in 9.56, 85.4-hana (for 'jahi'), 16.42, 39.131- vadha (for 'jahi'), 43.35, 68.16-hanet (for 'hanyat'), 15.22 jighamsa (for ''jaghana' or 'avadhit'), 83.23 vadhisyati (for 'hanisyati'), han (ni+) in 14.8-nighnami (for 'nihanmi'), 85.12-nighnanti (for 'nihanti'), 9 han (vi+) in 2.31, 32-vighnate (for 'vihanyate'),

130 STUDIES IN THE UPAPURANAS rud in 119.37-rurodanti (for 'rudanti'), stu (conjugated as a bhvadi-ganiya root) in 2.43 stavati (for 'stauti' or 'staviti'), 2.95, 11.29-stavate (for 'stute' or 'stavite'), 80.19-stavase (for 'stuse' or 'stavise'), va in 55.7-vayante (for 'vanti'), sas in 72.5-sasat (for 'asat'); (c) hvadi-ganiya da in 93.123-dadanti (for 'dadati'), 11.50, 86.28- dada (for 'dehi'), 91.74-dadet (for 'dadyat') (the forms 'dadate' in 37.2, 'dadase' in 83.15, and the like having been derived probably from the root 'dad' meaning 'to give'-cf. Panini's Astadhyayi iii.1.139 dadati-dadhatyor vibhasa, and Whitney, Sanskrit Grammar, 672 and 794 j), dha in 2.50, 116.iii.18 (antar+)-antaradhiyata (for 'antaradhita'-Lun, third person, singular), 48.22 (vi+)-vidhet (for 'vidadhyat'); (d) divadi-ganiya as (vi+ni+) in 44.1-vinyasat (for 'vinyasyat'), budh in 8.38-buddha (Lot, second person, singular, for 'budhyasva'), sam in 6.ii.44, 8.30-samate (for 'samyati'), 8.29- samet (for 'samyet'), vyadh in 43.36-vyadha (for 'vivyadha'), yudh in 39.82-ayodhyanta (for 'ayudhyanta'), 43.32 (vi)-vyayodhayat (for 'vyayudhyata'); (e) svadi-ganiya ap (pra+) in 6.ii.40-prapayisyate (for 'prapsyati'), sadh in 1.24-sadhate (conjugated as bhvadi-ganiya, for 'sadhnoti'), 4.18-sadhyanti (conjugated as divadiganiya, for 'sadhnvanti' or 'sadhayanti'), ur in 39.131-varam varaya govinda (for ... vrnu ..'); (f) tudadi-ganiya krt in 9.56-karta (for 'krnta'), 7.67 (vi+ni+)daityan hi vinikrntaya (for .. vinikrnta'), muc in 4.70, 7.87-mocate (conjugated as bhvadi-

THE SAKTA UPAPURANAS 131 ganiya, for 'muncati'), 119.35-mumocati (for 'muncati'), 15.18-vimoksa (for 'vyamuncat' or 'vyamukta'), vid in 2.30, 9.43-vindyat (conjugated as rudhadiganiya, for 'vindet'); (g) rudhadi-ganiya bhid in 13.21-bhindate (for 'bhinte'), bhuj (to protect, to enjoy) in 9.12-bhunjate (for 'bhunakti'), 4.10-bhunjatha (for 'bhunktha'), 9.8 bhunjami (for 'bhunajmi'), 2.42, 2.49, 5.16, 9.56, etc.- bhunja (for 'bhungdhi'), and so on, chid in 7.68-chinda (for 'chindhi'), 12.13-chindayet (for 'chindyat'), 12.20-chindeta (for 'chindita'); (h) tanadi-ganiya kr in 35.27 karanti (for 'kurvanti'), 9.8-kurvamah 35.27-karanti (for 'kurmah'), 7.78-kurva (for 'kuru'); (i) kryadi-ganiya bandh in 7.66, 9.56-bandha (Lot, second person, singular, for 'badhana'), grah in 9.56, 85.4, 116.iii.14-grhna (Lot, second person, singular, for 'grhanna'), 95.35-grhya (for 'grhana'), jna in 85.34 janayanti (for 'jananti'), li in 95.43-liyanti (for 'linanti'); (j) curadi-ganiya cint in 2.32-cintate (for 'cintayate'), dhu (ava+) in 4.64-avadhunayat 289 (Lan, third person, singular), kath in 1.31-kathisyati (conjugated as bhvadiganiya, for 'kathayisyati'), 82.1-katha (for 'kathaya'), kirt in 99.50-kirtanti (for 'kirtayanti'), pid in 24.18-pidati (conjugated as bhvadi-ganiya, for 'pidayati'), tad in 41.3, 4-tadayat (without the augment a in Lan, for 'atadayat'). (iv) Irregular Causative (nijanta) and Denominative verbal forms, viz., bhojiyat (for 'bhojayet') in 33.67, damsapayati (for *** The reading 'avadhunayat' seems to be a mistake for 'avadhullayat'.

132 STUDIES IN THE UPAPURANAS 'damsayati') in 9.68, mocapayati (for 'mocayati') in 9.68, niyamase (for 'niyamayasi') in 36.10, nrtyapayati (for 'nartayati') in 9.68, patayat (without the augment 'a' in Lan, for 'apatayat') in 41.8, pratisthayet (for 'pratisthapayet') in 32.8, 50.i.45, 50.ii.3, presayat (for 'presayet') in 40.14, samaye (Vidhi-lin, third person, singular, for 'samayet') in 7.72, tosati (for 'tosayati') in 2.40, utthapayat (for 'udasthapayat') in 39.154, vadayat (for 'avadayat) in 39.158, varate (for 'varayate') in 37.57, vinasisye (for 'vinasayisyet') in 1.3, homaye (Vidhi-lin, third person, singular, for 'homayet') in 26.36-satatam homaye'nale-should always perform 'homa' in fire, ksamayeta (for 'ksamapayet') in 31.27, 36.29. (v) Wrong use of Denominative verbs, as in 39.108 panya-striva yatha lobhat kamukanam varayate (for variyati')-just as a harlot wants to have, out of greed (for money), lustful persons as her husbands. (vi) Wrong formation of Krdanta words, viz., (a) Present Participles (in 'satr' and 'sanac'), such as bhunjayat (for 'bhunjat') in 8.40-bhunjayan punar anyas ca striyo ratna-vibhusitah, utthat (for 'uttisthat') in 85.43-kvacit patantam utthantam . . ., • 6. . bhramanta (for 'bhramat') in 85.38-bhramantavrttakutilam chadayanta (for 'chadayat') in 43.39 chadayanto disah sarvah purayanto navambubhih / amay-opari pata sah //, jvalanta (for 'jvalat') in 14.27 jvalanta-kopah, 84.14-jvalantagni-, purayanta (for 'purayat') in 43.39 (cited above), vahanta (for 'vahat') in 85.39-vahantaih, anicchamana (for 'anicchat') in 9.43-anicchamanapi tatha grhita panina kare (for 'anicchanty api ...'), cintamana (for 'cintayat') in 36.32-vedartham cintamanena visnuna dhyayamana (in active voice, for 'dhyayat') in 10.vi.4--yas tu kaya-krtan bhogan dhyayamanas tu sevate, ghatamana (in active voice, for 'ghnat') in 32.38-ghatamana ripum devi (for 'ghnati ...'), . . . " *

THE SAKTA UPAPURANAS 133 hasamana (for 'hasat') in 34.15-agatah hasamanah, japamana (for 'japat') in 39.50, 56, kridamana (for 'kridat') in 4.78, pathamana (for 'pathat') in 39.51, pravisamana (for 'pravisat') in 12.37-prathamam pravisamana bhumim yastir hanti rastram (for 'prathamam pravisanti bhumim .'), rudamana (for 'rudat') in 9.46-tena sa vimana drsta rudamana tu kanyaka (for . . rudati tu...'), tarjamana (for 'tarjayat') in 32.37-tarjamanam hatam murdhni naga-pasena vestitam (for 'tarjayantam ."), ghatamana (in the passive voice, for 'hanyamana') in 8.31-satravo ghatamana hi ksiyante hy avicaranat, 43.41-ghatamanasya; (b) Past Participles (in 'kta'), such as kaman devi ... ... '...... bhidita (for 'bhinna') in 52.7- ... astadha bhiditena chadmita in 1.39-chadmito guruna punah-again deceived by Guru (the preceptor of gods), ghatita (noncausative) in 11.41-visnuna ghatitah kecit ... (for '...... hatah ..'), 39.68-evam sa ghatitah (for ... sa hatah '), krntita (for 'krtta') in 119.9, pathata (for 'pathita') in 32.38-39- // sthapita pujita sakra smarata pathatapi va / prayacchati subhan , pranamita (for 'pranata') in 33.23 sarva-surasura-pranamitam sambhum prcchita (for 'prsta') in 13.2-kathitam yogam narada-prcchitam, 39.26-prcchito vrsa-vahanah, 93.15-prcchitah sivah, 92.15 (pra+)-devya praprcchitam, 39.38 (sam+)samprcchitas tatah sarve rangita (for 'ranjita') in 91.38, smarata, smarita and smarita (for 'smrta') in 32.39 (cited above in connection with 'pathata' for 'pathita'), 39.69-smaritam bhaktina, 43.58-yenaham smaritam tvaya; ... ... . . . . " (c) Gerunds (in 'ktva' and 'lyap'), such as ... apasthapayitva (for 'apasthapya') in 3.18, bhunjayitva (for 'bhuktva') in 9.45, chinditva (for 'chittva') in 86.22, kridayitva (for 'kriditva') in 87.14, namaskrtva (for 'namaskrtya') in 10.i.14, racitva (for 'racayitva') in 86.19, sammatikrtva (for 'sammatikrtya') in 7.96, smaritva (for

134 STUDIES IN THE UPAPURANAS 'smrtva') in 39.45, tyajitva (for 'tyaktva') in 17.31, ujjhayitva (for ''ujjhitva') in 13.13, uttha (for 'utthaya') in 85.58, vacitva (for 'uktva') in 28.7, vyapayitva (for 'vyapya') in 7.97, 39.147, yastva (for 'istva') in 58.13, bhojya (for 'bhojayitva') in 35.11, grhya (for 'grhitva') in 39.107, 112, khyapya (for 'khyapayitva') in 7.19, ksepya (for 'ksiptva') in 39.113, pasya (for 'drstva') in 14.22, presya (for 'presayitva') in 39.101, pujya (for 'pujayitva') in 5.13, 33.70, ruhya (for 'rudhva') in 39.71, samaruhya (for 'samruhya') in 39.112, snapya (for 'snapayitva') in 33.70, sthapya (for 'sthapayitva') in 4.4, 33.73, tarpya (for 'tarpayitva') in 2.73, tyajya (for 'tyaktva') in 14.12, and so on; (d) miscellaneous other primary bases, such as bhunjaniya (from the root 'bhuj') in 9.36-na hi anicchata (=anicchanti) bala bhunjaniya kadacana, darsakarin (for 'darsaka') in 10.iii.6-7, dhyapana (for ''dhyana') in 10.v.12-ity ete dhyapan-opaya rsibhih parikirtitah, samjitr (conqueror) in 15.3-yamasya vahner api samjitaram, vaivaha (for 'vivaha') in 27.6- ... yajna-vaivahasthapane-in sacrifice, marriage and consecration. (vii) Wrong use of Gerunds, as in ... ... 43.3 mahadevi ghoram hatva devena visnuna dattaparajita candre //, 55.4-ratrav indradhanur drstva ... smasane dhumo jayate, 95.36 bhuktva tu vipulan bhogan pascan mokso bhavisyati. (viii) Wrong use of Tenses and Moods: (a) Lat used for Lit, as in 15.18-vavarsa plavate (for 'plavayate') sarvam danavim vahinim tada-then rain poured down and flooded the entire army of the Danavas; (b) Vidhi-lin (Potential) used for Lan, as in 2.69-atandrita-manah sakra dharmadini na hapayet (for 'ahapayat'=ajahat), 2.71-72-ghrte va darpane vapi mukham pasyed (for 'apasyat') dadau ca gam // tatah sabham samasthaya pasyet (for 'apasyat') karyani karyinam /, 2.73-sabha-mandapam asthaya pasyet (for 'apasyat') svani balani ca, 2.76-mukto hy astadasair A

sa THE SAKTA UPAPURANAS 135 dosaih kuryad (for 'akarot') rajyam mahasurah, 13.67- casvina-prathamahe girindram avarohayet (for 'avarohayat'=avarohat), 13.69 karam prasarayed (for 'prasarayat') devyah -stretched out (his) hand towards Devi, 13.70-samdhyayam vijaya gatva karanayam nivedayet (for 'nyavedayat'), 13.72-ghoro 'pi svapnan pasyeta (for 'apasyat'), 40.13-15-mundam sampidayed (for 'samapidayat') devi presayat (for 'praisayat') saktim khadgena tadayet (for 'atadayat') //, 40.17-vayavyam praksiped (for 'praksipat') devi .., and so on; ... // indraya ... indram (c) Asir-lin (Benedictive Mood) used for Vidhi-lin, as in 12.54-ketor nirata yajane bhuyad vipra-kanyas ca, 67.53 vastram karpasikam dhriyat (for 'dharet', i.e. 'dharayet'). (d) Lrt (future) and Vidhi-lin (Potential) mixed up, as in 1.3-vinasisye (for 'vinasisyet'200 <'vinasayisyet' formed by wrong conjugation of the root 'nas'). (ix) Causative (nijanta) forms of verbs used for the non-Causative (anijanta), as in 2.69-hapayet (for 'ajahat'), 6.ii.40-prapayisyate (for 'prapsyati'), 9.56-bhedaya (for 'bhindhi'), 12.41 ghatitah (for 'hatah'), 13.67-avarohayet (for 'avarohat'), 20.18-tathastam prapite (for 'prapte') surye .. 43.59 mocitam (for 'muktam'), 43.61-ghatayitva (for 'hatva³), 66.6- ... devatah sarva vestayanti (for 'vestante') and so on. . . . . . . . . (x) Non-Causative (anijanta) forms of verbs used for the Causative (nijanta), as in 2.94- ... tasya deva-devo janardanah / dadarsa (for 'darsayamasa') svam tanum // Janardana, the god of gods, showed his own person to him ..., 9.51 tada tasya dadarsa tam-then (Nandin) showed him that 290 A similar form (karayisyet) occurs in Mahavastu i.267.9 (prose). See Edgerton, op. cit., I, p. 152 (31.37 and 40). See also Whitney, Sanskrit Grammar, 938 (pp. 333-4) for similar forms in the Rgveda, Aitareya-brahmana, Gopatha-brahmana, Mahabharata, etc.

136 STUDIES IN THE UPAPURANAS (Padamala Vidya), 7.6-stavenainam tutosa (for 'tosayamasa'), 15.18-(meghah) vavarsa plavate (for 'plavayate') sarvam danavim vahinim tada, 27.8-ya samate (for 'samayate') ca hutasanam-and which extinguishes the fire, 42.16 ... brahma-hatyadi-patakam / samate sa na samdeho vidya ... // (samate=samayate), 52.11 -snanam sisyaya kartavyam (for 'snapanam sisyasya kartavyam), 88.6-saktibhir jatam (for 'janitam'), and so on. (ix) Confusion as regards voice: (a) Active forms of verbs used in the passive voice, as in ... 2.31-32-vighnate (for 'vihanyate'), 2.77-so 'pi tatena drstavan (for 'drstah'), 21.8-krtavan sarva-devais ca pujah (for 'krtah '), 11.9-(vidya) bharadvajena praptavan (for 'prapta'), 15.4-bhayebhya muncanty avicaranena . . . tvam asrita vita-bhaya bhavanti mucyante ...'), 17.27-te 'pi papat pramuncanti (for 'pramucyante'), 90.4-visnuh sakrena prstavan (for 'prstah'), and so on. (for * ... (b) Passive forms of verbs used in the active voice, as in 83.2-tvam punah kathyate (for 'kathayasi') ... 3.2-tvam ca ... katham purvam prakathyase (for 'prakathayasi'), 4.86-tasya kah sakyate (for 'saknoti') yuddhe, 9.11-sa katham mucyate pathah (for muncati panthanam'), 9.74-yas tu devim na pujyate (for 'pujayate'), 43.40 sah (sarah) / rathanagasva-padatam hanyamanam (for 'ghnan') sahasradha //, 51.6-avidhau yah sivam pujyeta (for 'pujayeta'), and so on. (c) Active and passive voice mixed up, as in 2.23-sakrena ca samayantam drstva devam pitamaham / tyaktva simhasanam turnam dandavat patito bhuvi // 3.5 ... tena arjitam / jambum sakam tatha krauncam salmalim atha pauskaram // 4.42-43-indrena tau samayatau drstva yama-hutai

THE SAKTA UPAPURANAS 137 sanau mahaksobham samasthaya gaja-rajam ruroha sah // 8.4 ... ghorah / ... krto yatnah ... // • 8.11-dvahsthena tam samayantam drstva brahmasutottamam / pravisto yatra vai raja ghoro ghoraparakramah // 11.22 tena dvipadhipan jitva divam utsahate jaye, drstva vajra-dandena mahamaya- 15.18-tam samudbhavam / vayum mumoca 17.3-evam prstas tada brahma / vimrsya kathyate sarvam Il 35.32-rajna vanena vidhina dhvaja-yastim ... samucchrayet, and so on. (d) Wrong formation of passive verbs, as in 24.19-pidante nata-nartakah (for 'pidyante nata-'). XV. Wrong formation of Taddhitanta words, viz., darva (for 'darava'-wooden) in 25.14, 25.19, 32.39, sakhayatva (for 'sakhitva') in 39.146, sphatikamaya (for 'sphatika' or 'sphatikamaya') in 93.226, tapavan (for 'tapasvan³) in 2.37, tavanta (for 'tavat') in 50.i.27, vaimana (for 'vimana') in 64.3-vata-ramhasa-vaimanaih, vim- satima (for 'vimsatitama') in 10.ii.3-tac ca vinsatimam tattvam ........., yathasaktitah (for 'yathasakti') in 35.17- te yathasaktitas tosyah and so on. ..., XVI. Defects in versification: (i) Want of restriction with regard to the number of syllables in a Pada, as shown by (a) the Padas having more syllables than necessary, as in . . . . 2.7-sanakah sanatkumaras ca, 2.22-krtva kratu- satam vidhivat 3.16 hilihilo bhadranama ca, 4.78 _yama-mahisam ivaparam, 7.29-kunda-kutmalavad- abhasa, 8.7-yena kenacid upayena, 13.57-ingudodumbara-m-kharjura-, 27.2 manahsila-kusthakarpura-, 84.10, 86.1-kavacinah sottara-cchadah,

