Introduction

The word ‘purana’ means old, ancient. The Puranas are old texts, usually referred to in conjunction with Itihasa (the Ramayana and the Mahabharata). ¹ Whether Itihasa originally meant only the Mahabharata—with the Ramayana being added to that expression later—is a proposition on which there has been some discussion. But that’s not relevant for our purposes. In the Chandogya Upanishad, there is an instance of the sage Narada approaching the sage Sanatkumara for instruction. When asked about what he already knew, Narada says he knows Itihasa and Purana, the Fifth Veda. ² In other words, Itihasa–Purana possessed an elevated status. This by no means implies that the word ‘purana’, as used in these two Upanishads and other texts too, is to be understood in the sense of the word being applied to a set of texts known as the Puranas today. The Valmiki Ramayana is believed to have been composed by Valmiki and the Mahabharata by Krishna Dvaipayana Vedavyasa. After composing the Mahabharata, Krishna Dvaipayana Vedavyasa is believed to have composed the Puranas. The use of the word ‘composed’ immediately indicates that Itihasa–Purana are smriti texts, with a human origin. They are not shruti texts, with a divine origin. Composition does not mean these texts were rendered into writing. Instead, there was a process of oral narration, with inevitable noise in the transmission and distribution process. Writing came much later.

Frederick Eden Pargiter’s book on the Puranas is still one of the best introductions to this corpus. ³ To explain the composition and transmission process, one can do no better than to quote him:

The Vayu and Padma Puranas tell us how ancient genealogies, tales and ballads were preserved, namely, by the sutas, ⁴ and they describe the suta’s duty . . . The Vayu, Brahmanda and Visnu give an account, how the original Purana came into existence . . . Those three Puranas say— Krsna Dvaipayana divided the single Veda into four and arranged them, and so was called Vyasa. He entrusted them to his four disciples, one to each, namely Paila, Vaisampayana, Jaimini and Sumantu. Then with tales, anecdotes, songs and lore that had come down from the ages he compiled a Purana, and taught it and the Itihasa to his fifth disciple, the suta Romaharsana or Lomaharsana . . . After that he composed the Mahabharata. The epic itself implies that the Purana preceded it . . . As explained above, the sutas had from remote times preserved the genealogies of gods, rishis and kings, and traditions and ballads about celebrated men, that is, exactly the material—tales, songs and ancient lore—out of which the Purana was constructed. Whether or not Vyasa composed the original Purana or superintended its compilation, is immaterial for the present purpose . . . After the original Purana was composed, by Vyasa as is said, his disciple Romaharsana taught it to his son Ugrashravas, and Ugrashravas the souti ⁵ appears as the reciter in some of the present Puranas; and the sutas still retained the right to recite it for their livelihood. But, as stated above, Romaharsana taught it to his six disciples, at least five of whom were brahmans. It thus passed into the hands of brahmans, and their appropriation and development of it increased in the course of time, as the Purana grew into many Puranas, as Sanskrit learning became peculiarly the province of the brahmans, and as new and frankly sectarian Puranas were composed.

Pargiter cited reasons for his belief that the Mahabharata was composed after the original Purana, though that runs contrary to the popular perception about the Mahabharata having been composed before the Puranas. That popular and linear perception is too simplistic, since texts evolved parallelly, not necessarily sequentially.

In popular perception, Krishna Dvaipayana Vedavyasa composed the Mahabharata. He then composed the Puranas. Alternatively, he composed an original core Purana text, which has been lost, and others embellished it through additions. The adjective ‘purana’, meaning old account or old text, became a proper noun, signifying a specific text. To be classified as a Purana, a Purana has to possess five attributes—pancha lakshmana. That is, five topics must be discussed—sarga, pratisarga, vamsha, manvantara and vamshanucharita. The clearest statement of this is in the Matsya Purana. A text like the Bhagavata Purana also mentions these five attributes, but adds another five, making it a total of ten. Unlike the Ramayana and the Mahabharata, there is no Critical Edition of the Puranas. ⁶ Therefore, citing chapter and verse from a Purana text is somewhat more difficult, since verse, if not chapter, may vary from text to text. With that caveat, the relevant shloka (verse) should be in the fifty-third chapter of the Matysa Purana. Sarga means the original or primary creation. The converse of sarga is universal destruction, or pralaya. That period of sarga lasts for one of Brahma’s days, known as kalpa. When Brahma sleeps, during his night, there is universal destruction.

