This book is a reworking of an earlier work The Ganesa Purana. Volume One . Upasanakhanda, Otto Harrassowitz, Wiesbaden, 1995. I have revised the translation for this new edition, the Introduction has been substantially shortened and the notes to the translation have been severely truncated.

I have tried to make the translation less literal than in the previous volume and I have been able to correct some mistakes in the initial translation, especially following some good suggestions made by reviewers such as Ludo Rocher and John Brockington.

1. THE GANESA PURANA

The Ganesapurana (GnP.) is a Puranic text probably composed in the fourteenth or fifteenth century. Its contents derive partially from earlier literature dealing with Ganesa, but also draw on the body of myths found in the mahapuranas and upapuranas. It is divided into two books, called Upasanakhanda (Ukh.) and Kridakhanda (Krkh.), which are complementary with each other from the perspective of contents. Each book contains many myths and sections of text detailing rituals to be performed to Ganesa, though 'these are much less common in the Krkh. In addition, the forty-sixth chapter of the Ukh. contains a sahasranamastotra and chapters one hundred and thirty-nine to one hundred and fifty of the Krkh. contain the Ganesagita. The presence of both these plus the large body of myths makes this an almost complete devotional text for the worship and theology of Ganesa. Whether the text was developed as a literary expression of one of the Ganapatya sects, cannot be determined because of lack of any direct connection between it, and the other Purana dealing with Ganesa-the Mudgalapurana-, and the presence of a distinct sect of Ganesa worshippers located at a specific time and place probably in Maharashtra. Yet there must have been an audience for the text, a growing audience judging from the large number of manuscripts dating from about 1760 until the end of the nineteenth century. In addition, institutional support to fund the initial composition of the text must have come from one or more patrons, conceivably, though speculatively, the family of Moraya Gosavi in Cincwad.

One should not read the extant GnP. in order to find out about the early history .of the development of Ganesa as a god, I nor should it be read to discover the meaning of the myths about the early birth of Ganesa, his functioning as Parvati's house-guard and his clash with Siva, subsequent beheading and acquisition of the head of an elephant. 2 As for the first, Ganesa was already a very popular god by the time his Purana was composed and had many sacred places associated with him, though the founding of some of these is provided in the GnP. itself. For the second, the myths of his childhood as they are expressed in the mahapuranas are reinterpreted in the Ukh. and are directly present in the Krkh, but in an extremely muted manner, with the tension between Ganesa and Siva being underplayed, though still in evidence.

Rather the GnP. must be read as the product of a time when Ganesa had already assumed a status as one of the most popular Hindu gods over much of the sub-continent. The appearance of both it and the Mudgalapurana at similar periods suggests a need was felt by certain groups of Ganesa worshippers to establish Ganesa textually as a god of status not unlike Visnu and his more popular avataras, Siva, Hanuman or the goddess. I am not saying there occurred a competition between the worshippers of these gods- especially given the likelihood that many individuals worshipped more than one of them-but that it was considered a god like Ganesa should be represented in a text or texts where he was unambiguously the principal god. This would be a text not only drawing together any myths that might have been developed to expose the unique activities of this god within the larger framework of received mythology, it would also present his theology (through the vehicle of myth and stotra) and collect together as much material pertaining to puja as was deemed necessary. It is true that a series of Tantras dealing with Ganesa had been in existence for some time, but knowledge of them would surely have been restricted to a small circle of adepts.

Regrettably we have no direct evidence as to how it was used in a recitational sense, though I heard in October 1983 that recitations from it were being performed in a Ganesa temple in Pune. The sahasranama stotra has been recorded in compact disc form and vernacular versions of the text have existed since the early eighteenth century. I have easily been able to purchase several summaries of the text in Marathi and copies of the Tamil Vinayakar Purana can easily be purchased in temple book stalls. Despite this the GnP. seems not to have been widely known outside of a certain intellectual elite and even now, when all the signs suggest Ganesa is becoming a popular god all over India, the text remains largely unknown.

II. DATING AND PLACE OF COMPOSITION OF THE GANESA PURANA

Determination of the date of any Sanskrit text, except for the most recent is always difficult. Not only is it necessary to date the beginning and end of the compositional process, it is also necessary to date the stages corresponding to the development of new recensions if these arise. In the case of the GnP., the manuscript evidence I have looked at does not suggest the evolution of the text through various stages. Of course this applies only to the manuscript tradition. Any evidence pertaining to the presence of an oral tradition is completely lost to us. The manuscript evidence is consistent in presenting a very late date for the text and does not allow us to establish a stemma for the manuscripts of which there are a very large number.