138 STUDIES IN THE UPAPURANAS 50.i.10-rddhir vrddhir unnatih siddhih, and so on, (b) the Padas wanting in syllables, as in 2.52-tam ayantam tu srutva, 5.12-tvam devarse viprendra, 7.64-mahavidyujjihvaya, 15.6-tam devi drstva bhasmi-prayatam (Upajati), 29.8-sakala-bhudivam bhutam, 57.21-sa ca mangalam rupam, and so on. (ii) Indiscriminate use of (metrically) long and short syllables in a Pada, as in 2.1-nrpavahana-mahatma, 2.8-sandilyo maharsir vahni-, 3.25-bhogah kulikah sauvarnah, 4.90 (second Pada) vinasah kena kriyate, 6.ii.20-pita-padmarunahema-, 6.ii.25-pasu-mrga-paksi-tiryak-, 7.14-karnapurita-isave, 7.27-ardhendur iva lalata-, 8.32-sabdasparsa-rasa-rupa-, 8.36-dandair vaca irsyasuua-, 54.9- rakta-karavira-puspaih, chinna-bahu-karakantha-siroru-vinistambhitah, 89.7 (first Pada)-kartavyam jita-dvandvena, and so on. 85.58 - (iii) Wrong pronunciation or non-pronunciation of vowels and consonants necessitated: (a) Long vowels to be pronounced as short, as in 7.14 jaya trisula-hastaya (in which the 'u' of '-su-' is to be pronounced as short), 7.22-rg-yajuh-samatharvanam (in which the 'a' of '-ma-' is to be pronounced as short); similarly in 8.3 (third Pada)-brhaspatina cakhyatam, 13.76 (third Pada)-tamo 'ndhakare kantare, and so on. (b) Short vowels to be pronounced as long, as in 5.14-tava ghora-suta vajra (in which the final 'a' of '-suta' is to be pronounced as 'a'), 6.ii.20-pitapadmaruna-hema- (in which the 'a' of '-na-' is to be pronounced as 'a'); similarly in 6.ii.26-lava-syandatruti-mesa-, 7.14-jaya sardra-gaja-carma-, karnapuritaisave, 7.27-ardhendur iva lalata-, 15.22-sarena haimadala-patritena, 32.2-sammarjanam upalepam, and so on. (c) Consonant-groups to be pronounced as single consonants, as in 6.ii.25-pasu-mrga-paksi-tiryak- (in which '-paksi-' * ✓ ☑

THE SAKTA UPAPURANAS 139 is to be pronounced as '-pakhi-'), 6.ii.33 (fourth Pada)- vetta tvam vedani ca tvam (in which 'tvam' is to be pronounced as 'tam'), 9.33-varuti puspa-granthini (in which '-gra-' is to be pronounced as '-ga-'); similarly in 27.14-satatyad avicchedini, 41.15-dahyamanah sasyandanah, 55.14-payasam dadhi-ksirajyam, 67.67- catuske ghrta-dyotitaih, 79.iii. 1-vratam me stustidam, 79.iv.8-patanti viyaccarini, 80.21-prcchami jagaddhetave, 85.58-siroru-vinistambhitah, 87.21 brahmana siva-skandena, and so on. mana- (d) Non-pronunciation of intermediate or final vowels of words, as in 2.7-sanakah sanatkumaras ca (in which 'sanat-' is to be pronounced as 'sant-'), 2.22-krtva kratu-satam vidhivat (in which 'vidhivat' is to be pronounced as 'vidhvat'), 3.16-hilihilo bhadranama ca (in which 'hilihilo' is to be pronounced as 'hilhilo'), 4.78-yamamahisam ivaparam (in which 'yama-' is to be pronounced as 'yam-'), 13.57-ingudodumbara-m-kharjura-matulungaih sa-dadimaih (in which the final 'a' of '-kharjura-' is not to be pronounced), 32.34-parasu-cakra-dharau canyau damaru-darpana caparau (in which 'parasu-' and 'damaru-' are to be pronounced respectively as 'parsu-' and 'damru-'), 32.36-kurvanti kalakalaravaih (in which 'kalakala-' is to be pronounced as 'kalkala-'), 51.2-mangalah sakambhari kali (in which 'mangalah' is to be pronounced as 'manglah'), and so on. That the above mode of pronunciation was actually in practice, is shown by the occurrence of the word 'parsu' (for 'parasu') in a number of verses, viz., 39.15 (parsuramena), 43.8 (parsuna) and 43.19 (parsuh). (e) Non-pronunciation of the final Anusvara and Visarga of words, as in 15.7-tvam bhumi-vayu kham jalam hutasanam (in which the Anusvara of 'kham' is not to be pronounced), 82.20-vibhuti-kridam samksepat (in which the Anusvara of '-kridam' is not to be pronounced), and so on; 3.25-bhogah kulikah sauvarnah (in which the

140 STUDIES IN THE UPAPURANAS Visarga of 'kulikah' is not to be pronounced), 27.21 amrtam havih kamikam (in which the Visarga of 'havih' is not to be pronounced), 31.39-sa sukham yasah saubhagyam (in which the Visarga of 'yasah' is not to be pronounced), 36.3-pulahadibhih sattamaih (in which the 'Visarga' of '-pulahadibhih' is not to be pronounced), and so on. In 'kavacinah sottara-cchadah' (84.10, 86.1), both the 'a' and Visarga of '-nah' are to be overlooked in pronunciation. So, 'kavacinah' is to be pronounced as 'kavacin'. (It may also be pronounced as 'kavcinah'). That the final Anusvara and Visarga of words were sometimes not pronounced (for the sake of metre) will be evident from the instances cited above (in section X). See also 65.52 -nigranthayah (for '-nirgranthayah'), 72.55-ity uvac-osana svayam, 84.26-asya sakti-dvitiyaham. (iv) Very frequent presence of hiatus (caused by absence of Samdhi) in verses, as in loke 1.14 varnasrama-sthitir yatra acarasya devi atma-bhav-anurupatah, 1.35- . . . . 1.36- ... tu rsinam . . . . 1.54- ' 1.26- yati ihapadam va adyadhyaya-trayam ..., 2.1-nrpavahana-mahatma agastyasyasramam gatah, 2.2- yatra rsayah ramante ekatah sada, 2.14- 2.5- . . . . 2.17 agastyas tisthate arghya-, yatha rtvig gurus caiva rsi-varair yuktah, 2.25- ... . . . > ramye pratipujam tu asan- . . . . 2.29 evam uktas tu indrena ..., 2.22 brahma veda-garbhaya utpatti- 2.34-para va apara vatha ..., 2.38-tena aradhitah purvam and .... so on. .... (v) Unwarranted Svara-bhakti (Anaptyxis) already made in the text, or required to be made in pronunciation, as in 10.ix.11-darasanam api punyam yoga-margasthitanam (for 'darsanam ..'), 15.6-tam devi drstva bhasmi-prayatam (in which 'bhasmi-' is to be pronounced, with Svara-bhakti, as 'bhasami-').

THE SAKTA UPAPURANAS XVII. Use of wrong words, such as bhuksa (for 'bubhuksa') in 33.6-bhuksormi-trsitam, 141 gavi, gavi (for 'go') in 64.1-vrsam gavim ..., 103.2- 103.8 subham hema- ... sukha-doham gavim mayim gavim karayet . . . . .... lancha (for 'lanchana') in 35.38 lanchena lanchitah . . . . ... vahananomali (for 'navamalika') in 62.2-nomali-kusumaih puja ་ varakin (for 'varaka') in 4.21-brahmanena varakina, and so on. See also 6.ii.19 ff. (for the words 'prasamati', 'bhavati', 'virodhaki, etc.). XVIII. Use of verbal roots in peculiar meanings, such as daks (in the sense of 'paying honorarium-daksina- to') in 33.79- kanya bhojyeta saktya dakseta vacayet, 99.43-bhojaniyas tatha daksed go-bhu-danahiranyatah. XIX. Use of nouns as adjectives, as in 4.11-majara-musikam yadvad yuddham dhvanksolukam yatha, 4.12-mahisasvam yatha yuddham yatha danti-mrgadhipam, 4.92 asva-mahisa-marjara-akhukakolukam... yuddham vanim 39.121-uvaca sausthavam XX. Use of synonymous words, as in 2.2-yatra veda-dhvanih sabdah, 2.3-vidya-vedakavettarah, 2.60-svadu-jal-odakah (dirghikah), 2.61 svabhava-prakrti-sthitah, 4.71-sakhilam sarvam, 13.76- tamo'ndhakare kantare, 14.23-vajrasani-vajra-kopah, 17.16-sarvan samastan api pidayitva, 33.7-vancita-moksamarga-rahitam, 43.14-prthivim vana-kananam, 83.100 ghanta-ninada-sabdena, 91.42-sarvartu-kusumaih puspaih, and so on. XXI. Ungrammatical construction of sentences, of

142 STUDIES IN THE UPAPURANAS which only a very few instances are cited below: 2.92 mantri-vaky-anal-otthena uddyotita vasum prati / matir danava-nathasya vrat-aradhanam ayayau // 4.32-33-yugantakas tatha cakre / mahisam yama-bhangaya mahisasya mahabalam // 4.42-43-indrena tau samayatau drstva ksobham samasthaya gaja-rajam ruroha sah // 4.70 na yuktam satru-paksasya vrddhim datum kadacana, ... | maha- 4.71-nipatya sakhilam sarvam mulam yasya na khanyate, 4.86-tasya kah sakyate yuddhe sabalasya nipatitam, 4.90-narayanastra-brahmastra-saivas canye 'tha varunah / tasya evamvidhah satror vinasah kena kriyate // 6.ii.18-tada tvaya mahadeva jnatva saktim mahabalam... vayam / tutosa ... ... 9.14-yada hi vyasanasaktam nrpam buddhiviparyaye vijnaya sa tadamatyah prakrtam darsayed bhayam // na I 9.23-stri-svarupa yada kincin maya vani samskrta / tathapi mama ksantavyam balanam hi na rustatam // (in which '-svarupa' has been used for '-svarupaya', and 'rustatam' for 'rusyatam'), 9.29-vayam tvam ca tathagaccha pratyaksam anu- sasyatam, 9.31-yathapi bhavato nasman vacayam api bhasate, 9.51-aradhya nandina purvam deva-devam jagadgurum yogabhyasena mahata tada tasya dadarsa tam // 11.46-47-yatha na saktah samare daityan yoddhum pitamaha / satrunam paribhutanam saranam tva gata vayam // yathaham cintayamasa sakra visnu-divaukasam / 16.2- susenah pratyabhasata / maya tvam sarva-devanam buddhir nabhyadhiko matah // 39.70-atha devair gate svargam ..., 39.76-atha tad-danu-rajendrah pitr-vairanalo 'bhavat, 3

THE SAKTA UPAPURANAS 39.134 purvam so'ham hato maya, 143 43.41 na samkhya vidyate tata ghatamanasya danavan (for ... tata hanyamananam danavanam³). The peculiar features of the language of the Devi-purana, noted above, are perhaps sufficient to show that they have much in common with those of Buddhist hybrid Sanskrit found in the Mahavastu, Saddharma-pundarika, Lalitavistara and a good number of other works on Buddhism. There is little scope for doubt that they are due chiefly to the influence of (early) Prakrit and Apabhramsa, which have controlled to a very great extent the formation of stems used in this work and are responsible for the peculiar features of its euphonic combination, declension, conjugation, formation of compounds, use of numbers, genders and cases, and so on. As a matter of fact, it is mainly Prakrit and Apabhramsa which form the basis of the un-Paninian Sanskrit of this Purana 291, which occupies a unique position in the Purana literature from the linguistic point of view. 291 That Prakrit and Apabhramsa had much to do with the language of the Devi-purana is evident from the rules of Prakrit grammars with regard to formation of stems, euphonic combination (Samdhi), formation of feminine bases, declension, conjugation etc. A few instances may be cited here: In his Prakrta-prakasa Vararuci says that the final 'r' of words is replaced by 'ara', in case any of the case-terminations follows (rta arah supi-5.31), that the word 'matr' has its final 'r' changed into 'a' under the same conditions (matur at-5.32), and that the words 'pitr,' 'bhratr' and 'jamatr' have, for their final 'r', 'ara' (and not 'ara') in all the cases (pitr-bhratr-jamatrrnam arah-5.34) and also 'a' in Nominative singular only (a ca sau-5.35). Hemacandra also says in his Prakrit grammar Siddha-hema-sabdanu- sasana that the (feminine) words 'svasr' etc. have the 'da'(=a) Pratyaya (and thus give rise to the stems 'sasa', 'duhia' etc.--svasrader da-3.35), that the final 'r' of words are substituted by 'ara' in the presence of any of the case-endings (arah sy-adau--3.45), that in the same circumstances the word 'matr' has its 'r' changed into 'a' and 'ara', with the former of which this word means 'a mother' and with the latter, 'a mothergoddess' (a ara matuh-3.46, bahulakaj janany-arthasya a devatarthasya tu ara ity adesah-commentary), and that (such) words (as 'pitr', 'bhratr' and 'jamatr') meaning 'designation' have their final 'r' changed into 'ara' in all the cases, but into 'a' in Nominative singular only (namny arah-3.47, a sau na va-3.48, in which the word 'naman' means 'samjna' and not "nomen" or 'noun' as held by Pischel, P. L. Vaidya and others). Similar statements have also been made in Trivikramadeva's Prakrta- sabdanusasana ii.2.41-svasrgat dal, and ii.2.49-52-arah supi, matur a ara, samjnayam arah, a sau va. By following these processes of change the Devi-purana forms its stems 'data', 'stota', 'trata', 'hotara', 'vettara', 'pita', 'pitara", 'mata", 'matara", 'matara', and so on, in imitation of those in Prakrit.

144 STUDIES IN THE UPAPURANAS Now, the question arises as to whether the Devi-purana is indebted to the Buddhist works for its peculiar language, In Vararuci's Prakrta-prakasa there is a rule (4.6--antyasya halah) which states that the final consonant of a word is dropped (in Prakrit). Hemacandra and Trivikramadeva also give similar rules in their Siddha-hema-sabdanusasana (1.11-antyavyanjanasya, 1.12-na srad-udoh) and Prakrta-sabdanusasana (i.1.25-antya-halo 'srad-udi) respectively. So, the Devi-purana has such stems as 'atharva', 'bhasma', 'brahma', 'adha', 'ambha', 'sapta', 'astadasa' etc. on the basis of their parallels in Prakrit. See also Prakrta-prakasa 4.7-striyam at and 4.9-na vidyuti, Siddha-hema- sabdanusasana 1.15-striyam ad avidyuti, and Prak rta-sabdanusasana i.1.29-avidyuti striyam at, which provide for the formation of Prakrit stems like 'saria', 'sampaa' etc. Consequently, the Devi-purana has 'apada', 'mrda', 'sampada', 'sruca', etc. (for 'apad', 'mrd', 'sampad', 'sruc', etc.). According to Vararuci, Hemacandra and Trivikramadeva, the pronoun 'idam' is replaced by 'ima' in all the cases (idama imah-Prakrta-prakasa 6.13, Siddha-hema- sabdanusasana 3.72, and Prakrta-sabdanusasana ii.2.76). So, the Devi-purana has the stem 'ima' (for 'idam'), which it declines as 'imaih' in Instrumental plural (masculine and neuter). In treating of euphonic combination Hemacandra says that Samdhi between the vowels of two words (of a compound) is optional (see Siddha-hema-sabdanusasana 1.5- padayoh samdhir va). Vararuci and Trivikramadeva also give similar rules in their Prakrta-prakasa (4.1--samdhav acam aj-lopa-visesa bahulam) and Prakrta-sabdanusasana (i.1.19-samdhis tv apade). So, for Sanskrit 'vyasarsih,' Prakrit has 'vasesi' and also 'vasa-isi' (with hiatus). Similarly, the Devi-purana has 'rasa-anya-kriya-', '-samkh- abja-ananta-', '-marjara-akhu-', etc. (with hiatus). According to Vararuci, Hemacandra and Trivikramadeva, a vowel, when followed by another, is often dropped (see Prakrta-prakasa 4.1 cited above, Siddha-hema- sabdanusasana 1.10-luk, Prakrta-sabdanusasana i.1.4-lopah). Thus, for Sanskrit 'tridasesah' Prakrit has 'tiasiso' (from 'tiasa+iso'3) often as 'bhunj' (as in 'bhunjate', 'bhunjatha', 'bhunjami, etc.). For change of the root 'kr' to 'kara' and 'jna' to 'jana' in conjugation, see Prakrtaprakasa 8.12-13-rto 'rah, krnah kuno va, and 8.23-jno jana-munau, Siddhahema-sabdanusasana 4.234-r-varnasy-arah, 4.7-jno jana-munau, and Prakrta- sabdanusasana ii.4.66-ara uh, ii.4.130-jana-munau jnah. So, the Devi-purana has the forms 'karanti', 'janayanti' etc. It is needless to multiply examples.