In measuring time, there is the notion of a yuga (era) and there are four yugas—satya yuga (also known as krita yuga), treta yuga, dvapara yuga and kali yuga. Satya yuga lasts for 4,000 years, treta yuga for 3,000 years, dvapara yuga for 2,000 years and kali yuga for 1,000 years. However, all these are not human years. The gods have a different timescale and these are the years of the gods. As one progressively moves from satya yuga to kali yuga, virtue (dharma) declines. But at the end of kali yuga, the cycle begins afresh, with satya yuga. An entire cycle, from satya yuga to kali yuga, is known as a mahayuga (great era). However, a mahayuga is not just 10,000 years. There is a further complication. At the beginning and the end of every yuga, there are some additional years. These additional years are 400 for satya yuga, 300 for treta yuga, 200 for dvapara yuga and 100 for kali yuga. A mahayuga thus has 12,000 years, adding years both at the beginning and at the end. One thousand mahayugas make up one kalpa. A kalpa is also divided into fourteen manvantaras, a manvantara being a period during which a Manu presides and rules over creation. Therefore, there are 71.4 mahayugas in a manvantara. Our present kalpa is known as the Shveta Varaha Kalpa. Within that, six Manus have come and gone. Their names are (1) Svayambhuva Manu, (2) Svarochisha Manu, (3) Uttama Manu, (4) Tapasa Manu, (5) Raivata Manu and (6) Chakshusha Manu. The present Manu is known as Vaivasvata Manu. Vivasvat, also written as Vivasvan, is the name of Surya, the sun god. Vaivasvata Manu has that name because he is Surya’s son. Not only the Manus, but the gods, the ruler of the gods and the seven great sages, known as the saptarshis (seven rishis), change from one manvantara to another. Indra is a title of the ruler of the gods. It is not a proper name. The present Indra is Purandara. However, in a different manvantara, someone else will hold the title. In the present seventh manvantara, known as Vaivasvata manvantara, there will also be 71.4 mahayugas. We are in the twenty-eighth of these. Since a different Vedavyasa performs that task of classifying and collating the Vedas in every mahayuga, Krishna Dvaipayana Vedavyasa is the twenty-eighth in that series. Just so that it is clear, Vedavyasa isn’t a proper name. It is a title conferred on someone who collates and classifies the Vedas. There have been twenty-seven who have held the title of Vedavyasa before Krishna Dvaipayana and he is the twenty-eighth. His proper name is Krishna Dvaipayana—Krishna because he was dark and Dvaipayna because he was born on an island (dvipa). This gives us an idea of what the topic of manvantara is. This still leaves pratisarga, vamsha and vamshanucharita. The two famous dynasties/lineages were the solar dynasty (surya vamsha) and lunar dynasty (chandra vamsha) and all the famous kings belonged to one or other of these two dynasties. Vamshanucharita is about these lineages and the conduct of these kings. There were the gods and sages (rishis) too, not always born through a process of physical procreation. Their lineages are described under the heading of vamsha. Finally, within that cycle of primary creation and destruction, there are smaller and secondary cycles of creation and destruction. That’s the domain of pratisarga. In greater or lesser degree, all the Puranas cover these five topics, some more than the others.