That so many appear at a particular time (between 1700 and 1850) is in itself a significant development in the dispersal of the text and is one definite indication of the popularity of the god and the systematization of attempts to portray him as a popular god. The existence of both the GnP. and the MudP, with their huge body of myths, metaphysical component and substantial ritual element, reflect a felt need to codify all aspects of Ganesa worship and to universalize a god who must have already been universalized, if we can judge from the dispersion of iconographical evidence. What this means is that any scholarly study of the GnP. will be required to date the text and date the extensive appearance of manuscripts.

We cannot know the prior history of the development of the GnP. if it did in fact exist prior to its earliest manuscript evidence. Obviously, using the epics as a parallel example, a given text can exist either in oral or written form long before its earliest manuscripts but to prove this we need to have external evidence of the kind that includes testimonia. Only in the latest of the Nibandhas (1708) do we find evidence of the use of the GnP. as a source and in the earlier Nibandhas, sources for the Ganesacaturthi rituals are listed as several of the mahapuranas which deal extensively with Ganesa.

The second book of the GnP must be read as a complement to the first book, the Upasanakhanda. Where the first introduces the mythology of Ganesa by focussing on how a person becomes his devotee and the subsequent worship this requires, the second takes a more direct theological aim. It offers a comprehensive set of narratives dealing with the god's adventures in the world and how he cheats certain classes of people, saves others, reveals himself to yet a third category and provokes a fourth. These adventures are placed together and organized through the application of a number of thematic frames exercizing control over a diverse range of content, which despite its diversity, has a certain uniformity about it. Of these the avatara frame is one of the most important and easily recognized, and conveys a kind of a universal ambience to the god's activities. Its ubiquity in the Krkh means that the Ukh can be read as focusing on the individual in relation to Ganesa, whereas the Krkh focuses on the triple-world in relation to him. And to extend this further, the majority of myths In the Ukh are in part built around the motif of the consequences of an individual's failure to worship Ganesa first before any undertaking. This motif is far less common in the Krkh, without, however, being totally absent.

In the GN and Wai editions of the GnP, the second khanda is called Uttarakhanda, in contrast with the Purvakhanda, the alternative name for the Upasanakhanda. However, in the colophons the title haphazardly alternates between Krida and Uttarakhanda. Yet irrespective of whether the word lila or Krida is used, the reader/hearer is never allowed to lose sight of the games being depicted in the myths populating the Krkh. As both avatara and boy, Ganesa is depicted in game playing activities from the beginning to the end of the GnP, and even when the .austere Visvarupa forms of the god are described at so many spots in the text, the idea of play is present and, like the BhagP, the Krkh can he read as a meditation on lila.

I. Literary Organization of KRKH

Like the Ukh the Krkh can easily be divided into individual myths and larger Narrative Units. The Ganesagita (138-148), for example, stands as a separate generic unit by itself, even though most Gitas do appear within larger textual units than themselves. An overall organization of the contents of this khanda is given, typically, as early as the first chapter and it is followed relatively consistently and the reader/hearer is reminded time and again of its relevance.

In period after period Ganesa has different names, different vehicles, different deeds, different qualities and he destroys different demons. In the Krtayuga he is mounted on a lion, he is ten-armed and named Vinayaka. His body looks splendid, it is huge, he gives gifts to everyone and he is independent. In the Tretayuga he is mounted on a peacock and he has six arms, a white skin and he is famous in the triple-world under the name of Mayuresvara. In the Dvaparayuga he is coloured red, mounted on a rat and four-armed. He is called Gajanana and gods and humans worship him. But in the Kaliyuga he is smoke-coloured, mounted on a horse and has two hands. He is called Dhumraketu and destroys the barbarian armies. He has killed many demons and I am now going to tell you about that, sage.

This is only a rough guide to the reading of the contents. The sections of the text which correspond to the four yuga frame are the following: Krta (2,1,23-2,72,36), Treta (2, 73, 3- 2,126,63), Dvapara (2,127,7 - 2,137,41) and Kali (2, 149, 1541) ; and the respective demonic figures are Devantaka and Narantaka, Sindhu and Sindura, but with no such figure occurring in the Kaliyuga. Chapters 138-148 comprise the GG and from 149-155 all of the major interlocutory frames are completed and the conditions under which the Purana was first heard are given.