THE SAKTA UPAPURANAS 145 and, if not, how this type of incorrect Sanskrit came to be adopted in such an important work of the Hindus. The references to the decadent state of Buddhism in several places of this work and its mention of the Tantrik Bauddhas and of the Buddha as one of the ten incarnations of Visnu, tend to create the impression that in its language the Devi-purana was influenced by the Mahavastu and other Buddhist works mentioned above. But a critical study of the Devi-purana and a consideration of several matters relating to Indian religion, literature and culture encourage us to hold a perfectly different view. We have already seen that the Devi-purana, as we have it now, is the result of several revisions and abridgments to which its original text was subjected in different ages before attaining its present form, and that this work in its present state (excepting a very few interpolated passages and verses) comes down from the sixth century A.D. and most probably from its latter half. So, there can be little doubt about the fact that, in spite of its incorporation of matters bearing stamps of comparatively late dates, the Devi-purana contains chapters (or parts thereof) and verses which were composed much earlier than the sixth century A.D. Thus, it is evident that the language of this Purana had a much earlier beginning, which finds strong support in the Tantric Mantras and Vidyas recorded in it as well as in the striking similarity it has in some respects with the Mahavastu. It can hardly be denied that the various Tantric Vidyas, and more particularly the Tantric Mantras, like the Vedic ones, are not in every case new creations made by individual authors but have long traditions behind them. The frequent reference to Agamas and Tantras as authorities and the mention of a good number of them by name show that a rich Tantric literature had developed before the rise of the Devi-purana and that this work derived its Vidyas and Mantras from these Agamas and Tantras. As a matter of fact, the Devi-purana expressly says with regard to a particular 'Mahavidya' called Pada-mala (or Mantra-mala) that after extracting this Vidya from 'a crore of works' (koti-granthat) Siva spoke it out to Viresa (Nandin) in the 10

146 STUDIES IN THE UPAPURANAS Mulatantra 292. The Devi-purana even connects this Vidya with the Atharva-veda by calling it 'atharva-pada-dipani' 293. So, the history of the Vidyas and more particularly of the Tantric Mantras (both of which are replete with incorrect forms based on or influenced by Prakrit) goes far back to the pre-Christian days. Consequently, the period of origin of the hybrid Sanskrit of the Devi-purana has to be pushed farther back, probably up to, or even beyond, the time of first composition of the oldest Buddhist hybrid text, viz., Mahavastu, the nucleus of which is believed to have 'originated as far back as in the 2 nd century B.C., even though it was enlarged in the 4 th century A.D. and perhaps still later, by additions and interpolations' 294. This early origin of the language of the Devi-purana is supported by certain facts, which may be stated as follows. After critically examining the language of Buddhist hybrid texts Edgerton divides these works into three main classes 295, viz., class i consisting principally of the Mahavastu, and also of a short citation made in Santideva's 296 Siksasamuccaya from the lost Bhiksu-prakirnaka, and a short Jataka printed in the Appendix to Aryasura's 297 Jatakamala, class ii, of the Saddharma-pundarika, Lalitavistara, Gandavyuha, Kasyapa-parivarta, Dasabhumika-sutra, Rastrapala-pariprccha, Sukhavativyuha (both the longer and the shorter text), Samadhiraja-sutra, Suvarna- 29ª Devi-purana 9.65-devy uvaca kailasa-pitha-madhyastham viresam paramam prabhum/ ukta ya ca mahavidya mulatantre tvaya prabho/ koti-granthat samahrtya sarva-karma-pravartaki/ (The printed text of the Devi-purana reads 'samahitya' for 'samahrtya' contained in many Manuscripts ). 293 Ibid., 9.64- tatha te 'ham pravaksyami srnu tattvena vasava/ siddhanta-veda-karmanam atharva-pada-dipanim/ anaya tu sama vidya na bhuta na bhavisyati// 294 Winternitz, History of Indian Literature, II, p. 247. 295 See Edgerton, op. cit., I, p. xxv. 206 Santideva belonged to the seventh century A.D.-Winternitz, History of Indian Literature, II, pp. 365 ff.; Keith, History of Sanskrit Literature, p. 72. 297 Aryasura probably lived in the fourth century A.D.-See Winternitz, History of Indian Literature, II, p. 276; Keith, History of Sanskrit Literature, p. 68.

THE SAKTA UPAPURANAS 147 bhasottama-sutra, Udanavarga (containing no prose), and nearly the entire Siksa-samuccaya (consisting mainly of quotations from older texts mostly extinct at present), and class iii, of the Mula-sarvastivada-vinaya, Divyavadana, Avadana-sataka, Pratimoksa-sutra, Arya-manjusri-mulakalpa, Bodhisattvabhumi, Astasahasrika-prajnaparamita, Satasahasrika-prajnaparamita, Mahamayuri, Bhiksunikarmavacana, and a few other works (including the Lamkavatara-sutra). As regards the language of these three classes of works, Edgerton says that in the works of class i 'the prose parts are thoroughly hybridized, showing as many MiddleIndicisms as the verses', that in the works of class ii 'the verses are hybridized, as in the first class, but the prose contains relatively few signs of Middle Indic phonology or morphology', and in the texts placed in class iii 'even the verses (if any; ) are substantially as Sanskritised as the prose parts' and 'the entire text of these works resembles linguistically the prose parts of the second class'. These strata in the Sanskritisation of the Buddhist hybrid texts. indicate that about the second century B.C. when the Mahavastu first came into existence, hybrid Sanskrit as a language of religious texts was quite popular and that, with the progress of time and the consequent changes in the country, Sanskrit attained greater popularity even with the Buddhists, who consequently set their hands to Sanskritising their hybrid texts more and more, evidently for effecting better circulation of their religious views. Now, the date of original composition of the Saddharma-pundarika and the Lalitavistara is most probably to be placed in the second or even the first century A.D.298; and the verses of these two • • • Though the whole of the Saddharma-pundarika only gives expression to a later phase of Buddhism', the nucleus of this work is most probably to be dated as carly as the first century A.D., because it was quoted by Nagarjuna (who probably lived towards the end of the second century A.D.) and was translated into Chinese first in 223 A.D. and again by Dharmaraksa in 286 A.D. There were also other translations in later times. Winternitz, History of Indian Literature, II, pp. 303-4. Also Elliot, Hinduism and Buddhism, II, p. 52, and Nalinaksha Dutta in his Introduction (pp. xvxvii) to his edition of the Saddharma-pundarika. The Lalitavistara belongs to the first two centuries of the Christian era but also contains materials of later dates.-See Winternitz, History of Indian Literature, II, pp. 255-6; P. L. Vaidya's Introduction (pp. xi-xii) to his edition of the Lalitavistara.

148 STUDIES IN THE UPAPURANAS works as well as of the others constituting class ii have Prakritisms as frequently as the Mahavastu, but their prose parts show signs of the influence of Prakrit on a small number of occasions. A greater tendency towards Sanskritisation is found even in the early works placed in class iii. For instance, the Avadana-sataka, which was composed about the first or second century A.D. 299, has very few Prakritisms in its verses, its prose portions being almost clear of these. Similar peculiarities are also found in the language of the other works of the third class. So, it can hardly be denied that from about the beginning of the Christian era Sanskrit had begun to find favour with the Buddhists and to make its way into the Buddhist works in a greater degree, with the result that in works like the Pratimoksa-sutra as found in Anukul Chandra Banerjee's edition (Calcutta, 1954), prose and verse have become almost totally free from Prakritisms. As regards our Devi-purana, it is remarkable that the prose portions show as much influence of Prakrit as the verses; and in this respect it is similar to the Mahavastu, which also, unlike the other hybrid texts of the Buddhists, has Prakritisms equally in its verses and prose parts. There is another point of agreement between the Mahavastu and the Devi-purana Unlike the other Buddhist hybrid texts the former has its Nominative and Accusative singular ending in 'u' in rare cases 300. Similarly, the Devi-purana, as we have already seen, appears to have the final 'a' of a word replaced by 'u' on a single occasion (in Devi-purana 7.12-kala-maha-kutavisa-kanthastha-jirnave). So, the history of the language of the Devi-purana can reasonably be traced back to the time of origin of the Mahavastu. But the dissimilarities between the language of the Devi-purana and the Mahavastu are so varied and serious and outnumber the similarities to such an extent that the former 2** This date of the Avadana-sataka is based mainly on its translation into Chinese in the first half of the third century A.D. and on its mention of Dinara as a minted coin (laksanahatam dinara-dvayam).-Winternitz, History of Indian Literature, II, p. 279, and P. L. Vaidya's Introduction (p. x) to his edition of the Avadana-sataka. 300 See Edgerton, op. cit., I, pp. 49 (8.20-21) and 51 (8.30). Also ibid., pp. 12 (1.95) and 26 (3.51 ff.).

THE SAKTA UPAPURANAS 149 work can never be said to have been influenced in its linguistic peculiarities by the latter. In order to show the nature of the Prakritisms in the Mahavastu, two extracts are quoted below from its verse and prose protions. (a) apare pi parasparam pradustamanasamkalpa ayasehi nakhehi patenti tiksnani ca asipatrani hastesu pradurbhavanti gatrani cchindanti na ca kalam karonti yavat sanam papaka karma na pariksina // kalasutre mahanarake adraksit satvam kalasutrena sutritanga nihataksiyantam parasuhi pi vipatiyantam karapatrehi vipatiyantam // Mahavastu Avadana, p. 5. (b) imam lokam paralokam satvanam agatim gatim / cyuti-upapatti-samsaram sambuddho svayamaddasa // avajjanto saphalatam karmanam prana-samsritam / yathasthanam vipakam ca svayam avabudhye munih // so abhijnaya akhyasi narakan asta gautamah / pratyaksa-dharma bhagavam sarva-dharmesu caksumam // Mahavastu Avadana, p. 9. That these sorts of Prakritisms are not peculiar with the Mahavastu only, will be evident from the following quotations made from the Saddharma-pundarika, Lalitavistara, Gandavyuha and Sukhavativyuha: (a) yehi mahi sobhatiyam samantat parsas ca catvara sulabdha-harsah / sarvam ca ksetram imu samprakampitam sadbhir vikarehi subhisma-rupam // sa caiva rasmi purima disaya astadasa-ksetra-sahasra-purnah / avabhasayi eka-ksanena sarve suvarna-varna iva bhonti ksetrah // yavan avici paramam bhavagram ksetresu yavanti ca tesu sattvah | satsu gatisu tahi vidyamanah cyavanti ye capy upapadyi tatra // karmani citra vividhani tesam gatisu drsyanti sukha dukha ca / hina pranita tatha madhyama ca

150 STUDIES IN THE UPAPURANAS iha sthito addasi sarvam etat // Saddharma-pundarika, p. 5 (Gathas 3-6). atitam adhvanam anusmarami acintiye aparimitasmi kalpe / yada jino asi prajana uttamas candrasya suryasya pradipa nama // saddharma deseti prajana nayako vineti sattvana ananta-kotyah / samadapeti bahu-bodhi-sattvan acintiyan uttami buddha-jnane // Saddharma-pundarika, p. 18 (Gathas 57-58). (b) athapi va trnasya tuli bhumito grhitvana dadati maya aturana sarvi bhonti nirjara / saukhya-prapti nirvikara gehi gehi gacchisu bhaisajya-bhuti vaidya-raji kuksi-sampratisthite // yasmi kali maya-devi svatanum niriksate adrsati bodhi-sattva kuksiye pratisthitam yathaiva candra antariksa tarakai parivrtam tathaiva nathu bodhi-sattva-laksanair alamkrtam // Lalitavistara, p. 55 (verses 31-32). (c) tvat-prabhavata aham mahamate bodhi prasthitu hitaya dehinam / tatra niscayu ananta-gocaro (d) yo mama bhavati tam srnohi me // Gandavyuha, p. 54 (Gatha 1). ye eta naya samudran avatirna sthihitva buddhabhumi ye / te bhonti sarva-darsi siksanto lokanathanam // ... Gandavyuha, p. 57 (Gatha 7). ... 1 surya-mani-girisa-candra-abha na tapita bhosisu ebhi sarva-loke || rupam api anantu sattva-sare tatha api buddha-svaro ananta-ghosah / silam api samadhi-prajna-viryaih sadrsu na te 'stiha loki kascid anyah // gabhiru vipulu suksma-praptu dharmo-

THE SAKTA UPAPURANAS 'cintitu buddha-varo yatha samudrah / tenonnamana na casti sastuh khila-dosan jahya ato 'dhikalam // 151 Sukhavativyuha, pp. 7-8 (Gathas 1-3). It will be noticed that in these extracts, as also in the other relevant parts of the hybrid texts, Prakrit phonology and morphology have been followed to a very great extent. Thus, the conjunct consonants have often been assimilated, single consonants have not rarely been dropped or replaced by others, non-final vowels of words have been subjected not infrequently to various changes or eliminated in some cases, new vowels have been inserted at times, non-Sanskrit endings (Vibhaktis) have very often been used in declension and conjugation, and so on. The Devi-purana, on the other hand, has Prakrit influence in a much limited sphere. It allows assimilation of conjunct consonants not in its written text but only in pronunciation, evidently for the sake of metre; it does not drop or change any single consonant unless it is final; it scrupulously avoids the use of nonSanskrit endings of verbal inflection, and has non-Sanskrit case-endings only in extremely rare cases 301; and so on. As a matter of fact, the Devi-purana uses Prakrit practically as a basis for its un-Paninian Sanskrit but does not allow it to exert too much influence on the latter and thereby to obliterate its character. So, the language of this unique Purana follows a tradition very different from that of the Buddhist hybrid texts. The occasional similarity in language, which the Devi-purana has with the Arya-manjusri-mulakalpa and the Lankavatara-sutra, does not deserve any serious consideration, because these two Buddhist texts come from comparatively late dates 302 and the points of their linguistic 301 As in '-prakridaya' (for '-prakridayai') in Devi-purana 8.20, 'nandaya' (for 'nandayah'-Genitive case) in Devi-purana 93.273, 'trtiyam" (for 'trtiyayam') in Devi-purana 99.12, and 'samharakaraya rudra-deha-bhavaya' in Vocative feminine in Devi-purana 16.18. 30% Regarding the period of original composition of the Arya-manjusri-mulakalpa Benoytosh Bhattacharya says: "The Manjusri-mulakalpa deals with the formulae and practices which lead to the material prosperity of the followers of Mahayana, and probably belongs to the early centuries A.D. but decidedly after the time of the composition of the Amitayus Sutra or the Sukhavati Vyuha which ushered in the conceptions of Amitabha or Amitayus and Avalokitesvara for the first time in Mahayana. The

152 STUDIES IN THE UPAPURANAS dissimilarity with the Devi-purana far outnumber those of their similarity with it. There is another matter which cannot be overlooked here. Long before the rise of the Mahavastu and other Buddhist hybrid texts the Hindus accepted the standard of correct Sanskrit set up by Panini in his famous grammar Astadhyayi; and the language of their Sanskrit works, written after Panini's time, has been controlled to a very great extent by the rules laid down by that great grammarian. On the other hand, the Buddhists were looked down upon and called Pasandas (heretics), with whom the Hindus following the Vedic tradition did not like even to meet or talk. Thus, the Visnudharma, which was composed some time during the third century A.D. 303, severely denounces the heretics (including the Buddhists) 304, identifies the founders of the heretical faiths with the dreadful Mahamoha born of the malevolent rite (krtya) performed by the demons Sanda and Marka for the destruction of the gods 305, and characterises the language of the religious works of these heretics as 'Meccha-bhasa 306. The Visnu-purana gives out the Bauddhas, Arhatas (Jainas), and other anti-Vedic sectaries Amitayus Sutra was first translated into Chinese at a period between A.D. 148 and A.D. 170 and hence the time of its composition may be fixed at about 100 A.D. or a little later. The Manjusri-mulakalpa in that case would be only about a hundred years later than the Amitayus Sutra." (See B. Bhattacharya's Introduction, pp. xxxiv-xxxvi, to his edition of the Sadhana-mala, Volume II). But according to M. Winternitz, the arguments, put forward by B. Bhattacharya in favour of such an early date of the Arya-manjusri-mulakalpa, 'are by no means convincing.' (See Winternitz, History of Indian Literature, II, p. 635). That this work was composed earlier than the second quarter of the tenth century A.D., admits of no doubt, because it 'was translated into Chinese as a Maha-Vaipulya-Sutra between 980 and 1000 A.D. and into Tibetan as a Tantra in the 11 th century.' (See Winternitz, op.cit., II, p. 397). J. N. Farquhar ascribes it to the tenth century A.D. (See J. N. Farquhar, pp. 272 and 398). 303 For the evidences in favour of this date of the Visnudharma see Volume I, pp. 137-143. 304 Ibid., pp. 124 and 147-149. In chap. 105 of the Visnudharma the following heretical sects have been named: Utkocas (v.l. 'utkaucah' for 'utkocah'), Saugatas, Mahayanists, Kapilas, Bhiksus, Sakyas, Sravakas, Nirgranthas, and Siddhaputras. (For the relevant verses of the Visnudharma see Volume I, p. 150). 306 For a summary of the relevant story contained in Visnudharma, chap. 25, see Volume I, pp.128-9. 306 For the relevant verse of the Visnudharma see Volume I, p. 150,

THE SAKTA UPAPURANAS 153 as 'Pasandas' and 'Nagnas' (naked, i.e. bereft of the raiment or covering consisting of the triple Veda) and fabricates stories to show how the demons led themselves to ruin by forsaking the Vedic Dharma and taking to the religious rites and practices of the Bauddhas, Jainas and other heretics, and how king Satadhanu incurred sin by speaking with a heretic just after his holy bath in the Ganges during his observance of a fast on the full-moon day of the month of Karttika and consequently became a dog, a jackal, a wolf, a vulture, and so on in his successive rebirths. The Purana (III.18.95-102) then says: "esa pasanda-sambhasa-dosah prokto maya dvija / tasmat pasandibhih papair alapa-sparsane tyajet / visesatah kriya-kale yajnadau capi diksitah // kriya-hanir grhe yasya masam ekam prajayate / tasy-avalokanat suryam pasyeta matiman narah kim punar yais tu sa tyakta trayi sarvatmana dvija / paranna-bhojibhih papair veda-vada-virodhibhih // pasandino vikarmasthan baidala-vratikan sathan / haitukan baka-vrttims ca van-matrenapi narcayet // durad apastah samparkah sahasyapi ca papibhih / pasandibhir duracarais tasmat tan parivarjayet // ete nagnas tavakhyata drstya sraddh-opaghatakah / yesam sambhasanat pumsam dina-punyam pranasyati // ete pasandinah papa na hy etan alaped budhah / punyam nasyati sambhasad etesam tad-din-odbhavam //" "Such is said to be the sin of speaking with a heretic, [O] twice-born man, "So, [one] must avoid discourse with and contact of the heretical sinners, especially at the time of observing [a religious rite] and also after getting initiated for [the the performance of] sacrifice etc. "A wise man is to look at the sun after beholding one in whose house there is omission of [Vedic] rites for a month. "What, again, [O] twice-born man, [can he be required to do after beholding] those sinning anti-Vedic

154 STUDIES IN THE UPAPURANAS parasites who have forsaken the triple Veda completely? "One must not honour even with a word [of the mouth] the heretics, the performers of acts not meant for them, the followers of the [deceitful] course of conduct of cats, the scoundrels, the sceptics, and those [hypocrites] who behave like cranes, "far from having any intercourse with these sinful heretics of ignoble practices; so, one must avoid them by all means. "These persons, spoken of to you, are the Nagnas, who spoil a Sraddha ceremony by their very sight, and conversation with whom destroys the [religious] merit of persons earned in a day. "These are the sinning heretics, with whom a wise man must not speak, [because] the [religious] merit obtained [by the speaker] on a particular day will be lost through conversation with them." In his commentary on Yajnavalkya-smrti 1.103 Apararka ascribes to the 'Brahmanda-purana' a number of verses in which the Bauddhas, Jainas, Kapalikas and others have been decried as 'following bad practices and being foreign to the acts of purification' (duracarah saucacara-bahiskrtah) and classed with the outcasts. The Sattrimsan-mata, as quoted in Devanabhatta's Smrti-candrika (II, p. 310), says: "bauddhan pasupatan jainan lokayatika-kapilan / vikarmasthan dvijan sprstva sacelo jalam aviset / kapalikams tu samsprsya pranayamo 'dhiko matah //" "After touching the Bauddhas, Pasupatas, Jainas, Lokayatikas, Kapilas, and those twice-born [people] who take to unlawful acts, [a man] should get into water with [all his] clothes on; but in case of contact with the Kapalikas [performance of] pranayama is prescribed as [an] additional [duty]." Going to suggest an alternative explanation of two Sutras of Jaimini 307 for the rejection of the scriptures of the Buddhists and other anti-Vedic sectaries as authorities on 307 Viz., Purva-mimamsa-sutra i.3.3-virodhe tv anapeksam syad asati hy anumanam, and i.3.4-hetu-darsanac ca.