There are Puranas, and there are Puranas. Some are known as Sthala Puranas, describing the greatness and sanctity of a specific geographical place. Some are known as Upa-Puranas, minor Puranas. The listing of Upa-Puranas has regional variations and there is no countrywide consensus about the list of Upa-Puranas, though it is often accepted that there are eighteen of them. The Puranas we have in mind are known as Maha-Puranas, major Puranas. Henceforth, when we use the word Puranas, we mean Maha-Puranas. There is consensus that there are eighteen Maha-Puranas, though it is not obvious that this number of eighteen existed right from the beginning. The names are mentioned in several of these texts, including a shloka that follows the shloka cited from the Matsya Purana. The listing is also included in the last sections of the Bhagavata Purana itself. Thus, the eighteen Puranas are (1) Agni (15,400); (2) Bhagavata (18,000); (3) Brahma (10,000); (4) Brahmanda (12,000); (5) Brahmavaivarta (18,000); (6) Garuda (19,000); (7) Kurma (17,000); (8) Linga (11,000); (9) Markandeya (9,000); (10) Matsya (14,000); (11) Narada (25,000); (12) Padma (55,000); (13) Shiva (24,000); (14) Skanda (81,100); (15) Vamana (10,000); (16) Varaha (24,000); (17) Vayu (24,000) and (18) Vishnu (23,000). A few additional points about this list. First, the Harivamsha is sometimes loosely described as a Purana, but strictly speaking, it is not a Purana. It is more like an addendum to the Mahabharata. Second, Bhavishya (14,500) is sometimes mentioned, with Vayu excised from the list. However, the Vayu Purana exhibits many more Purana characteristics than the Bhavishya Purana does. There are references to a Bhavishyat Purana that existed, but that may not necessarily be the Bhavishya Purana as we know it today. That’s true of some other Puranas too. Texts have been completely restructured hundreds of years later. Third, it is not just a question of Bhavishya Purana and Vayu Purana. In the lists given in some Puranas, Vayu is part of the eighteen, but Agni is knocked out. In some others, Narasimha and Vayu are included, but Brahmanda and Garuda are knocked out. Fourth, when a list is given, the order also indicates some notion of priority or importance. Since that varies from text to text, our listing is simply alphabetical, according to the English alphabet.

The numbers within brackets indicate the number of shlokas each of these Puranas has, or is believed to have. The range is from 10,000 in Brahma to a mammoth 81,100 in Skanda. The aggregate is a colossal 409,500 shlokas. To convey a rough idea of the orders of magnitude, the Mahabharata has, or is believed to have, 100,000 shlokas. It’s a bit difficult to convert a shloka into word counts in English, especially because Sanskrit words have a slightly different structure. However, as a very crude approximation, one shloka is roughly twenty words. Thus, 100,000 shlokas become two million words and 400,000 shlokas, four times the size of the Mahabharata, amounts to eight million words. There is a reason for using the expression ‘is believed to have’, as opposed to ‘has’. Rendering into writing is of later vintage, the initial process was one of oral transmission. In the process, many texts have been lost, or are retained in imperfect condition. This is true of texts in general and is also specifically true of Itihasa and Puranas. The Critical Edition of the Mahabharata, mentioned earlier, no longer possesses 100,000 shlokas. Including the Harivamsha, there are around 80,000 shlokas. The Critical Edition of the Mahabharata has of course deliberately excised some shlokas. For the Puranas, there is no counterpart of Critical Editions. However, whichever edition of the Puranas one chooses, the number of shlokas in that specific Purana will be smaller than the numbers given above. Either those many shlokas did not originally exist, or they have been lost. This is the right place to mention that a reading of the Puranas assumes a basic degree of familiarity with the Valmiki Ramayana and the Mahabharata, more the latter than the former. Without that familiarity, one will often fail to appreciate the context completely. Specifically for the Bhagavata Purana, more than passing familiarity with the Bhagavad Gita—strictly speaking, a part of the Mahabharata—helps. ⁷

Other than the five attributes, the Puranas have a considerable amount of information on geography and even geological changes (changes in courses of rivers) and astronomy. Therefore, those five attributes shouldn’t suggest the Puranas have nothing more. They do, and they have therefore been described as encyclopedias. Bharatavarsha is vast and heterogeneous and each Purana may very well have originated in one particular part of the country. Accordingly, within that broad compass of an overall geographical description, the extent of geographical information varies from Purana to Purana. Some are more familiar with one part of the country than with another. Though not explicitly mentioned in the five attributes, the Puranas are also about pursuing dharma, artha, kama and moksha, the four objectives of human existence, and are about the four varnas and the four ashramas. The general understanding and practice of dharma is based much more on the Puranas than on the Vedas. Culture, notions of law, rituals, architecture and iconography are based on the Puranas. There is beautiful poetry too, included in parts of the Bhagavata Purana.