Such an organization of events is not uncommon, being seen already in the; ViP, and earlier still in the Mbh, but some utilize the temporal frame provided by the pancalaksana scheme. What is perhaps unique about the GnP is its stated intention to use the yuga scheme and the very perceptible way it carries through this intention. The ViP does not trace the role of Visnu extensively through the entire text in the same way as the GnP. Running alongside this is the Balacarita The Life of the Boy. This name occurs in thirty-six out of one hundred and fifty-five colophons of the Kshh, overwhelmingly in the first seventy chapters, after which it does not occur at all. Within these chapters it is concentrated between chapters nine and forty-six, which deal with Ganesa's childhood from two separate perspectives: his upbringing in the hermitage of Kasyapa and Aditi, on the one hand, and his time spent with the king of kasi, on the other. The whole of the Krkh is a Balacarita in the sense that the god is presented as a boy overwhelmingly. As an adult he hardly appears at all, an observation valid also for the Ukh. Yet the linear temporality of many of the chapters is organized in line with Ganesa's age measured by year or samskara performed. This has a much more vivid and immediate impact on the narrative as the temporality of much of the action corresponds to the time of a particular samskara of the boy. In short, the periods of his boyhood break up the narration into short periods of narrated time, contrasting fundamentally with the huge time spans of the yugas.

Strictly speaking the Balacarita is subordinate to the yuga scheme, but that scheme is nowhere as developed as the Balacarita, nor does its presence emerge so closely to the surface of the text. It is, finally, arguable that the Balacarita is a generic type, pro to type in earlier texts such as the BhagP, though different from the Ganapatikhanda of the Brahmavaivarta Purana; which could conceivably have been taken as another model. I suspect though, that the childhood of Krsna in the tenth book of the BhagP, with the erotic material absented, would have been the most appropriate model as it ties in the theme of childhood with that of the avatara, a combination not replicated in the Ganapatikhanda.

II. AVATARA

What ties together both temporal frames just outlined is the avatara myth which, closely intertwined with Ganesa's games, forms the principle plot of the Krkh. The depiction of the avatara in the Krkh is absolutely dear and possesses no ambiguity. As an example are the following words of Ganesa when he addresses Kasyapa after he has killed Devantaka and Narantaka.

Formerly propitiated by austerities, I became your son, and I removed the Earth's burden and appropriately destroyed the very powerful daityas, oppressors of the triple world, Devantaka and Narantaka. The gods and the sadhus were protected and others were supported. (2, 72, 35-36)

Here are the central elements of the avatara plot: the overburdened Earth, the powerful demonic forces, the oppression of the gods and the propitiation of the god to become the avatara. Other descriptions are closer to the wording given in the famous passage from the Bhg 4, 7-8.

Why did Ganesa have to be depicted as an avatara, considering that this particular role is never brought into association with him in the classical Puranas where his mythology is initially strongly developed? Two possibilities could be suggested for this and it is significant that the avatara role is also heavily stressed in the MudP. In the first instance, Ganesa is a creator and destroyer of obstacles, either of which can be given emphasis in any of the myths in the GnP. Since, formally speaking, all myths involve the creation of a lack and its removal, the avatara myths derive their plot line from the (often inadvertent) establishment of a lack in dharma and the removal of this lack by the re-establishment of dharma. This is consistent with the theme of the god who removes obstacles and, equally, it falls within the required role of the avatara. Secondly, in a text which aims to provide a Puranic style text for the Ganapatyas or some other very prominent lineage of Ganesa devotees, the text would have to be complete in what it offers about Ganesa, but also ensure the image presented of him is one that would make him comparable to the other major gods and goddesses who are accorded avataras in the maha and Upapuranas. This responds to both sectarian and generic pressures.

III. BOYHOOD AND DIVINE IDENTITY

The compound Balacarita is a keyword whose various meanings reach far beyond the actual meanings of the words "boy" and "life." The boy is depicted brilliantly in several different guises: as a growing boy being taken through the sa1fl$kiiras, as a playmate, as someone who takes many disguises, as a son, as an object of devotion, sometimes as a young warrior and as an object of derision. In the interaction of all these themes a complex-image of the boy develops, an image strongly influenced by the way those figures with whom he interacts perceive the divine/human split. This does not mean the text is offering us a simple binary opposition which would take precedence as an heuristic device. Rather it is the implications of the interaction of all these distinct images and-the status of those who perceive them that is vital in reading the complex persona of the boy.