THE SAKTA UPAPURANAS 155 Dharma, Kumarila-bhatta points out that the texts compiled by Sakya (i.e. the Buddha) and the propounders of the systems of Samkhya, Yoga, Pancaratra, Pasupata and the like, in which there are statements about Dharma and Adharma, are not accepted as authorities by those who know the Veda, that these texts have in them a certain mixture of the Vedic Dharma which serves as a thin cloak to hide their true character, that they are meant for achieving selfish ends such as a good following, pecuniary gain, respect, fame, etc., and that, in spite of their occasional recording of statements about non-killing, truthfulness, self-control, charity, kindness, etc. in agreement with those of Sruti and Smrti, they are opposed to the Veda. So, Kumarila says, the said two Sutras establish the unacceptability of the above-mentioned works as authorities on Dharma for their anti-Vedic character as well as for the worldly motives lying behind them, because the rejection of the authority of these works has not been established in any other Adhikarana and also because they are sufficiently known and have wide circulation "like the knowlegde of the imports of [such incorrect forms of] words [as] 'gavi' etc."308 Again, after referring to the statement of the Puranas that Sakya and others would become the causes of confusion of Dharma in the Kali age, Kumarilabhatta says in his commentary on Purva-mimamsa-sutra i.3.7 that, like very small quantities of real gold-dust, camphor etc. put in masses of artificial ones, the few Vedic truths, mixed up in these works 30º yad va, yany etani trayividbhir na parigrhitani kimcit-tan-misra-dharma-kancukacchaya-patitani lokopasamgraha-labha-puja-khyati-prayojana-parani trayi-viparit-asambaddha-drsta-sobhadi-pratyaks-anuman-opaman-arthapatti-praya-yukti-mul-opanibaddhani samkhya-yoga-pancaratra-pasupata-sakya-nirgrantha-parigrhita-dharmadharma-nibanvisa-cikitsa-vasikaran-occatan-onmadanadi-samartha-katipaya-mantr-ausadhidhanani kadacitka-siddhi-nidarsana-balena ahimsa-satyavacana-dama-dana-dayadi-sruti-smrtisamvadi-stokartha-gandha-vasita-jivika-pray-arthantar-opadesini,. tesam evaitac-chruti-virodha-hetu-darsanabhyam anapeksaniyatvam pratipadyate/ na caitat kvacid adhikaranantare nirupitam na cavaktavyam eva gavy-adi-sabda-vacakatva-buddhivad atiprasiddhatrat yadi hy anadaren-aisam na kalpyet-apramanata/ asakyaiveti matv-anye bhaveyuh sama-drstayah// 1 Tantra-varttika, pp. 114-5. For this extract see also Mimamsa-darsana, p. 194 where it reads, '-grantha-' (for '-nirgrantha-').

156 STUDIES IN THE UPAPURANAS with numerous other statements misrepresenting Dharma, lose themselves in the latter, and thus, like milk kept in a bag of dog-skin, remain unacceptable to us until they are discovered in the recognised Dharmasastras (already mentioned in the Tantra-varttika). Kumarila adds that when these truths are clearly known from these Dharmasastras, the works of Sakya and others become useless for us. So, Kumarila concludes, no compilation outside the recognised Sastras such as the Veda etc. is to be recognised as an authority on the Vedic Dharma. 309 At another place of his Tantra-varttika Kumarila-bhatta says that the scriptures of the Sakyas, Jainas and others, abounding in incorrect words and being thus bad and unreliable compilations, cannot be recognised as Sastras and that they are composed in incorrect words consisting mostly of those of Magadha, Daksinatya and their Apabhramsas. Thus, Kumarila says, the objects denoted by such words as are themselves unreal, cannot be accepted as real, and the words having deteriorated forms, cannot be recognised as eternal.³ 310 It is needless to refer to any more authorities. From the Visnudharma and other works cited above it is evident that the followers of the Vedas and the varnasramadharma bore a very strong feeling of hatred towards the 309 smaryante ca puranesu dharma-vipluti-hetavah/ kalau sakyadayas tesam ko vakyam srotum arhati// yatha krtaka-karpura-suvarnadisu diyate/ yad bijam tad api vyaktam agrahyatvat praliyate// tena karm-anurupya-samanyato drstarth-apatti-balat tad-abhipraya-kalpita-dharm- abhasa-madhya-patitam san-mulam apy ahimsadi sva-drti-niksipta-ksiravad anupayogy avisrambhaniyam ca tan-matr-opalabdham bhavatity avasyam yavat-pariganitadharmasastrebhyo nopalabhyate tavad agrahyam bhavati/ yada sastr-antaren-aiva so'rthah spasto 'vadharyate/ tada tenaiva siddhatvad itarat syad anarthakam// tasmad yavat-pariganita-vedadi-sastra-vyatirikta-nibandhanam tad-dharma-pramanatvena napeksitavyam iti.-Tantra-varttika, p. 127. 310 asadhu-sabda-bhuyisthah sakya-jain-agamadayah/ asan-nibandhanatvac ca sastratvam na pratiyate// magadha-daksinatya-tad-apabhramsa-pray-asadhu-sabda-nibandhana hi te/ 'mama vi hi bhikkhave kammavac ca isi save/ tatha ukkhitte lodammi uvve atthi karanam/.... anuppatti-karanam' ityevam-adayah/ ... tatas casatya-sabdesu kutas tesv artha-satyata/ drst-apabhrasta-rupesu katham va syad anadita// Tantra-varttika, p. 171.

THE SAKTA UPAPURANAS 157 Buddhists and their scriptures and had no respect for their language, which they called 'Mleccha-bhasa' and denounced as being quite unfit for religious works. Under these circumstances the Devi-purana, which is a religious work of the Hindus and holds the Vedas in high esteem as the sources of Dharma, cannot be expected to have derived its peculiar language from the Buddhist religious texts. If, in spite of the strong feeling of antagonism and hatred between the Hindus and the Buddhists, this Purana was written by somebody in such undesirable language in imitation of the Buddhist works, it would hardly be able to attain wide popularity and to occupy and enjoy for long a position of high authority even among the Smrti-writers including Laksmidhara, Apararka and others who do not recognise the scriptures of the Pancaratras, Pasupatas and similar other sectaries as sources of Dharma³11. As a matter of fact, we do not know of a single scholar, Smarta or otherwise, who has raised the slightest objection against its language. It is true that Vallalasena does not utilise its contents in his Danasagara and rejects its authority with the following remark: "tat-tat-puran-opapurana-samkhya-bahiskrtam kasmalakarma-yogat / pasanda-sastr-anumatam nirupya devi-puranam na nibaddham atra//"312 "Finding that the Devi-purana, which was excluded from the [recognised] lists of Puranas and Upapuranas due to its connection with (i.e. treatment of) foul acts, followed 311 As a matter of fact, Laksmidhara does not draw upon the Pancaratra Samhitas, Saivagamas etc. in his Krtya-kalpataru, nor does he care to mention the names of these works even by way of rejection in course of selecting the authorities on Dharma at the beginning of the Brahmacari-kanda. Apararka, on the other hand, discusses at length the question of accepting the religious texts of the Saivas, Pasupatas, Pancaratras and other similar sectaries as sources of Dharma and comes to the following conclusion: These works are not to be decried, nor are their prescriptions to be carried out into action in their totality; and one may follow in practice only such statements of these works as agree with, or do not go against, the directions of the Vedas. But those texts of these sectaries which are of human authorship (pauruseya), are to be rejected without any reservation. (See Apararka's commentary on Yajnavalkya-smrti 1.7). *1* Danasagara, p. 7, verse 67,

158 STUDIES IN THE UPAPURANAS the scriptures of the Pasandas, it was not drawn upon here." But the word 'pasanda', as used in this verse, definitely means the 'Tantriks', and not particularly the 'Buddhists'³13; and this meaning of the word finds strong support in Vallalasena's statements made in the introductory verses of the Danasagara about the Bhavisya-purana and five other Puranic works which he considered spurious. Going to explain why he did not utilise those parts of the Bhavisya-purana which dealt with the procedures of religious rites to be observed on the Astami and Navami Tithis, Vallalasena says: "saptamy-avadhi puranam bhavisyam api samgrhitam atiyatnat / tyaktv-astami-navamyoh kalpau pasandibhir grastau //3314 313 In his Introduction (p. xxiii) to his edition of the Danasagara Bhavatosh Bhattacharya takes the word 'pasanda-sastra' in the said verse to mean 'the Buddhistic Scriptures.' 314 See Danasagara, p. 7, verse 59. The printed edition wrongly reads 'saptamy-aiva' (for 'saptamy-avadhi') in the first line, and '-navamyau' (for '-navamyoh') in the second. Of these two readings the former is clearly arbitrary as well as nonsensical, being supported neither by any of the Manuscripts used in this edition nor by sense or grammar, and the latter, though contained in two Manuscripts, is equally unsuitable for the same reasons. So, we have to prefer the readings 'saptamy-avadhi' and '-navamyoh' given by two Manuscripts of the Danasagara, viz., the India Office Manuscript (for which see Julius Eggeling, Part III, p. 544) and the Vangiya Sahitya Parisat Manuscript (No. 1374, fol. 3 a). It is remarkable that these two readings can be made to stand in correct grammatical relation with the other words of the said verse and to create a perfectly relevant meaning in it, and that the former (viz., 'saptamy-avadhi') finds considerable support in the corrupt reading 'saptamy-avava" of Manuscript B written in Bengali scripts. Moreover, in the printed edition of the Bhavisya-purana the Brahma-parvan (which consists of chapters 1-215) is found to deal with the 'kalpas' (procedures of religious rites) to be followed on the different Tithis from Pratipad to Saptami in the following chapters: Chaps. 16-18-Pratipat-kalpa, chapters 19-20-Dvitiya-kalpa, chap. 21-Trtaya-kalpa, chapters 22-31-Caturthi-kalpa, chapters 32-38-Pancami-kalpa, chapters 39-46-Sasthi-kalpa, chapters 47-215 Saptami-kalpa (including stories of Samba and others relating to Sun-worship). Vallalasena's statement shows that the Manuscript of the Bhavisya-purana, current in Bengal in his days and used by him in his Danasagara, contained chapters on the Astami- and the Navami-kalpa and that these chapters had already been influenced by the Pasandas in Vallalasena's time. That chapters on the Astami-kalpa and Navami-kalpa did once occur in the Bhavisya-purana, is also shown by some Manuscripts of this work (for which see Haraprasad Shastri, V, Nos. 3739 and 3741, Haraprasad Shastri, II, No. 151, and R. L. Mitra, Nos. 1742 and 2553) as well as by those verses, ascribed to the 'Bhavisya-purana' in Hemadri's Caturvarga-cintamani (Volume II) and in other Smrti Nibandhas, which deal with the A

THE SAKTA UPAPURANAS 159 "The Bhavisya Purana also has been drawn upon with much care up to [the part dealing with] the Saptami [-kalpa], leaving off [the parts on] the Astami- and the Navami-kalpa [which were] taken up by the Pasandins." In naming the said spurious Puranas and giving a short account of their extents and contents Vallalasena says: "tarksyam puranam aparam brahmam agneyam eva ca / trayovimsati-sahasram puranam api vaisnavam // sat-sahasra-mitam laingam puranam aparam tatha / diksa-pratistha-pasanda-yukti³15 -ratna-pariksanaih // mrsa-vamsanucaritaih kosa-vyakaranadibhih / asamgata-katha-bandhaih paraspara-virodhibhih // tan-minaketanadinam bhanda-pasanda-linginam / loka-vancanam alokya sarvam evavadhiritam ||'3316 vows and worships (including that of the goddess Durga) to be observed on the Astami and Navami Tithis but do not occur in the present Bhavisya-purana We have already seen that the Bengal Smrti-writers are found to ascribe to the 'Bhavisya-purana' a good number of verses on Durga-puja, some of which seem to betray Tantric influence but are not traceable in the printed text of this Purana. It should be mentioned here that it is the Bhavisyottara, and not the present Bhavisya, which is found to deal with Durgapuja in chap. 138, that this work, though claiming to belong to the Bhavisya-purana as its final part and printed as such with the latter, is an independent work of a different date, and that its chap. 138, which deals with the Mahanavami-vrata requiring the worship of Durga on the Astami and more particularly on the Mahanavami Tithi, contains no trace of Tantric influence, even the Mantras prescribed for use in Devi-worship being Puranic, and not Tantric. (See Bhavisyottara 138.41 b-42 a-puja-mantran pravaksyami puranoktan aham tava / yaih pujitah prayacchanti kirtim ayur yaso balam //). (For a more detailed discussion on the suitability of the said two readings see our article on a 'Critical Examination of Some Readings of the Danasagara' published in Our Heritage, Volume VIII, 1960, Part 2. 315 The word 'pasanda-yukti' means 'usages or practices of Tantriks', and not 'arguments of the Buddhists' as said by the learned editor of the Danasagara in his Introduction, p. xxiii. It is well known that the Puranas, as we have them at present, were meant for religious synthesis as well as for the popularisation of the varnasramadharma and the authority of the Vedas in society. On the other hand, the Buddhists have in all ages been openly anti-Vedic. So, the Puranas are found to decry Buddhism often in very strong terms, although in the comparatively late Puranic works the Buddha has been respected not as such but as an incarnation of Visnu come down to the earth for deluding people and making them forsake the right path of the Vedic Dharma. Under these circumstances, it was not, and could not be, possible for the Buddhists to have anything to do directly with the Puranas or to introduce their 'arguments' into these works for helping the propagation of their religious faith. It is true that the present Puranas are sometimes found to refer to some of the main teachings of the Buddha, but they do so invariably for denouncing Buddhism and not for advancing its cause. 31* See Danasagara, p. 7, verses 63-66. In the sixth line (asamgata-katha- etc.) the printed edition reads '-katha-bandha-

160 t STUDIES IN THE UPAPURANAS "[There is] another Tarksya (i.e. Garuda) Purana, as well as a Brahma and an Agneya [Purana], also a Vaisnava Purana of twenty-three thousand [Slokas], "[and] likewise another Lainga Purana consisting of six thousand [Slokas], [which deal] with initiation, consecration [of images etc.], usages of Pasandas, and testing of gems, "with false accounts of the families [of kings and sages], with lexicography, grammar etc., and with irrelevant statements having mutual contradiction. are "Finding people to be deceived [with these] by those [well-known] Minaketanas and others, who hypocrites, Pasandas and pseudo-ascetics, all [the abovementioned Puranas] have been rejected with contempt." It is hardly necessary to say that in the above-mentioned verses of the Danasagara on the Bhavisya and the spurious Puranas, the word 'pasanda' or 'pasandin' has been used by Vallalasena to mean the Tantriks, and not particularly the Buddhists; and this meaning of the word becomes evident from the facts that the Buddhists had practically nothing to do directly with any of these Puranas or with Durga-puja dealt with in them, and that the present Garuda- and Agni-purana (now available in printed forms), which, as Vallalasena's enumeration of the contents of the spurious Puranas shows, are undoubtedly the same as his second 'Tarksya' and 'Agneya' Puranas, prominently betray the influence of the Tantras (including particularly the Pancaratra Samhitas). As a matter of fact, on one occasion Vallalasena himself expressly takes the word 'pasandin' to mean 'those people who instruct such Dharma as is opposed to that of the Veda'317. We have already seen that according to Kumarilaparaspara-virodhatah' and this reading is supported by all the three Manuscripts, except in the case of '-virodhatah', for which the Indian Office Manuscript reads '-virodhitah'. But as, with this reading, the said line cannot be properly construed with the rest and does not give any cogent meaning, we have preferred to read it as '-katha-bandhaih paraspara-virodhibhih.' 317 See Danasagara, p. 57- visnudharme-pasandino vikarmasthan nalapec caiva nastikan / pasandino veda-viparita-dharm-opadestarah /

THE SAKTA UPAPURANAS 161 bhatta, the scriptures of the Pancaratras, Pasupatas and similar other sectaries were 'opposed to the triple Veda' (trayi-viparita). Following Kumarila, Medhatithi also names the Bhojakas 318, Pancaratras, Pasupatas and others as persons denying the authority of the Vedas and look upon their scriptures with equal disrespect 319. In Kurma-purana I.12.250-259 Devi has been made to declare Sruti and Smrti, and no other works, as the sources of Dharma consisting of sacrifices etc., and to decry 'the delusive (mohan- atmaka, mohanartha) literatures of the Kapalas, Bhairavas, Yamalas, Vamas, Arhatas, Kapilas, Pancaratras, Damaras, and others of these types' as 'opposed to Sruti and Smrti' (sruti-smrti-viruddhani) and meant 'for infatuating in their subsequent births those persons who mislead others by means of bad scriptures'; similarly, in Kurma-purana I.16.115-117 Siva and Kesava are said to have 'compiled the delusive scriptures (moha-sastrani) [such as] Kapila, Nakula, Vama, Bhairava, Purva-pascima, Pancaratra, Pasupata, and also others by thousands' for deluding those who had become extra-Vedic as a result of Gautama's curse 320; in Kurma-purana II. 37.146-7, the scriptures of the Vamas, Pasupatas, Somas, 318 They formed a class of Sun-worshippers.-See Volume I, pp. 95-99 319 See Medhatithi's commentary on Manu-smrti 2. 6. 320 The story of Gautama, as given in Kurma-purana I. 16. 95-122, runs as follows: In ancient times there was once an extremely severe drought, which continued for a very long time and compelled the sages living in Daruvana to approach Gautama for maintenance. Gautama supported them with food for twelve years, after which there was heavy rain and the earth regained its former state. As the difficulty was over, the sages met Gautama for permission to return to Daruvana for austerities, but the latter requested them to stay with him for some time more. Being eager to leave the place, the sages created a black cow by virtue of their magic power and presented it to Gautama. But as soon as Gautama touched this magic cow, it fell down dead. Thus associating Gautama with sin, the sages left for Daruvana. Very soon Gautama saw through the plan and cursed the sages to be extra-Vedic and guilty of heinous crimes (mahapatakin) and to be born as such again and again. Being in a difficult situation the sages praised Samkara and Visnu with 'popular hymns' (laukikaih stotraih) for immunity from sin; and as a result these two gods compiled for their deliverance the delusive scriptures of the Kapalas, Nakulas, Vamas, Bhairavas, Purva-pascimas, Pancaratras, Pasupatas, and 'thousands of others' and asked the sages to follow the prescriptions of these works and go to hell again and again, until, with the exhaustion of their sin, they would be born as human beings and attain the state of pious men by worshipping God. The sages did as desired by Samkara and Visnu, compiled other scriptures, taught them to their disciples, and deluded people by undergoing rebirths. 11