Perhaps one should mention that there are two ways these eighteen Puranas are classified. The trinity has Brahma as the creator, Vishnu as the preserver and Shiva as the destroyer. Therefore, Puranas where creation themes feature prominently are identified with Brahma (Brahma, Brahmanda, Brahmavaivarta, Markandeya). Puranas where Vishnu features prominently are identified as Vaishnava Puranas (Bhagavata, Garuda, Kurma, Matsya, Narada, Padma, Vamana, Varaha, Vishnu). Puranas where Shiva features prominently are identified as Shaiva Puranas (Agni, Linga, Shiva, Skanda, Vayu). While there is a grain of truth in this, Brahma, Vishnu and Shiva are all important and all three feature in every Purana. Therefore, beyond the relative superiority of Vishnu vis-à-vis Shiva, the taxonomy probably doesn’t serve much purpose. The second classification is even more tenuous and is based on the three gunas of sattva (purity), rajas (passion) and tamas (ignorance). For example, the Uttara Khanda of the Padma Purana has a few shlokas along these lines, recited by Shiva to Parvati. With a caveat similar to the one mentioned earlier, this should be in the 236th chapter of Uttara Khanda. According to this, the Puranas characterized by sattva are Bhagavata, Garuda, Narada, Padma, Varaha and Vishnu. Those characterized by rajas are Bhavishya, Brahma, Brahmanda, Brahmavaivarta, Markandeya and Vamana, Those characterized by tamas are Agni, Kurma, Linga, Matsya, Skanda and Shiva.

Within a specific Purana text, there are earlier sections, as well as later ones. That makes it difficult to date a Purana, except as a range. Across Purana texts, there are older Puranas, as well as later ones. Extremely speculatively, the dating will be something like the following. (1) Agni (800–1100 CE); (2) Bhagavata (500–1000 CE); (3) Brahma (700–1500 CE); (4) Brahmanda (400–600 CE); (5) Brahmavaivarta (700–1500 CE); (6) Garuda (800–1100 CE); (7) Kurma (600–900 CE); (8) Linga (500–1000 CE); (9) Markandeya (250–700 CE); (10) Matsya (200–500 CE); (11) Narada (900–1600 CE); (12) Padma (400–1600 CE); (13) Shiva (1000–1400 CE); (14) Skanda (600–1200 CE); (15) Vamana (450–900 CE); (16) Varaha (1000–1200 CE); (17) Vayu (350–550 CE); (18) Vishnu (300 BCE to 450 CE); and (19) Bhavishya (500–1900 CE). Reiterating once again that there is no great precision in these ranges, by this reckoning, the Vishnu Purana is the oldest and some parts of the Bhavishya Purana are as recent as the nineteenth century.

As mentioned earlier, there is no Critical Edition for the Puranas. Therefore, one has to choose a Sanskrit text one is going to translate from. If one is going to translate all the Puranas, it is preferable, though not essential, that one opts for a common source for all the Purana texts. The common source for the Bhagavata Purana, and the subsequent Purana translations, is the one brought out by Nag Publishers, with funding from the ministry of human resource development. ⁸ A transliterated Sanskrit text is available through the University of Gottingen. ⁹ In Devanagari, the text is available at https://sanskritdocuments.org/doc_purana/bhagpur.html?lang=sa, with the transliterated Sanskrit text at http://www.sanskritweb.net/sansdocs/bhagpur.pdf. The Oxford Centre for Hindu Studies at the University of Oxford has a lot of useful resources under the umbrella of the Bhagavata Purana Research Project. ¹⁰ Among other things, this research project has published a book that is a bit like an abridged translation of the Bhagavata Purana. ¹¹ The Bhagavata Purana is divided into twelve skandhas. The word skandha means trunk, or largish branch. In this context, it means section or segment. Out of these twelve skandhas, the tenth is the longest and is also the most read. Therefore, there are translations of the Bhagavata Purana that are limited to only the Tenth Skandha. Edwin Bryant’s rendering of the Tenth Skandha is almost like such a translation. ¹² For the entire Bhagavata Purana, there are unabridged translations in Indian languages. However, to the best of my knowledge, there are only five unabridged translations in English: (1) Manmatha Nath Dutt; ¹³ (2) Bhaktivedanta Swami Prabhupada; ¹⁴ (3) Ganesh Vasudeo Tagare; ¹⁵ (4) Swami Tapasyananda; ¹⁶ and (5) C.L. Goswami and M.A. Shastri. ¹⁷ In 1901, Purnendu Narayana Sinha published a version that was close to a translation, but fell just short because it followed the path of retelling. ¹⁸ The Sanskrit texts used in these five translations vary a bit and the one used in this translation also varies a bit from the ones used in these five. The Bhagavata Purana is believed to have 18,000 shlokas. The table below shows what this particular Sanskrit version has: just over 14,000 spread across 335 chapters. One should not jump to the conclusion that a large number of shlokas are missing. A few are indeed missing. But sometimes, it is also a question of how one counts a shloka. With the content remaining identical, the text may be counted as one shloka in one place and as two shlokas elsewhere. Our numbering is exactly the same as in the Sanskrit text we have followed. Hence, even though there may be no difference in content between our version of the text and, say, that used by Swami Prabhupada, the numbering will vary a bit. (Sometimes, there are minor differences in the Sanskrit text though.) However, there are some shlokas that are indeed missing and Ganesh Vasudeo Tagare is a good source for translations of these missing shlokas, because those missing shlokas have been separately translated there. When we have enumerated all the chapters, there are two identifying numbers that have been provided. The first number refers to the skandha, the second to the number of the chapter within the skandha. Thus, Chapter 4(30) will be the thirtieth chapter in the fourth skandha.

In the translations of the Bhagavad Gita, the Mahabharata, the Harivamsha and the Valmiki Ramayana, we followed the principle of not using diacritical marks. The use of diacritical marks (effectively the international alphabet of Sanskrit transliteration) makes the pronunciation and rendering more accurate, but also tends to put off readers who are less academically inclined. Since diacritical marks are not being used, there is a challenge of rendering Sanskrit names in English. Sanskrit is a phonetic language and we have used that principle as a basis. Applied consistently, this means that words are rendered in ways that may seem unfamiliar. Hence, the name of the jewel, often written as Kaustubha, will appear as Koustubha here. This is true of proper names, and, in a few rare cases, of geographical names. The absence of diacritical marks causes some minor problems. How does one distinguish Vasudeva Krishna from Krishna’s father, Vasudeva? Often, the context will make the difference clear. If not, we have written the son as Vaasudeva and the father as Vasudeva. In translating, the attempt has been to provide a word-for-word translation, so that if one were to hold up the Sanskrit text, there would be a perfect match.