162 STUDIES IN THE UPAPURANAS Langalas, Bhairavas, and other extra-Vedic sects have been said to be delusive (mohana) and opposed to the Vedas (veda-vada-viruddhani); in Kurma-purana II. 16.15 one following the Vedic Dharma is urged not to show respect, even by a word (of the mouth), to the Vamacaras, Pancaratras and Pasupatas who are Pasandins (i.e. follow an anti-Vedic course of conduct) and always do acts not meant for them; and in Kurma-purana II. 21.32 the Buddhist saints, Nirgranthas, Pancaratras, Kapalikas, Pasupatas, and similar other sectaries have been called 'Pasandas'. Thus, Vallalasena's use of the word 'pasanda' or 'pasandin' and the fact that in spite of his deep hatred of the Pasandas he says nothing expressly about the language of the Devi-purana, show that this avowedly pro-Vedic royal Smarta of great eminence cannot be taken to mean that the Devi-purana was influenced in its language by the Buddhists. From what has been said above it is evident that the language of the Devi-purana follows a tradition which originated not later than that followed by the Buddhists in their hybrid texts and was quite distinct from the latter. As a matter of fact, this tradition comes down from a much earlier date, which finds strong support in the occurrence in this Purana of a good number of such Vedic and epic forms and usages as had gone out of use long ago, and it was followed in ancient times by the adherents to the Pancaratra, Pasupata and similar other religious systems as known and practised by those Hindus who had little regard for the Vedas and the varnasrama-dharma. As regards the compilation of the earliest pro-Vedic Pancaratra work the Mahabharata says in its Narayaniya section 321 that the seven illustrious Citrasikhandin Rsis 'unanimously proclaimed on the great mountain Meru an excellent Sastra which was made consistent with the four Vedas' 322 and meant for the popuConsisting of chapters 335-351 of the Santi-parvan. These chapters correspond to chapters 322-339 in the Poona critical edition. 322 Mahabharata XII. 335, 27 b-28 ye hi te rsayah khyatah sapta citra-sikhandinah // tair eka-matibhir bhutva yat proktam sastram uttamam / vedais caturbhih samitam krtam merau mahagirau //

THE SAKTA UPAPURANAS 163 lace, 323 that it consisted of one lac of verses 324 and dealt with the best 'loka-dharma' (religious duties of the people), 325 and that as it was to direct them in their activity as well as inactivity, it was furnished with (the teachings of) the Yajus-, Rk-, Sama and Atharva-veda. 326 Although no Pancaratra Sastra ascribed to the seven Citra-sikhandin Rsis has come down to us nor is there any definite evidence in favour of its actual existence, the very mention of making this work, which was meant for the populace, conform to the four Vedas, is important in that it implies the non-Vedic, and perhaps also in some respects anti-Vedic, character of the ideas and practices of the Pancaratra system as understood and followed by the common run of people. The Mahabharata itself differentiates the Pancaratra system from the Vedic, saying: 'samkhyam yogah pancaratram vedah pasupatam tatha / jnanany etani rajarse viddhi nana-matani vai //'9327 "Samkhya, Yoga, Pancaratra, Vedas and Pasupata- know these, O royal saint, as [branches of] knowledge holding different views." That before the compilation of the said pro-Vedic Sastra by the seven Citra-sikhandin Rsis the Pancaratras had little regard for the Vedas and the varnasrama-dharma, is also shown by the extant Pancaratra Samhitas of early The first two lines are the same as Poona critical ed. XII. 322. 26, the third being given there as spurious. a³ Mahabharata XII. 335. 32 b-lokan samcintya manasa tatah sastram pracakrire. (This line is the same as Poona critical ed. XII. 322. 29 c). 824 Ibid., XII. 335. 39-krtam sata-sahasram hi slokanam idam uttamam / lokatantrasya krtsnasya yasmad dharmah pravartate // (=Poona critical ed. XII. 322. 36). 325 Ibid., XII. 335. 29 a-asyaih saptabhir udgirnam loka-dharmam anuttamam. (In the Poona critical ed. this line has been given as spurious). See also verse 39 cited in the immediately preceding footnote. *** Ibid., XII. 335.40 pravrttau ca nivrttau ca yasmad etad bhavisyati / yajur-rk-samabhir justam atharvangirasais tatha // For this verse see Poona critical ed. XII. 322. 37. The Poona ed. reads 'yonih' (for 'yasmat'). 337 Mahabharata XII. 349. 64. (This verse, as occurring in Poona critical ed. XII. 337. 59, reads 'yogam' for 'yogah'). See also ibid., XII. 349. 1-samkhyam yogah pancaratram vedaranyakam eva ca/ jnanany etani brahmarse lokesu pracaranti ha // (=Poona critical ed. XII. 337. 1, which reads 'yogam panca-').

164 STUDIES IN THE UPAPURANAS origin which continue more or less the non-Vedic, and in some cases also anti-Vedic, tradition of the ancient Pancaratras. 328 About the Pasupatas also the Mahabharata expresses a very similar view. We have already seen how it distinguishes the Pasupata system from the Vedic. 329 At another place of this work Siva is found to say that in ancient times he created the Pasupata Vrata which 'was contradictory to, though in some cases agreeing with, the rules of varnasramadharma' and which was 'denounced by the unwise. '330 In his Smrti-candrika (II, p. 311) Devanabhatta quotes from the 'Brahmanda-purana' a verse which says: "saivan pasupatan sprstva lokayatika-nasitkan / vikarmasthan dvijan sudran savasa jalam aviset //" "[After] touching the Saivas, Pasupatas, Lokayatikas, Nastikas, those twice-born [people] who have taken up the duties not meant for them, and the Sudras, [one] should get into water with [all the] garments on." In the same work there is another verse, ascribed to the 'Sattrimsan-mata', in which, as we have already seen, the Pasupatas, Lokayatikas, Kapilas and Kapalikas have been denounced in a similar way. 331 These and other condemFor an idea of the non-Brahmanical character of the ideas, rites and practices of the Pancaratra Samhitas, see our analysis of the contents of the Jayakhya-samhita given in our Puranic Records, pp. 218-223. 329 See Mahabharata XII. 349. 64 quoted above. 330 Ibid., XII. 284. 121-124 avedat sad-angad uddhrtya samkhya-yogac ca yuktitah / apurvam sarvatobhadram visvatomukham avyayam / abdair dasardha-samyuktam gudham aprajna-ninditam // varnasrama-krtair dharmair viparitam kvacit samam | gatam tair adhyavasitam atyasramam idam vratam // maya pasupatam daksa subham utpaditam pura / These verses have been quoted by Apararka in his com. on Yajnavalkya-smrti 1.7 (pp. 17-18) with the readings 'saktitah' (for 'yuktitah'), 'arthair dasardhaih samyuktam' (for 'abdair ctc.'), 'satantaih' (for 'gatam taih'), 'anyasramam' (for 'atyasramam'), 'smrtam' (for 'vratam'), 'purvam' (for 'daksa'), and 'yogam' (for 'subham'). In the Poona critical ed. the above verses have been taken as spurious and placed in Appendix I (No. 28, lines 401-407) with the readings 'dasaha-' (for 'dasardha-'), 'gatantair' (for 'gatam tair'), and "yogam' (for 'subham'). Still, their early origin cannot be denied, and this is clearly indicated by their occurrence, with variation in readings, in the Vayu-purana (30. 293-5), Brahma-purana (40. 108-110), Linga-p (II. 20. 9-11), etc. 331 For this verse see above.

THE SAKTA UPAPURANAS 165 natory statements made in the comparatively early Sanskrit works about the Pasupatas show that among the early worshippers of Siva-Pasupati there was a large section which was no less irreverent of the Vedas and the varnasramadharma than the early Pancaratras. 332 It is natural, therefore, that these Pancaratras and Pasupatas and also similar other sectaries among the Hindus should have their scriptures written in an unorthodox language. Speaking of those Brahmin sages who, at the advice of Siva and Visnu, followed the teachings contained in the scriptures of the Vamas, Pasupatas, Somas, Langalas and others, the Kurma-purana says that after being cursed by Gautama to be extra-Vedic and guilty of Mahapatakas, these sages pleased the two gods 'with popular hymns' (laukikaih stotraih). At another place this Purana says that in the Kali age people would praise their deities 'with popular songs' (laukikair ganaih).333 It is highly probable that by 'laukika stotra' and 'laukika gana', the Kurma-purana means such hymns and songs as were composed in a popular language like that of the Devi-purana That the peculiar language of the Devi-purana was not confined to this work only, but was used in writing their scriptures by those other sects among the Hindus which were stamped out as 'extra-Vedic' by the staunch followers of the Vedas, is shown by the earliest extant Pancaratra Samhitas such as the Pauskara and the Jayakhya, more particularly the former, which was written, like the Jayakhya, in Northern India but comes from an earlier date. On the strength of doctrinal and palaeographical evidences Benoytosh Bhattacharya assigns the Jayakhya-samhita to about 450 A.D.334 So, the Pauskara-samhita is to be dated earlier. The high age of this work is further shown by the fact that it is named in the Satvata-samhita, 335 which, again, is mentioned in the 332 There is hardly any scope for doubt that the extra-Vedic ideas, rites and practices of these early Pasupatas come down more or less through the early Agamas. 333 Kurma-purana I. 29. 24--gayanti laukikair ganair daivatani naradhipa. Jayakhya-samhita, Foreword, pp. 26-34. 334 335 evam-adyais tu vidhivad-bhogair nanavidh-otthitaih / yah sthitas trividhe sarge vibhavah paramesvarah // pauskarakhye ca varahe prajapatye mahamate / suksmatvena ca nihsesam pratyekasmin hi artate // 132-133.

166 STUDIES IN THE UPAPURANAS Ahirbudhnya.336 In its printed edition this work consists of forty-two chapters covering as many as 272 pages. In order to show how strikingly its language agrees with that of the Devi-purana in the variety as well as frequency of grammatical and other irregularities caused mainly by Prakrit and Apabhramsa, we cite below a few instances from the first five chapters covering a little more than 31 pages only. I. Unlawful changes made occasionally in the stems: (i) Substitution of the final r (of masculine words) with a, as in 'savita' (for 'savitr') in Pauskara-samhita 3.69 savitakhyam ca savitram (ii) Loss of the final single consonants of masculine · and neuter words, as in 'sasi' (for 'sasin') in Pauskara-samhita 337 4.155- ...sasim ratrikaram smaret; 'adhva' (for 'adhvan') in 5.3-sad-adhvam syat trtiyam ca (in which 'sadadhvam' qualifies 'mandalam' in the immediately preceding verse); 'raja' (for 'rajas') in 1.6- ... purayitva rajaih subhaih, 5.53-- vyoma bahyam rajena tu. For a few more instances see under III (i) below. ... (iii) Loss of the final syllables of words denoting cardinal numbers, as in 'vimsa' (for 'vimsati') in Pauskara-samhita 5.145-dvitiyenatha vimsadin ...; 'saptavimsa' (for 'saptavimsati') in 5.269- saptavimsa-padotthitam. ... (iv) Loss of the final vowels of words, as in 'pad' (for 'pada') in Pauskara-samhita 5.170-tulyam caturvidhakaram pad-bhedena yathasthitam; 'vimsat' (for 'vimsati') in 5.154 catur-abhyadhikan vimsad-amsakan 5.178 -satani panca-bhaganam vinsadbhis tu navadhikaih; 'ekonavimsat' (for 'ekonavimsati') in 5.118-bhagair ekonavimsadbhih . . . ; 'pancavirsat' (for 'pancavimsati') 936 samkalpa-nirmita saktir ya tat-tat-karya-gocara / satvate sasane sarvam tat tad uktam mahamune || 337 Pauskara-samhita Pauskara-samhita. - 5. 59. . . . .

THE SAKTA UPAPURANAS in 5.64-kosthakaih pancavimsadbhih kamalam pancavimsadbhih vimsadbhih .. " 167 5.119- . . . . 5.140-pankajam pancaII. Lengthening of short vowels (evidently for the sake of metre), as in Pauskara-samhita 5.152, 260-caturasitibhir bhagaih (for 'caturasitibhir ...'); 5.271-saptadasa-padad bhutam (for 'saptadasa- etc.). ... III. Samdhi made against the rules of Sanskrit grammar: (i) Double Samdhi (?), as in 'tatordhvam' (for 'tatah +urdhvam') in Pauskara-samhita 4.20-madhyamam viddhi tanmanam tatordhvam tu smrtam mahat; 'siropari" (for 'sirah+upari') in 4.83- tac-chiropari vinyaset; 'rajottame' (for 'rajah+uttame') in 5.201-rajasa purayet pascat vyoma-bahye rojattame // (ii) Irregular Samdhi, as in 'samekavarna' (for 'sama +ekavarna') in Pauskara-samhita 2.18-samekavarna sarpadyair nirmukta ... An instance of irregular Samdhi with r as an intevening Samdhi-consonant (or hiatus-bridger), viz., jati-r-inkulikabhena' (for 'jatinkulikabhena') in Pauskara-samhita 5.76-dasakam dasakam vipra bhaganam atha ranjayet / jati-r-inkulikabhena vyoma-bahyat tu purayet // IV. Hiatus created by absence of Samdhi in verses, as in Pauskara-samhita 1.30-tathaivayatananam ca asramanan mahamune, 2.17. hy apare amra-kananam, 5.66-dvaropagasya madhyat tu astau .... V. Wrong use of Number (i) in Dvandva compounds, as in 'udag-yamye' (for 'udag-yamyayoh') in 5.162, 'purva-pascimam' (for 'purvapascimau') in 3.4 and 3.16, 'daksinottaram' (for 'daksinottarau') in 3.38, 51, 52, 'pascimottare' (for 'pascimottarayoh') in 3.68;

168 5.47- STUDIES IN THE UPAPURANAS ... (ii) in verbs (irrespective of the Number of their subjects), as in Pauskara-samhita 5.35- . . . bhaved dvarany ath-asrayah, ekavimsottaram satam | utpadyante ca bhaganam //, 5.66 ... astau sobha-dvaye bhavet, 5.71 krtva dvidasadha ksetram sannavaty-adhikam satamutpadyante 'msakanam tu ... 11; - ... (iii) in adjectives, as in Pauskara-samhita 2.20-atyusna ya ca hemante rasanva (? rasala) jaladagame / kamsya-bhandasvana ghanta-vina-vamsa-ravah subhah || (in which 'ya' has been used for 'mahi' in the immediately preceding verse), 5.1-yagas tu deva-devesa bhadradini kramad vada, 5.84-ragais tu pitha-konani pascimadi tatah punah; (iv) in words denoting cardinal numbers in most cases, as in Pauskara-samhita 3.28-astavimsatibhir yukta-dvy-angulair dhanusi srnu/astadasangula vipra dvy-angulair dasabhir vina, 3.29-trimsadbhir dvy-angulair una 4.8 caturvimsatibhir brahman hastas tair angulair bhavet, 4.10-11-adya-dvadasabhir manam angulaih parikirtitam dvitiyam astadasabhis caturvimsatibhih param //, 5.64 kosthakaih pancavimsadbhih 5.118-bhagair ekonavimsadbhih 5.139-ekavimsatibhir bhagaih 5.152 caturasitibhir bhagaih | ... sattrimsadbhis tu kosthakaih //, 5.178-trayovimsatibhir bhagaih also 5.206, 208, 226, 241, 247 etc. (for similar use), 4.6-tabhir liksastakair yukam viddhi ... .... ... • ... • . . ." .... and (v) in other cases such as 'yagas tu bhradradini vada' (for 'yagams tu etc.') in Pauskara-samhita 5.1, 'saptadasapadad bhutam (for 'saptadasa-padebhyah etc.') in 5.271, '... . . catvarimsat-pade sthitam' (for 'catvarimsatpadesu etc.') in 5.280. VI. Unauthorised change of Gender: (i) Masculine words used as neuter, viz., 'adhara' in Pauskara-samhita 5.16 buddhy-adharam atah proktam .... 5.258 buddhy-adharam athocyate; 'akara' in 4.104 kartavyam jyesthakaram ganasya ca; 'amsa' in 5.186-

THE SAKTA UPAPURANAS saptavimsaty-athamsani dasamsani . . ., ' 169 5.196-samsodhy-aika- 5.232 dve catus-sat tathamsani 5.266-ekonavimsaty-amsani . . . . ...; 'bhaga' in 3.30- sankor dvadasa-bhagam yad vijnatavyam tad angulam / tasya yat sasthamam bhagam boddhavyam dvy-angulam tu tat //, 5.38-pitha-dik-sthita-bhagani . . . . 5.41 paritas tatra bhagani sodhayitva 5.81 ekaikadasa- 5.102 triny-ardham panca-bhagani .... bhagani 5.265, 268, 301- . . . . bhagani ...; 'danda' in 4.120- 4.204 vedi-toranasarvany adheya-dandani ... ... ... . . . . dandani ....; 'desa' in 3.18-tatha pascima-bhagastham desam (for '-bhagastho desah ...'), 5.61--tat-konamadhya-desani ranjayet; 'jaya' in 5.10-atah pancadasam yagam jayam nama jaya-pradam, 51.3-yam drstva bhavinam sasvad bhaved bhuta-jayam dvija, 5.151 idam uktam jayam nama ...; 'kona' in 5.28, 50, 249- .konani 5.39 tad-upasthana-konani... 5.68 pitha-konani sarvani ..., 5.70- ... tadvat konam . . ., . . ., ... iti smrtam, 5.77- pitha-konani ranjayet, 5.84 ragais tu pitha-konani 5.94 pithasya sarvakonani purayet, 5.95-konany athopapithasya tenaiva pariranjayet; 'pralaya' in 4.92- pralayani ca yojya ca; 'samcaya' in 1.5-evam-adi yada bhuyat karunam vakya-samcayam; 'stambha' in 4.63--stambhani ca suvrttani ...; 'sugriva' in 3.73-indras cendrapadakhyas ca sugrivam kusuma-dvija; 'vighna' in 3.49- madhye vighnam bhavet sphutam; 'yaga' in 2.3 pravartante ca yagani 5.5, 9, 10, 14, 97, etc. yagam (for 'yagah"), 5.18-asyaparam mahayagam dhruvakhyam ca prakirtitam; and so on. . . . . (ii) Neuter words used as masculine, viz., 'sthana' in Pauskara-samhita 5.45-dvara-kanth-opagau sthanau ; 'dvara' in 5.106-dvaradin ... susamsthitan, 5.203-dvaran nrpopalabhena sobham pitena purayet. (iii) Feminine words used as neuter, viz., 'asri" in Pauskara-samhita 5.78-nrpopalabhen-asrini sobhani vyoma-bahyavat:

170 STUDIES IN THE UPAPURANAS 5.114-sesac cho 'sobha' in 5.78 (cited above), 5.96-pitena copasobhani 5.106 sitena copasobhani bhopasobhani bhagair dvarani karayet. .... . . . > (iv) Difference in Gender between nouns and their adjectives, as in Pauskara-samhita 1.33-bhava utpadyate ksipram tan-madhye sraddhayanvitam, 4.78 ... kuryac chirsaganam mahat, 4.85 stambha-samtatim kimcit samucchritam dattva ... ... //, 4.97- ... kuryat prakaram ... mahat ||, 5.1-yagas tu devadevesa bhadradini kramad vada, 5.160-evam isana-dik proktam 5.202-jatya hinkulikabhena ca bhagams tu • .... . . ., 5.252-dve catus-sat (v) Difference in Gender between nouns and their pronouns, as in Pauskara-samhita 1.8-... tam tu viddhi caturvidham (in which 'tam' has been used for 'mandalam'), 2.14- tath-anye mrga-jatayah, 2.18-19-vapi-kupadikam khatam isane yasya cottare / subha-laksana-yukta sa sarva-siddhi-kari mahi, 4.6 tabhir liksastakair yukam viddhi ... VII. Wrong declension, as in 'prajapatim' (for 'prajapatih') in Pauskara-samhita 3.74; 'sasim' (for 'sasinam') in 4.155-brhac-chasaka-prsthastham sasim ratrikaram smaret; 'ekadasani" (for 'ekadasa') in 5.249... konany ekadasani tu; 'ekonavimsadbhih' (for 'ekonavimsatibhih') in 5.118; 'vimsadbhih' (for 'vim- satibhih') in 5.178 (cited above); 'pancavimsadbhih' (for 'pancavimsatibhih') in 5.64, 119, 140; 'ekacatvaribhih' (for 'ekacatvarinsadbhih') in 5.278- ... ekacatvaribhih padaih; 'tule' (for 'tulayam') in Pauskara-samhita 3.27-tule tu mesa-sadrsi vrscike dvadasangula. VIII. Incorrect use of case-endings: (i) The first case-ending used (a) for the second, as in Pauskara-samhita 1.51-anyas ca viddhy anaharo vela tasyendra vartate (for 'anyam ca

THE SAKTA UPAPURANAS ... 171 anaharam ...'), 4.77-bandhas ca sadrsam dadyat (for 'bandham ca ...'), 5.1-yagas tu devadevesa vada (for 'yagams tu etc.'), 5.39- paripurayet / dvarani casrayah (for ... casrih'), 5.62 ranjayet ... /dvarani ... vasrayah (for '... vasrih'); (b) for the seventh, as in Pauskara-samhita 5.161-tat tu pascima-dik kuryat (for '... pascima-disi ...'). ... (ii) The second case-ending used for the sixth, as in Pauskara-samhita 5.154... bhagam catustayam (for ... bhaganam .....'?). (for (iii) The third case-ending used for the sixth, as in Pauskara-samhita 5.131-gatrakabhyam samipam tu 'gatrakayoh ...'). (iv) The fifth case-ending used for the sixth, as in Pauskara-samhita 4.78 bhinnad upari ... (for ... bhinnasya upari ...'). ... (v) The sixth case-ending used (a) for the fourth, as in Pauskara-samhita 1.8-pujitam darsayet tasya (for ... tasmai ...'); (b) for the seventh, as in Pauskara-samhita 1.4-dayam kuru tv anathasya (for ... anathe ...'). ... (vi) The seventh case-ending used (a) for the third, as in Pauskara-samhita 4.206-mrda gomayamisrayam upalipya (for misraya upa- ...') '... ... (b) for the sixth, as in Pauskara-samhita 2.15-janayaty asu cahladam sva-kantasv iva darsanam (for ... sva-kantanam iva '), 3.18-tasyam abhyantarastham va ... (for 'tasyah abhya- ..'), 3.53-diksu caivantaralani (for 'disam caiva ...'). IX. Irregularities in the formation of compounds: (i) Occasional insertion of 'm' as a connecting link between the component parts, as in 'torana-m-manam'

172 STUDIES IN THE UPAPURANAS (for 'torana-manam') in Pauskara-samhita 4.125-etad dvadasamamsena torana-m-manam abjaja, and 'pankti-dvaya-m-jatam' (for 'pankti-dvaya-jatam') in 5.101-bhagam pankti-dvayam-jatam samcarartham tu sodhayet. (ii) Occasional insertion of the indeclinable 'atha' between the component parts, as in 'saptavimsaty-ath- amsani" (for 'saptavimsaty-amsan') in Pauskara-samhita 5.186-saptavimsaty-ath-amsani kona-desad visodhayet. (iii) Wrong use of Samasantas, as in 'sad-adhvam' (for 'sad-adhva', a Bahuvrihi compound qualifying 'mandalam') in Pauskara-samhita 5.3-sad-adhvam syat trtiyam ca sad-adhvam adhunocyate, 5.40-sad- .... 5.33- adhvam ca maya proktam .. (iv) Various other irregularities, as in 'diva-ratram' (a Dvandva compound meaning 'a day and a night', for 'diva-ratri') in Pauskara-samhita 3.15-yasmin dese diva-ratram nityam eva samam dvija, and 3.16 samam yatra diva-ratram ekasmin vaisuve dine, 'purvapascimam' (for 'purva-pascimau') in 3.4- 3.4- yatnatah purva-pascimam / vijneyam 11, and 3.16-tatranena visesena jnatavyam purva-pascimam, 'purva-pascimat' (for 'purva-pascimabhyam') in 3.35, 'daksinottaram' (for 'daksinottarau') in 3.38, 3.51, 3.52, 'pascimottare' (for 'pascimottarayoh') in 3.68- daksine pascimottare / istv-aivam kona-padmesu ... ... ... 11, 5.145- 'coratanka-bhayadike' (for '-bhayadikaih' connected with '-vinirmukte') in Pauskara-samhita 2.7-dusta-prani-vinirmukte coratanka-bhayadike, 'sa-sthulaih' (for 'sa-sthaulyaih') in 4.76-drdha-kastha-samudbhutaih sa-sthulair bahubhis tatha, vinsadin' 'vinsadin' (for 'vimsaty-adin') in dvitiyenatha vimsadin satkam Satkam tad-asrisu, 'saptavimsa-padotthitam' (for 'saptavimsati-padotthitam') in Pauskara-samhita 5.269, 'pad-bhedena' (for 'pada-bhedena') in 5.170-tylyam caturvidhakaram pad-bhedena yathasthitam,

THE SAKTA UPAPURANAS 173 'vimsad-arsakan' (for 'vimsaty-amsakan') in Pauskara-samhita 5.154 catur-abhyadhikan vimsad-amsakan kamalodbhava, 'ekacatvaribhih' (for 'eka-catvarimsadbhih') in 5.278-satastakam tu bhaganam eka-catvaribhih padaih. X. Wrong formation and use of verbs: (i) Parasmaipadiya roots used in their Atmanepadiya forms, viz., bhvadi-ganiya 'stha' (meaning 'to remain') in Pauskara-samhita 1.50-yathaikas tisthate trptah (for " tisthati '), 3.92-samyuktam tisthate dehe .. (for '... tisthati ...'). • (ii) Atmanepadiya roots used in their Parasmaipadiya forms, viz., bhvadi-ganiya 'mud' in Pauskara-samhita 2.15- yatra modanti vai vyaghra-kunjarah saha manavaih (for '... modante ...'). (iii) Irregular conjugation of verbs (mostly with unauthorised change of Gana): (a) bhvadi-ganiya 'iks' (to see) in Pauskara-samhita 3.43 (with 'nir')tatha chaya niriksayet (for .. nirikseta'); (b) adadi-ganiya 'mrj' in Pauskara-samhita 5.216, 246, 252 mrjet (for 'mrjyat'); (c) hvadi-ganiya 'hu' in Pauskara-samhita 4.206-hunet (for 'juhuyat'); (d) divadi-ganiya 'as' in Pauskara-samhita 2.90, 4.60, 4.67, 4.74, etc. (with 'ni')-nyaset (for 'nyasyet'), 4.83, 4.127 (with 'vi+ni') vinyaset (for 'vinyasyet'); (e) svadi-ganiya 'ci' in Pauskara-samhita 4.56 (with 'vi')-vicinviyat (for 'vicinuyat'), 4.105-cinviyat (for 'cinuyat'). (iv) Wrong formation of Krd-anta words, viz., (a) Gerunds (in 'ktva' and 'lyap'), such as 'samjnatva' (for 'samjnaya') in Pauskara-samhita 1.42-krta-krtyam tu samjnatva 'samiksayitva' (for 'samiksya') in 4.60 samiksyitva tan sarvan ..., 'unmilayitva' (for 'unmilya') in 5.271, • • > 'lanchya' (for 'lanchitva') in Pauskara-samhita 3.56- ... lanchya

174 STUDIES IN THE UPAPURANAS purvokta-lanchanaih, 'lopya' (for 'lopayitva') in 5.246, 'pujya' (for 'pujayitva') in 2.27, 'purya' (for 'purayitva') in 3.79 and 82, 'sodhya' (for 'sodhayitva') in 5.250, 'yojya' (for 'yojayitva') in 4.92. (b) Other primary bases, such as 'aruruksakah' (for 'aruruksuh') in Pauskara-samhita 1.49-siddhas ca sidhyamanas ca trtiyas caruruksakah. (v) Wrong use of Gerunds, as in Pauskara-samhita 3.49-sankusutram dvidha bhanktva madhye vighnam bhavet sphutam, 5.63-sarva-kama-pradam caiva krtva 'mitraghnam ucyate / ksetram tridasadha krtva bhaganam jayate satam //, 5.100-bhavanti sobha-tulyani krtv-aivam gatrakani tu, 5.107-8-krtv-astadasadha ksetram | jayate vai caturvimsaty-adhikam krtva tu jayate, 5.128-ksetram dvadasadha krtva bhavec chatacatustayam, 5.151-2-dvir ekadasadha ksetram krtva samjayate dvija / caturasitibhir bhagair adhikam tu catuhsatam //, 5.155-evam amsaka-bimbam ca krtva dvaropamam bhavet, 5.271-sammarjy-onmilayitva ca bimbam tu prathamam bhavet, and so on. //, 5.118- (vi) Unauthorised use of Asir-lin (Benedictive Mood) for Vidhi-lin, as in Pauskara-samhita 1.5-evam-adi yadi bhuyat karunam vakya-samcayam / pravarteta tad-artham tu gurur mandala-pujane // XI. Wrong formation of Taddhitanta words such as 'dvadasama' (for 'dvadasa') in Pauskara-samhita 4.125-etad dvadasamamsena torana-m-manam abjaja, 'sasthama' (for 'sastha') in 3.30-tasya yat sasthamam bhagam boddhavyam dvy-angulam tu tat, 'traih' (for 'trih' formed with the Taddhita suffix 'suc', i.e. 's' added to the word 'tri') in 4.92-bhittau bhittau catus trair va pralayani ca yojya ca, 'vistrtam' (formed with the Taddhita suffix 'tal' i.e. 'ta' added to the wrongly formed word 'vistr' ?) in 4.4- astrader bhumivistrtam (for ... -vistrtim³). ...

THE SAKTA UPAPURANAS XII. Use of an adjective as a noun, as in word 175 sa-sthulair 'sthula' in Pauskara-samhita 4.76 drdha-kastha-samudbhutaih bahubhis tatha (in which the word 'sa-sthulaih' is an adjective used as a noun for 'sthaulya'). XIII. Use of a noun as an adjective in Pauskara-samhita 5.154- panktya bhagam catustayam(?). XIV. Use of synonymous words, as in Pauskara-samhita 1.33-tadiyena prabhavena sraddh-adyadyesu kim punah, 2.7 ... cor-atanka-bhay-adike, 3.73 yavad-vayu-pad-avadhi. It is needless to cite more instances. The above citations will be found sufficient to indicate the nature of the language of the Pauskara-samhita and to show its striking similarity with the language of the Devi-purana It should be mentioned here that our citation of instances of grammatical and other irregularities from the first five chapters of the Pauskarasamhita is not exhaustive. The above collection has been made simply to give a general idea of the variety as well as frequency of the irregularities found in this work. Like the Pauskara-samhita the Jayakhya also is looked upon as one of the earliest and most authoritative works of the Pancaratra Agama. But it comes from a date later than that of the Pauskara and consequently lacks many of the linguistic peculiarities of the latter, evidently under the growing influence of Sanskrit. Still it retains a good number of such irregular forms and usages as amply testify to the original nature of the language of the Pancaratra scriptures. In order to give an idea of these irregularities, a few striking instances are noted below mostly from the first sixteen Patalas of the Jayakhya (which cover about half of the work consisting of thirtythree Patalas). (1) The Jayakhya-samhita uses the word 'tredha' (for classical 'tridha') in 16.176 ekadha va dvidha tredha and spells 'trtiya' once as 'tritiya' in 15.184-tritiyam • "

176 STUDIES IN THE UPAPURANAS " mula-khandanam and 'simha', 'nrsimha' and 'narasimha, very often as 'sihma', 'nrsihma' and 'narasihma' respectively. (See, for instance, Jayakhya-samhita 16.21, 37, 107 etc., 25.16, 27, 31, 38, 51, 54, 62 etc., 16.91, 25.42, 72, 86, 95 etc.; 29.5, 9 etc.; 2.12, 15, 143, 26.107, and so on). Sometimes it shortens the final long vowels of words, as in 'devata' (for 'devata') in 2.57--tada sambodhito brahma devatair munibhih saha, and sometimes it lengthens the short ones, as in 'sakti' (for 'sakti') in 14.66 -tasya sabdamayi sakti jvalavan nihsrtam mahat, and 'kusa' (for 'kusa') in 15.115-madhya-namram kusakandam grhitva ... ekaSometimes it drops the final single consonants of words, as in 'ekatrimsa' (for 'ekatrimsat') in 13.194 trimsaksaram param / mantram ||, 'sira' (for 'siras') in 16.128- sisyam sama-pada-sironnatam, and 'sarpi' (for 'sarpis') in sarpibhyam snanam ... ... 15.146-hiranya-madhu- (2) As regards Gender, the Jayakhya-samhita sometimes goes against the rules of Sanskrit grammar. Thus, it uses neuter words as masculine, as in 15.143 narasihmadayo vaktrah sphurat-taraka-samnibhah, and 16.139- ... sa hi karm-aspado yatah, and it allows nouns and their adjectives or pronouns to have different Genders, as in 14.66-tasya sabdamayi sakti jvalavan nihsrtam mahat (in which the words 'nihsrtam' and 'mahat', though qualifying 'sakti', i.e. 'saktih', have been used in the neuter Gender, for 'nihsrta' and 'mahati'), and 16.141-sadhu parna-pute sutram krtva tam mallake 'tha va (for ... sutram ' ... tat mallake " 1. (3) It contains instances of irregular formation of feminine bases, as in 2.66-sampravistas tatas caham imam padmodbhavim tanum (for padmodbhavam C 5-

THE SAKTA UPAPURANAS 177 (4) Samdhi has sometimes been effected irregularly either with the elimination of the final single consonant of a word followed immediately by the initial vowel of another word, as in 'payupastheti' (for 'payupastham+ iti') in 3.6-tac ca vak-pani-padakhyam payupastheti ca smrtam, and in '-sironnatam' (for '-sirah+unnatam³) in 16.128... sisyam sama-pada-sironnatam, or with the insertion of a Samdhi-consonant (or hiatusbridger), such as 'r', as in 'tri-r-angula' (for 'tri+angula') in 15.103-vistarasya tri-bhagena griva sasra tri-rangula. (5) In using Vibhaktis the Jayakhya-samhita sometimes violates the rules of Sanskrit grammar. Thus, it employs (a) the fourth case-ending in place of the sixth (in the sense of 'sambandha'), as in 2.3-svam atmanam puraskrtya sad-bhaktaya hitaya ca (for '... sad-bhaktasya hitaya ...'), (b) the sixth case-ending in place of the fourth and the seventh, as in 2.64-dattam mayabhayam tesam .. (for ... tebhyah tebhyah ...'), and 12.132 kramoditena vidhina tasya tusyamy aham mune (for ... tasmin tusyamy ...'), (c) the seventh case-ending (instead of the sixth) in connection with the word 'upari' (formed with an 'atas-artha pratyaya', viz, 'ril', i.e. 'ri'), as in 11.4 ksitav upari vinyasam | tasmims copari vinyasam //, 12.128... tasyam upari bhavayet, and 15.189 tatr-opari ghrtam dadyat and so on. ... . . . . (6) Its irregularity in the formation of compounds is found in the words 'ekatrimsaksaram' (a Bahuvrihi compound, for 'ekatrimsad-aksaram') in 13.194 (cited above), 'tri-r-angula' (for 'try-angula') in 15.103, 'susuptavasthagah' (for 'susupty-avastha-gah') in 16.175, 'payupastha' (for 'payupastham') in 3.6 (cited above), 'kusa-kandam' (for 'kusa-kandam') in 15.115, 'udak-pascima-aisana-pran- 12

178 STUDIES IN THE UPAPURANAS mukham' (with hiatus, for '-pascimaisana-') in 16.192, and so on. (7) On several occasions it uses Parasmaipadiya roots in their Atmanepadiya forms, and vice versa. Thus, it has 'samsmarasva' (for 'samsmara'-Lot, second person, singular) in 2.58, 'acarate' (for 'acarati') in 12.131, 'labhet' (for 'labheta') in 13.45, 'samarabhet' (for 'samarabheta') in 15.161, and so on. It has irregular conjugation of verbs in 'hanet' (for 'hanyat' Vidhi-lin, third person, singular of the root 'han') in 10.4, 'hunet' and 'juhet' (for 'juhuyat') in 15.183 (also 16.160) and 16.176 respectively, and so on. (8) Among the irregularly formed Krd-anta verbs there are (a) the Gerunds, such as 'likhya' (for 'likhitva') in 6.27, 'prerya' (for 'prerayitva') in 16.138, 'pujya' (for 'pujayitva') in 6.7, 'tadya' (for 'tadayitva') in 13.86, and so on, and (b) the Present Participles, such as 'anicchamana' (for 'anicchat') in 16.138-anicchamanasya balaj jantoh 'bhramayamana' (for 'bhramayat') in 13.85-smrtva bhramayamanas tam samvised yaga-mandiram, 'dhyayamana' (for 'dhyayat') in 16.95- dhyayamanas . . . . tam eva hi, 'patamana' (for 'patat') in 12.122-smrtva mantram tu tan-murdhni patamanam dvijambarat, and so on. The Satvata-samhita, which, with the Pauskara and the Jayakhya, formed the three jewels of the Pancaratra scripture, is undoubtedly an early work like the other two. It has been drawn upon by Ramanuja in his Sribhasya 338, and named and referred to in the Ahirbudhnya-samhita³39 338 See Sribhasya on Vedanta-sutra, II. 2. 42 (p. 1608)- .satvata-samhitayam uktam- brahmananam hi sad-brahma vasudevakhya-yajinam / vivekadam param sastram brahmopanisadam mahat // iti. This verse is the same as Satvata-samhita 2. 4 a and 5 a. 339 For the relevant verse of the Ahirbudhnya-samhita see footnote 336 above. 13

THE SAKTA UPAPURANAS 179 and the Bhagavata-purana 340. But its mention of the names of the Pauskara, Varaha and Prajapatya (i.e. Brahma) Samhita in one of its verses 341 with references to some of their topics, shows that it is to be dated later than all these three works. So, we can scarcely expect to find in it grammatical and other irregularities as much as in the Pauskara. Still it is not totally free from these, but there is hardly any necessity of noting them here. Although we have got very little information about the early Saiva Agamas and the Sakta Tantras, the language of the mystic Vidyas, Kavacas, Mantras, etc. contained in the extant Saiva and Sakta works shows that the early Agamas and Tantras were written in an unpolished language very similar to that of the Devi-purana and the Pauskara-samhita. In connection with imparting religious instructions to kings and others, the Siva-dharmottara says: "raga-dves-anrta-krodha-kama-trsnanusari yat / vakyam niraya-hetutvat tad durbhasitam ucyate // samskrten-api kim tena mrduna lalitena ca / avidya-raga-vakyena samsara-klesa-hetuna // yac chrutva jayate punyam ragadinam ca samksayah | virupam api tad vakyam vijneyam atisobhanam //" "Whatever utterance follows from love, hatred, falsehood, anger, desire and longing, is called bad speech, because of its being the cause of [residence in] hell. "What [can one have to do] with the words of attachment arising from nescience, which cause sufferings of rebirths, even though these [words] may be polished, gentle and charming ? "That speech, though deformed, is to be known as highly elegant, which, being heard, creates merit and destroys attachment etc." 340 I. 3. 8-trtiyam rsi-sargam vai devarsitvam upetya sah / tantram satvatam acasta naiskarmyam karmanam yatah // This verse unmistakably means our present Satvata-samhita, in which Narada appears as the speaker. 341 For the relevant verse of the Satvata-samhita see footnote 335 above. 142 Chapter 2, fol. 42 a.