The intention is also to offer a translation, not an interpretation. That sounds like a simple principle to adopt, and for the most part, is easy to follow. However, there is a thin dividing line between translation and interpretation. In some instances, it is impossible to translate without bringing in a little bit of interpretation. Inevitably, interpretation is subjective. We have tried to minimize the problem by (a) reducing interpretation; (b) relegating interpretation to footnotes and (c) when there are alternative interpretations, pointing this out to the reader through those footnotes. If a Purana text is just about the five attributes mentioned, the task isn’t that difficult. Take the Bhagavata Purana as an example. In Chapter 2(10), we are told about the ten characteristics of the Bhagavata Purana and these are sarga, visarga, sthana, poshana, uti, manvantara, isha-anukatha, nirodha, mukti and ashraya. These are (1) gross creation; (2) subtle creation; (3) preservation; (4) sustenance; (5) the addiction to the senses; (6) manvantaras; (7) accounts of the lord; (8) withdrawal; (9) liberation and (10) the ultimate refuge. Therefore, the text has much more than the five attributes cited. The word ‘bhagavat’ means the divine one, the illustrious one, the fortunate one, the prosperous one, the blessed one, the sacred one, the holy one. We are often more familiar with the nominative form, bhagavaan. Bhagavat/Bhagavaan occur frequently in our text and I have translated the word as ‘the illustrious one’. Since this is a Purana devoted to Vishnu, Bhagavat/Bhagavaan typically means Vishnu. But the text is not called Bhagavata Purana; it is in fact called Bhaagavata Purana. What does the word Bhaagavata mean? Taken as an adjective, the word means sacred and Bhaagavata Purana is simply the sacred Purana, a perfectly acceptable meaning. However, taken as a noun, Bhaagavata means a devotee or worshipper of Vishnu—and bhaagavata dharma is the dharma of devotion towards Vishnu. Therefore, Bhaagavata Purana is the text that describes this bhaagavata dharma that Vishnu devotees follow.

Hence, it isn’t only about stories. As mentioned earlier, the longest skandha, the tenth, is also the most popular and is read most often. This particular skandha describes Krishna’s exploits and pastimes. This includes rasa lila and this skandha has some exquisite poetry. The song of the gopis is an example of this. As is thus obvious, this skandha has its fair share of stories. But in the other skandhas, the emphasis is more on dharma and moksha and a mix of bhakti (devotion), advaita (monism), dvaita (dualism), samkhya and yoga, ¹⁹ with notions of avataras (Vishnu’s incarnations) brought in. There is the influence of pancharatra doctrines and texts. Literally, the term pancharatra means five nights, and is a reference to five nights over which sacrifices were performed. However, there were pancharatra doctrines too and these developed a metaphysical philosophy. If we so wish, we can call it a theology. Vaishnava dharma integrated this pancharatra tradition with the other strands that have just been mentioned. Since this is a synthesis of many different strands, there are bound to be nuances and different schools. What does the Bhagavata Purana teach? What is its philosophy? Beyond the obvious tenet of devotion to Vishnu and emphasis on bhakti yoga, this is extremely difficult to answer. As a text that builds a theology, the Bhagavata Purana is much more complicated than the Bhagavad Gita. Indeed, one could say it builds extensively on that Bhagavad Gita foundation. Down the years, there have been several commentaries on and interpretations of the Bhagavata Purana. Depending on the language and the timeline for beginning and ending the cut-off, there must be at least two hundred such commentaries and interpretations. Which of these schools or interpretations should one follow? Should one take cognizance of other possible interpretations? In principle, this dilemma can exist in the course of translating any Purana. But in no other Purana is it as serious, because no other Purana devotes such a large percentage of shlokas to philosophy.

The choice in this translation is conscious and is driven by the objective and the target audience. The target audience is the ordinary reader who desires a faithful rendering of the Sanskrit text. The target reader isn’t the academic who desires something like a survey of literature. Nor is the target reader a devotee of one particular Vaishnava school. To state it a bit more explicitly, Bhaktivedanta Swami Prabhupada, Swami Tapasyananda and C.L. Goswami and M.A. Shastri follow one particular line of commentary or interpretation or school. From an academic’s perspective, Ganesh Vasudeo Tagare is the best. His translation also includes an excellent introduction and a detailed note on several different commentators of the Bhagavata Purana, certainly covering the major ones. However, apart from the language of the Tagare rendition not being very smooth, that edition is also layered with complexities which can deter the ordinary reader. On the other hand, I have consciously done what is akin to a contemporary Manmatha Nath Dutt translation. Without distorting, I have chosen the simplest possible interpretation or translation that fits. In a few minor instances, this meant choosing an interpretation that was marginally different from those chosen by these preceding translators. It is possible to criticize the avoidance of complexities, but it is a conscious choice and the dissatisfied reader can always go on to read more copious commentaries and annotations. There is plenty in the Bhagavata Purana to immerse oneself in and this translation is not meant to be the final item on that reading list. It can at best be the first.