180 STUDIES IN THE UPAPURANAS The Siva-dharmottara allows to the Saiva teachers sufficient freedom in the use of language for religious instruction, saying: "samskrtaih prakrtair vakyair yah sisyam anurupatah / desa-bhasady-upayais ca bodhayet sa guruh smrtah //"343 "That person is regarded as a [religious] teacher who enlightens [his] disciple with refined and vulgar speeches (or, with expressions in Sanskrit and Prakrit) and by means of local dialects etc. according to [the latter's] fitness." It is to be noted that unlike Kumarila-bhatta, who refused to accept as Sastras those works which were composed in incorrect words (asadhu sabda) derived mostly from the Prakrit dialects such as 'Magadha, Daksinatya and their Apabhramsas', the Siva-dharmottara has a very liberal view as regards the medium of religious instruction and prefers the 'deformed" (virupa) speech to the refined one (samskrta) in case it rightly serves the purpose. The recognition of Sanskrit, Prakrit and the local dialects and also a mixture of all these as mediums of religious instruction, was likewise made by the Nandi-purana, an extinct 343 Chapter 2 (fol. 43 a). This verse has been quoted and ascribed to the 'Visnudharmottara' in Raghunandana's Vyavahara-tattva (for which see Smrti-tattva, II, p. 205) and to the 'Visnudharma' in Mitramisra's Viramitrodaya (Vyavahara-prakasa, p. 54). Both Raghunandana and Mitramisra take this verse as their authority for stating that even in teaching (adhyapana) there is no restriction about the use of correct words and that for the sake of proper understanding plaints etc. may be recorded in the courts of justice by using the popular words of the contending parties. (See Smrti-tattva, II, p. 205..... iti bhasa-sariram/ etac ca samskrta-desabhas-anyatarena yathabodham vaktavyam lekhyam va, murkhanam api vadi-prativadita-darsanat / ata evadhyapane 'pi tathoktam visnudharmottare-samskrtaih prakrtair vakyair yah sisyam anurupatah / desa-bhasady-upayais ca bodhayet sa guruh smrtah // Viramitrodaya, Vyavaharaprakasa, p. 54--bodhakais tu yesam ye sabdah prasiddhas tair eva bhasadi lekhaniyam / prativadi-sabhyadinam tad-anabhijnatve bhasa-dvayady-abhijnais te bodhaniya ityady uhyam / ata eva visnudharme 'dhyapane 'pi sadhu-sabda-niyam-anadara uktah samskrtaih prakrtair vakyair yah sisyam anurupatah / desa-bhasady-upayais ca bodhayet sa guruh smrtah //' iti /). As this verse does not occur in the present Visnudharma and Visnudharmottara, its ascription to these works seems to be wrong. But it is not possible to dogmatise that this verse did not occur in the texts of these works known to Raghunandana and Mitramisra. 1 1 I

THE SAKTA UPAPURANAS 181 work composed not later than the seventh century A.D., and this is shown by two of the verses 344 on Vidya-dana quoted by Laksmidhara, Apararka, Vallalasena, Hemadri, Anantabhatta and others. 345 From what has been said above it is evident that the type of un-Paninian Sanskrit, found in the Devi-purana, Pauskarasamhita etc., had a very early origin, with which the Buddhists had nothing to do in any way. 345 a As a matter of fact, the popular religions current among the Hindus in early days were recorded in this type of Sanskrit, which scrupulously avoided Prakrit phonology and morphology to a very great extent but widely used Prakrit and Apabhramsa as its basis. As these religions concerned the common run of people who were indifferent to the Vedic views of religion, life and conduct and also perhaps to Sanskrit learning, Prakrit and Apabhramsa, which were widely used and understood by them, 346 were very naturally allowed to play a great part in the making of the Sanskrit of their religious works. But Sanskrit had to be retained, however 'deformed (virupa) 344 samskaraih samskrtam vidyam prakrtaih prakrtam api / alapa-matrair vyakhyanair yas ca sisyan prabodhayet // desabhidhana-vinyasair bodhayec capi yo guruh / sa guruh sa pita mata sa ca cintamanih smrtah // (For these verses of the 'Nandi-purana' see Laksmidhara's Krtya-kalpataru, V, p. 218, Apararka's com. on the Yajnavalkya-smrti, p. 400, Hemadri's Caturvarga-cintamani, I, p. 552, and so on). 345 For references to the relevant works of these Smrti-writers see Chapter III, footnote 76 below. 34³ In case we do not admit the early origin and distinct tradition of the language of the Devi-purana, Pauskara-samhita etc. and these works are taken to be of dates not going beyond the third century A.D., we fail to understand why these works should be composed in a synthetic language in imitation of Buddhist hybrid Sanskrit in spite of the fact that the Buddhists themselves had begun to look upon their own hybrid Sanskrit with growing apathy and to Sanskritise their hybrid texts from a much earlier date going as far back as the beginning of the Christian era. 34# This is shown by the traditional use of different kinds of Prakrit in Sanskrit dramas, as well as by the fact that the learned and extremely intelligent Vaihasika, who was the prototype of the Vidusaka and who was engaged by kings and harlots for the extremely difficult task of psychological study of people, spoke not in Sanskrit but in some popular dialect, in order to hide his own identity and learning. For more information about this Vaihasika and the Vidusaka of Sanskrit dramas see our article on "The Vidusaka in Sanskrit Dramas -His Origin" published in the Journal of the Asiatic Society, Calcutta, Volume XIX, No. 1 (Letters), 1953, pp. 85-103,

182 STUDIES IN THE UPAPURANAS it might have to be made for practical purposes, and, even in spite of the deep association of Prakrit with the daily life of the people, it could not be replaced by the latter, nor could it be allowed to change its appearance by taking the latter's phonology and morphology, because the persons for whom the said works were meant were Hindus and it was Sanskrit (be it Vedic, epic or otherwise) which had long been accepted by the Hindus as the language of their religious works. gramIt is true that the un-Paninian Sanskrit, as now found particularly in the Devi-purana and the Pauskara-samhita, has many peculiar features in common with the language of the two great epics of India, viz., the Ramayana and the Mahabharata. But we cannot overlook the fact that these two epics remarkably lack the variety and frequency of matical and other irregularities, very often made under the influence of Prakrit and Apabhramsa, which we find in the Devi-purana and the Pauskara-samhita. This dissimilarity, which is deep as well as extensive, is very striking and shows definitely that the language of the Devi-purana, Pauskara-samhita, and similar other works was not derived from the epics. We have already seen that this peculiar language originated long before the rise of the Buddhist hybrid texts. So, the assumption of derivation of this un-Paninian Sanskrit from the epics would create insurmountable difficulties in explaining away the variations referred to above. It appears that even in the pre-Paninian period a kind of grammatically incorrect Sanskrit (of which a later and probably more improved form is presented by the Devi-purana etc.) had come into being among the less cultured sections of the Vedic and allied population as a result of the contact of the Vedic language (including pre-Paninian Sanskrit), on the one hand, and the early Prakrits and their Apabhrasta forms current in the lower strata of the society, on the other, and that this synthetic language, which was evidently evolved for practical purposes, could not be set aside totally by the cultured Aryans because of their daily necessity of maintaining religious, social, economic, political or cultural contact 9115

THE SAKTA UPAPURANAS 183 with the other sections of the population. 347 But, as we have already said, Aryan supremacy in Indian society demanded that this synthetic language should be more Sanskrit than Prakrit, whatever its basis may be, and it is for this reason that in the Devi-purana we find the Sanskritised forms of such Prakrit words as were originally non-Sanskritic or were derived from Sanskrit, but the Vibhaktis and Pratyayas are very often Sanskrit. The present Ramayana and the Mahabharata, on the other hand, attempted at religious synthesis by bringing about a compromise between the popular religious views and those of the orthodox followers of the Vedas, who had great faith in the Vedic sacrificial religion as well as in the varnasrama-dharma. As they were meant for guiding the unorthodox members of the Hindu society in their religious, social and individual life in conformity 347 In pointing out, in his Mahabhasya (pp. 12 ff.), the various purposes served by the study of grammar, Patanjali says that it ensures perfect protection of the Vedas through knowledge of elision, augmentation, and change of letters, places the Brahmin priests of sacrifice in a position to make the right use of the Vedic Mantras by effecting necessary modifications in these, by acquiring knowledge of these, and by removing doubts as regards their meanings by correct accentuation, and offers a means of earning knowledge of words within a short period. Patanjali further says that by studying this science a Brahmin (who is the custodian of the Vedas and has to perform sacrifices) learns to avoid the employment of originally wrong (mleccha) and mispronounced (apabhasita) words (like 'helaya' for 'he arayah'), and also of such words as become defective due to the use of wrong accents or letters. He then adds: "yo hi sabdan janaty apasabdan apy asau janati / yathaiva hi sabda-jnane dharmah, evam apasabda-jnane 'py adharmah / atha va bhuyan adharmah prapnoti / bhuyamso 'pasabdah, alpiyamsah sabda iti / ekaikasya hi sabdasya bahavo 'pabhramsah / tad yatha gaur ity asya sabdasya gavi goni gota gopotalik-ety-adayo bahavo 'pabhramsah | atha yo 'vag-yoga-vid ajnanam tasya saranam //" "One, who knows the [correct] words, also knows the incorrect ones. Just as merit accrues with the knowledge of [correct] words, in the same way demerit also arises with the knowledge of incorrect words. Or, greater becomes the demerit, [because] the incorrect words are more numerous, [whereas] the [correct] words are comparatively small [in number]. For, each individual word has many wrong forms, just as the word 'gauh' (meaning 'cow') has many deteriorated forms like 'gavi', 'goni", 'gota', 'gopotalika, etc. Now, one who does not know the formation and use of correct words, has ignorance as his resort." From Patanjali's statement it is evident that even the orthodox Brahmin followers of the Vedas widely used in their everyday life such words as were originally wrong or became wrong due to wrong pronunciation or want of knowledge of grammar; and this, in spite of their high education, was possible because of their close contact with the masses who were very often indifferent to or ignorant of the rules of Sanskrit grammar. See also Mahabhasya, pp. 87 and 89.

184 STUDIES IN THE UPAPURANAS with the prescriptions of the Vedas, their language, though chastened to a great extent, could not eliminate the popular elements fully but followed a middle course lying between correct Sanskrit and the highly incorrect synthetic language of the pre-Paninian days, an idea of which we get mainly from the Devi-purana For want of early records of this synthetic Sanskrit it is not possible for us to say anything definitely about its constituents. It is highly probable that during the long period of its growth, the popular dialects of different parts of India made their contributions to its making. The association of this synthetic Sanskrit with the popular religious literatures of the Hindus was a matter of such a hoary past and was consequently so well established in the country that in spite of their having a synthetic language of their own, which we have named as hybrid Sanskrit, the Buddhists composed their Tantras (except the Arya-manjusri-mulakalpa³48) in a kind of Sanskrit which agrees much more with the language of the Devi-purana than with their own hybrid Sanskrit. In these works Prakrit phonology and morphology are remarkably wanting, except in the Gathas (which must have been derived from older sources), and many irregular forms of words, like those in the Devi-purana, have been used. For instance, the Prajna-paramitasadhana 349 of Asanga 350 has 'atra' (with the Taddhita suffix 'tral' in place of the seventh case-ending) for 'asya' in 'atrapy upari and 'vama-daksine cotpale' (for '-daksinayos cotpalayoh') on p. 324; the Guhya-samajatantra, which is one of the earliest Buddhist Tantras written not later than the seventh century A.D.351, shows signs of laxity about the use of gender (as in 'jatikam trividham' , 348 In this work we find a number of the leading features of Buddhist hybrid Sanskrit as well as of the language of the Devi-purana and the Pauskara-samhita. 34 Sadhana-mala, No. 159, pp. 321-5. 350 Asanga probabhy lived in the fourth century A.D.-See Winternitz, History of Indian Literature, II, pp. 355 ff.; also p. 392. Benoytosh Bhattacharya places Asarga about 270-350 A.D.-See Sadhana-mala, II, Introduction, pp. xl and xciv-xcv. 351 See Sadhana-mala, II, Introduction, pp. xxvii ff., xxxv; Winternitz, History of Indian Literature, II, pp. 394 ff.

.. • THE SAKTA UPAPURANAS 185 on p. 33), and has puja' (for 'puja') in 'puja-hetuna' on p. 33, 'cintya' (for 'cintayitva') on pp. 52, 53, 'grhya' (for 'grhitva') on p. 52, 'bhavitva' (for 'bhavayitva') on p. 35, 'prabhavitva' (for 'prabhavya') on p. 53, 'cintet' (for 'cintayet) on p. 40, and 'dasa', 'panca', 'sapta' etc. (for 'dasama', 'pancama', 'saptama' etc.) on p. 152; the Trisamayaraja-sadhana³5 of unknown authorship has 'ime vajra-padah (for 'imani vajrapadani') on p. 2, 'klesa-duhkha-jalan' (for '-jalani') and 'sarvasukhan' (for '-sukhani') on p. 4, 'maham' (for 'mahatim') on p. 4, 'dentu' (for 'dadatu') on p. 2, 'dadahi' 353 (for 'dehi') on pp. 4 and 7, 'vidhunahi' (for 'vidhunu') on p. 4, and 'nyaset' (for 'nyasyet') on pp. 4, 5 and 6; the Trailokyavasamkara-lokesvara-sadhanas³54 of Saraha have 'bali' used in the neuter Gender on pp. 80 and 83 (idam balim grhna in which 'bali', though taken as neuter, has been declined as masculine), and 'dhyayat'355 (for 'dhyayet') on p. 80 and 'grhna' (for 'grhana') on pp. 80 and 83; the Vajratarasadhana³56 of Dharmakaramati has 'devati'357 (for 'devata') in 'dasa-devati-parivrtam' on p. 195 and in 'devatinam' and 'madhya-puta-devatinam' on p. 199; the Visva-matasadhana 358 uses the inflected words 'mata' and 'bhuvi' as basic words (for 'matr' and 'bhu') in 'visva-mata-sadhanam' and 'visva-matayai namah' on p. 240, and 'sakala-bhuvi-tale' on the same page respectively; the Kalpokta-tarodbhavakurukulla-sadhana 359 has 'mayukha' (for 'mayukha') and 'nirmalina' (for 'nirmala') on p. 343 (tasya citra-mayukhabhih krtva nirmalinam jagat), 'trata' (for 'tratri') on p. 344 (ksantavyam tat tvaya devi yatas tratasi dehinam), 'sodhya' (for 'sodhayitva') on p. 343, 'dadanti' (for 'dadati') on p. 346, 'tranana' (for 'trana') on p. 346 (mamapi trananarthaya kha-vajr-adyam dadahi me), and the compounds 'dhupa- 35% Sadhana-mala, No. 1, pp. 1-14. 353 For the form 'dadahi' see also ibid., pp. 231, 346 etc. 364 Sadhana-mala, Nos. 35-36, pp. 79-83. **5 For 'dhyayat' see also ibid., pp. 198, 327, 345, 349, 352 etc. *5* Sadhana-mala, No. 97, pp. 195-200. 357 For 'devati' see also ibid., p. 240. 250 Sadhana-mala, No. 113, p. 240. *59 Ibid., No. 17, pp. 343-7.

186 STUDIES IN THE UPAPURANAS puspa-tatha-dipa-gandha-naivedya-samcayaih (with the indeclinable 'tatha' as an intervening member) on p. 343, 'mahad-bhuta" (for 'mahabhuta') on p. 346 and 'yathecchaya' (for 'yatheccham') on p. 346; the Tarodbhava-kurukullasadhana 360 of Muktaka (circa. 1100 A.D.) has the word 'upasama' used in the neuter gender on p. 349 (..... prathamamase rogady-upasamam bhavati) and conjugates 'labh' as 'labhet' (for 'labheta') on the same page; the Suklakurukulla-sadhana 361 has 'samidhani' (for 'samidhah') on p. 369, 'jata-matrena' (for 'jata-matram') on p. 364, and 'cakrankitam sveta-phaninam punar narair nihsankam dhriyate' (for 'cakrankitah svetaphani punar narair nihsankam dhriyate') on p. 371; the Sadbhuja-kurukulla-bhattarikasadhana 362 has 'nimantrayami' (for 'nimantraye') and 'carisye' (for 'carisyami³) on p. 379; and so on. For instances of metrical irregularities in the Sadhana-mala reference may be made to p. 196 (asta-bhujam kanaka-varnam caturbuddha-mahamukutim with more syllables than eight in a Pada), p. 265 (... dvibhujaikajata-sadhanam), p. 343 (... kurukulla-parvate gatam, in which the word 'kulla' of two syllables has to be pronounced as 'kul' of one syllable), and so on. It is to be noted that the grammatical and other irregularities in the Tantric texts mentioned above, are much less varied and frequent than in the Devi-purana and the Pauskara-samhita and that even the Gathas, derived from older sources, have often been Sanskritised in such a way that the number of Prakritisms in them has gone down in many cases. These variations and the fact that the Buddhist Tantras cannot possibly be dated earlier than the fourth century A.D. show that it was the linguistic tradition of the Samhitas, Agamas and Tantras of the Hindus which influenced the language of the Buddhist Tantras, and not vice versa. . . . . . . . . From what has been said above it appears that the prototype of the synthetic Sanskrit now found in the Devi-purana 360 Ibid., No. 172, pp. 347-356. 381 Ibid., No. 180, pp. 362-371. 382 Ibid., No. 182, pp. 379-380. " 3

THE SAKTA UPAPURANAS 187 and the Pauskara-samhita attained the position of the medium of expression in religious and social matters in quite early days, that due to its wide popularity the Ramayana and the Mahabharata, and undoubtedly also the early Puranas, 362 a aiming at religious synthesis as well as mass enlightenment, could not but imbibe many of the peculiarities of this language, and that in view of the influence of this synthetic Sanskrit as well as of Prakrits and Apabhramsa on all grades of people including even the Vedic priests, Panini and others tried to set up the standard of correct Sanskrit by writing grammatical treatises. It is hardly necessary to say that it was not merely Prakrits and Apabhramsa with their peculiar phonology and morphology but a kind of widely circulated and deep-rooted incorrect Sanskrit based on or influenced by Prakrits and Apabhramsa, which stimulated and necessitated the setting up of a standard of correct Sanskrit and the compilation of grammatical works. It is highly gratifying to see that specimens, though of comparatively late dates, of this synthetic Sanskrit have been preserved for us in the Devi-purana and the Pauskara-samhita. As regards hybrid Sanskrit as found in the Buddhist texts mentioned above, it may be said that, being fully alive to the advantages of the synthetic Sanskrit of the Hindus, the Buddhists also adopted it in some of their works, but, finding that due to the prevalence of Sanskrit words and 362 That once the un-Paninian forms in the Puranas were more numerous and varied than those now found in the printed texts and comparatively late manuscripts of these works (in which the interfering hands of the scribes and editors cannot escape our notice) will be evident from the interesting collection of these made from printed texts as well as manuscripts by Anand Swarup Gupta and published in Puranam, Volume IV, Part 2, 1962, pp. 277-297. It may be stated here that there is no evidence to prove that the Puranas, or any parts or individual verses thereof, were originally composed in Prakrit and later on rendered into Sanskrit by the Puranic redactors (as said by F.E. Pargiter in his Purana Text of the Dynasties of the Kali Age, pp. xvii-xviii and 77-83, and Ancient Indian Historical Tradition, pp. 5-14, and by Anand Swarup Gupta in Puranam, IV, Part 2, p. 280). Equally baseless is Sri Gupta's surmise that the Puranas were originally composed in 'laukika-samskrta' (by which he means classical Sanskrit or the 'Bhasa' of Panini) but were later influenced by the 'loka-bhasa Prakrit' (loka-bhasaya prakrtena prabhavitam) in the-course of their circulation by Vyasa, Suta and others. (See Puranam, IV, pp. 279-280.)

188 STUDIES IN THE UPAPURANAS forms it had receded much from actual life, they made it more practical by introducing into it Prakrit phonology and morphology in considerable measures in order to fulfil the wish of the Buddha that preachings of Buddhist doctrines should be made in popular dialects and according to a pronunciation which people can understand. If, in compliance with their Master's wish, the Buddhists made their preachings in popular dialects and did not adopt any kind of popular Sanskrit for the purpose, then we would find their religious texts composed in hybrid Prakrit 362 b and not in hybrid Sanskrit abounding in Sanskrit words and forms. It may be mentioned here that the intention underlying the use of hybrid Sanskrit in the compilation of some Buddhist religious texts was clearly and unquestionably the popularisation of the religious views of the Buddha, and that the Buddhists were always on the look out for possible means for occupying an advantageous position over the Hindus, who were their greatest rivals in the field of religion. So, their hybrid Sanskrit must have had, in its origin, some connection with contemporaneous popular life and could not be their own creation, because in the latter case a newly created language of this type might satisfy the personal whim or taste of its creators but would never find favour either with the literate Hindus for its Prakrit phonology and morphology or with the illiterate masses for its predominant Sanskrit elements. Thus, from the linguistic point of view the importance of the Devi-purana is immense; and it opens up a new line of investigation, which, if made with a more thorough and critical study of the relevant sources, is expected to solve many problems relating to our language, religion and culture. As we have already indicated, the whole of the Devi-purana 362 b The prevalence of Prakrit phonology and morphology in the hybrid Sanskrit of the Buddhists shows that the authors of the hybrid texts were quite familiar with some Prakrits and Apabhramsa which were quite different from Pali, and which, as Asoka and Kharavela's inscriptions, Patanjali's Mahabhasya, Bharata's Natya-sastra, Asvaghosa, Bhasa and others' dramas, and similar other records of carly date amply indicate, were very closely associated with popular life in those days. 2 "

THE SAKTA UPAPURANAS 189 is not metrical. There are many prose passages in some of its chapters, such as chapters 9, 17, 33, 36, 77 and 116, and especially chapters 55-56 and 109-110. The frequent occurrence of grammatical irregularities in these prose parts points to the general integrity of the work. Tracts, claiming to be parts of the Devi-purana, are very small in number. In the Vangiya Sahitya Parisat collection there are two Manuscripts of the 'Devipuranokta-durgapujapaddhati' and 'Devipurane Virastami-vrata-katha' 363, and in the Dacca University collection there is a Bengali Manuscript of a Candika-khanda 364 claiming to be a part of the Devi-purana This Candika-khanda, which is quite different from the *Candika-khanda' drawn upon in the Prayascitta-khanda (p. 82) claiming to be a part of Hemadri's Caturvargacintamani, is a unique Puranic work dealing, in sixteen chapters, with the story of Mangala-candi as found in the Candi-mangala-kavyas of Bengal. The Bengali script of its Manuscript and its story and other contents show that it was written in Bengal at a comparatively late date. Before we finish our study of the Devi-purana we should say something about its contents as known from the references to and quotations from this work made in the Nibandhas. In course of our attempt at tracing the verses ascribed to it by the Smrti-writers of different parts of India we have found that hundreds of these verses do not occur in our present text of this Purana. Although some of these untraceable verses are undoubtedly local additions, they deal, often in details, with various topics mostly of religious interest, as will be evident from the following list of their contents. I. Worship of Devi (Durga) in images of different articles such as gold, silver, earth, etc., or in a sword or a trident 365. *** Chintaharan Chakravarti, p. 72. *** Manuscript No. 1617 A (consisting of 78 folios and containing a fly leaf bearing the date 1723 Saka). 365 Durga-bhakti-tarangini, p. 7.

190 STUDIES IN THE UPAPURANAS Characteristic features of an image of Devi having eight, ten or twenty hands 366. Names of leaves and flowers to be used in Deviworship 37. Praise of offer of Bilva-leaves to Devi³68 Proper time and method of Devi-bodhana, Bilvabhimantrana, Patrika-pravesa, Devi-pratima-sthapana, Devi-pratima-sthapana, Devi-visarjana, etc. 369 Bathing of Devi with ghee, milk, curd, honey etc. and with the waters into which gold, gems, camphor and Aguru have been put 370. Performance of Devi's ghrtabhiseka (holy bath with ghee) in the month of Asvina 371. Deviworship on the Astami Tithi connected with Navami 372. Lustration of Devi (devi-nirajana) 373. Names of sixteen Devis; performance of Mudras; and offer of weapons 374. Worship of Matrs³75. Worship of virgin girls (kumari-puja) with flowers etc.; feasts to be given to them on Suklastami, Navami and Caturdasi Tithis; and the virgin girls' placing of aksata (thrashed and winnowed rice dried in the sun) on the head of the giver of the feast 376. Worship of horses during Devi-worship³77. Characteristics of persons fit for consecrating an image of Devi 378 *** Ibid., pp. 14, 93 (12 lines). 367 Ibid., p. 23. Also Varsa-kaumudi, p. 170, Nityacara-pradipa, pp. 640-641, Acara-mayukha, p. 85. 368 Durga-bhakti-tarangini, p. 155. 389 Sulapani's Durgotsava-viveka, pp. 4, 7, 8, 11 etc.; Krtya-tattvarnava, folios 60 a-b, 62 b, 64 b; Varsa-kaumudi, p. 346; Sraddha-kaumudi, p. 248; Durga-puja-tattva, pp. 3, 44; Kalasara, pp. 88, 103, 105, 112 etc.; and so on. 370 Durga-bhakti-tarangini, pp. 142-7 (37 lines). 371 Krtya-ratnakara, p. 308. 372 Durga-bhakti-taramgini, p. 50. 373 Sulapani's Durgotsava-viveka, p. 25; Durga-puja-tattva, pp. 37-38; Smrtitattva, I, p. 102. 374 Durga-bhakti-tarangini, p. 71, Durga-puja-tattva, p. 25. 375 Durga-bhakti-tarangini, p. 91. 376 Ibid., pp. 100-101 (20 lines), Varsa-kaumudi, pp. 376, 447. 377 Caturvarga-cintamani, II. i, p. 906. 378 Apararka's com. (p. 16) on Yajnavalkya-smrti 1. 7, Nityacara-paddhati, pp. 509-510, Haribhakti-vilasa, p. 1180. According to the verses of the Devi-purana as quoted by Apararka, a consecrator of an image of Devi is required to be 'vama-daksina-vetta", 'matr-vedartha-paraga', 'pan-

THE SAKTA UPAPURANAS 191 Construction of images of Hara and Gauri according to Agama, and detailed description of the method of celebrating their marriage³79. Construction of an image of the Sun with his retinue 380. Performance of santi (pacificatory rite) in Asvina-krsnapancami for the good of horses 381. II. Performance of various Vratas, Pujas etc., such as Kunda-caturthi-vrata 382, Krsnastami-vrata 383, Nava-ratraSamkranti-vrata³86, Manasa-puja 387, and detailed description of the method of Mrtyunjaya-puja 388. vrata 384 Pusyarka-dvadasi-vrata³85, Elaborate treatment of 'dvari-bandha-vidhi'389. Statement of reasons for the name 'Aksaya Trtiya'390. Hoisting of flags on Bhadra-suklastami or Asvina- suklastami³91 Observances on the Caturthi Tithis of different months for the pleasure of Ganesa³2. Worship of different deities under different Naksatras (beginning with Asvini) for the fulfilment of different desires 393. Detailed description of the method of performing Laksa-homa 394. caratrartha-kusala', and 'matr-mantra-visarada'; according to those quoted in the Haribhakti-vilasa, he is to be 'pancaratrartha-kusala' and 'mantra-tantr-avadharaka'; and according to the verse quoted in the Nityacara-pradipa, he is to be 'pancaratrarthakusala" and 'matr-cakra-visarada'. 37 Krtya-ratnakara, pp. 108-110 (45 lines). 380 Apararka's com. (p. 570) on Yajnavalkya-smrti 1. 296 (the quotation consisting of 16 lines). 381 Kalasara, p. 80. 382 Krtya-kalpataru, VI, pp. 83-84; Caturvarga-cintamani, II. i, pp. 525-6. 283 Caturvarga-cintamani, II. i, pp. 823-6. **4 Tithi-nirnaya (of Candesvara), fol. 8 a (many lines). 385 Krtya-kalpataru, VI, p. 351. 386 Caturvarga-cintamani, II. ii, pp. 728-730 (34 lines). 387 Varsa-kaumudi, p. 437. 388 Caturvarga-cintamani, I, pp. 216-7 (14 lines). *** Ibid., I, pp. 1007-1014 (149 lines). 300 Krtya-ratnakara, p. 156. 3*1 Ibid., p. 285. *** Varsa-kaumudi, pp. 30-31 (20 lines). *** Caturvarga-cintamani, III. ii, pp. 878-9 (34 lines), 880-1 (25 lines). *** Vidhana-parijata, I, pp. 589-592 (51 lines).

192 STUDIES IN THE UPAPURANAS III. Gifts of artificial cows made of molasses, water and ghee in the months of Magha, Asadha and Bhadra 395. Gift of an image of Ganapati³96, of horses 397, of food, cows, milk, curd etc. 398, of shelter 399, and of dwelling houses 400. IV. Instructions on the feeding of Brahmins, offering of pindas to the deceased ancestors, and performance of Sraddha in holy places 401. Praise of the Ganges 402. Rubbing of the body in the Ganges with Nagakesara, camphor etc. 403 Offer of lamp and incense to her 404. Performance of Siva-worship, feeding of Brahmins, taking of food at night (nakta-bhojana), and similar other acts on the Caturdasi Tithi and Paurnamasi in Magha and Vaisakha on the bank of this river 405. V. Performance of Sraddha of persons meeting death in war or by throwing themselves into water or fire or from a high place 406 Prohibition of Magha-trayodasi-sraddha during Samkranti and eclipses 407. VI. Measurements of wells (with or without stairs) and tanks 408 395 Krtya-ratnakara, p. 513; Samaya-pradipa, folios 5 a, 36 a. 39* Dana-vakyavali, fol. 27 b. 397 Apararka's com. (p. 372) on Yajnavalkya-smrti 1. 210. *** Ibid., p. 386. 39 Caturvarga-cintamani, I, p. 668 (13 lines). 400 Ibid., I, pp. 673-4 (19 lines). 401 Pitr-bhakti, fol. 25 b, Sraddha-viveka (of Sulapani), p. 99, Ganga-vakyavali, pp. 129 and 245-6, Sraddha-viveka (of Rudradhara), pp. 14-15, Sraddha-cintamani, p. 130, Tirtha-cintamani, pp. 10, 11, Sraddha-kaumudi, pp. 268-9, Kalasara, p. 479, Smrti-tattva, I, pp. 324-5. 401 Ganga-vakyavali, p. 123. Also Calcutta Sanskrit College Manuscript, fol. 47 b, where there are three metrical lines which do not occur in the printed edition. 403 Ibid., Calcutta Sanskrit College Manuscript, folios 49 b-50 a (containing one verse not occurring in the printed edition). 404 Ibid., p. 205. 403 Tirtha-cintamani, pp. 247-250 (60 lines). 40€ Sraddha-cintamani, p. 23. 407 Krtya-cintamani, p. 25. *** Danda-viveka, p. 314.

THE SAKTA UPAPURANAS 193 Elaborate description of the methods of consecrating wells and gardens 409 and of planting trees 410. VII. A literate and highly placed Sudra's attainment of the sin of a Mahapataka by speaking immodestly 411. Results of a Brahmin or a similar person's dining in the house of a harlot or a similar woman, a Yati's intercourse with harlots etc., and a Siva-yogin's association with wine etc. 412 A king's enforcing of Siddhanta-sastras and Veda-sastras on his people¹¹³. Unfavourable results of giving no smoke to cow-pens for removing mosquitos and of milking a cow whose calf has died 414. Signs of a bull (viz., its exceptionally strong back and neck) capable of carrying a heavy burden 415. Punishment of those who hate the Siva-sastras or decry the Vedas; unassailable authority of Samkhya, Yoga, Pancaratra and Pasupata 416. Detailed description of the method of purifying the site for a house 417. Elaborate account of 'bhrgu-patana'418. Besides the untraceable verses with the contents mentioned above, there are numerous others which have been quoted individually or in small groups in the Smrtinibandhas. As they deal with miscellaneous topics of minor interest, they have been overlooked here. It is to be noted that all the verses referred to above were not known to and quoted by all the Nibandha-writers, but some of them 40º Caturvarga-cintamani, I, pp. 1003 (6 lines), 1004 (16 lines). 410 Ibid., I, pp. 1044-47 (54 lines). 411 Krtya-kalpataru, I, p. 194. 412-3 Danda-viveka, pp. 270, 271; Vivada-ratnakara, p. 631 (9 lines including the following two: tasmad raja samacaram dharmadharme vyavasthitah / siddhanta-veda- sastranam palanaya niyojayet //) 414 Krtya-kalpataru, III, p. 462. 415 Ibid., p. 197; Grhastha-ratnakara, p. 433. 416 Apararka's com. (pp. 10-11) on Yajnavalkya-smrti 1.7. 417 Smrti-tattva, II, pp. 415, 417, 419. 418 Caturvarga-cintamani, II. ii, pp. 964-9 (101 lines). 13

194 STUDIES IN THE UPAPURANAS were current in some particular part or parts of India, while others were known in other parts. This shows that the text of the Devi-purana was not the same everywhere but differed considerably in different provinces. It is true that some of these untraceable verses simply went under the name of the Devi-purana but did not form actual parts of its text, while some others, actually belonging to other Puranas, must have been ascribed wrongly to the Devi-purana We have already indicated that in spite of its Tantric elements the Devi-purana attained great prominence as an authoritative work. Its popularity became so much that some enthusiastic Saktas gave it out to be the original 'Bhagavata' Purana on account of its dealing with the glory, exploits, and method of worship of 'Bhagavati' (Durga). As a matter of fact, on p. 349 of his Krtya-ratnakara Cande- svara takes the word 'bhagavatan', occurring in a verse of the 'Bhavisya-purana', to mean 'bhagavati-bhaktan', and on pp. 357-8 he quotes Devi-purana, chap. 21, by naming the work as 'Devi-purana' at the beginning and as 'Bhagavata-purana' at the end and by giving the number of the chapter as forty-eight 419. It is hardly necessary to say that this claim failed to earn general support 420.

419 'bhagavata-purane devy-avatare navami-kriya-sucanam asta-catvarimso 'dhyayah'.-Krtya-ratnakara, p. 358.

420 Cf. the verse 'bhagavatyas ca durgayas caritam yatra vartate / tat tu bhagavatam proktam na tu devi-puranakam ||' occurring in chap. 23 of the Uttara-khanda of the Bengal Siva-purana and quoted by Saiva Nilakantha in his commentary on the Devi-bhagavata. (For further information on this verse see footnote 194 above).