Frederick Eden Pargiter
by Wikipedia
Accessed: 6/7/21
Frederick Eden Pargiter: Excerpt from The Puranas, by Ludo Rocher
The most important names in this period, from a methodological point of view, are Frederick Eden Pargiter -- whose historical evaluation of puranas has been admired and criticized, but never ignored -- and Willibald Kirfel -- whose strictly philological approach created a true school of purana scholars, especially in Germany.21 [21 Cf. Hermann BERGER (rev. of Losch's Rajadharma, ZDMG 113, 1963, 390): "Seit die in der Indologie etwas stiefmutterlich behandelten Puranas durch W. Kirfels bahnbrechende Arbeiten erstmals eine systematische Behandlung erfahren haben, beginnen immer mehr Indologen diesen fur die indische Geistesgeschichte so grundlegenden Texten ihr Interesse zuzuwenden." [Google translate: Since the Puranas, treated somewhat stepmotherly in Indology, were groundbreaking by W. Kirfels, Work for the first time have received systematic treatment more and more. Indologists wrote these texts, which are so fundamental to Indian intellectual history to turn their interest.] Their contributions, as well as the reactions they provoked, will be analyzed elsewhere in this volume....
1.4.1 The "Ur-Purana"
On the basis of the indices which Wilson prepared from various puranas, he could not fail being struck by the fact that individual puranas have numerous topics in common. When comparing the parallel passages in the Brahmaº and the Visnuº, he proposed, cautiously: "they appear to have been taken from some older work or works, from which the present Puranas are, probably, in part at least derived."1 [1 JRAS 5, 1839, 66.] One year later, the Preface to the translation of the Visnuº, subsequent to the discussion of the sectarian character of the puranas, states more boldly:"The identity of the legends in many of them, and still more the identity of the words -- for in several of them long passages are literally the same -- is a sufficient proof that in all such cases they must be copied either from some other similar work, or from a common and prior original."2 [2 WILSON 1840 = 1961: IV.]
Lassen, as always, echoed Wilson, but he referred unequivocally to the fact that, as far as the corresponding passages in individual puranas are concerned, the texts "have made use of a common prior source."3 [3 LASSEN 1847: 480.] In 1905 4 [4 In 1897, LUDERS applied a similar method to the legend of Rsyasrnga; see references sub Padmaº Patalakhanda.] A. A. Macdonell stressed the same idea more specifically with regard to what is supposed to be the main topic of the puranas, pancalaksana: "In that part of their contents which is peculiar to them, the Puranas agree so closely, being often verbally identical for pages, that they must be derived from some older collection as a common source."5 [5 1900: 299.]Pañcalakṣaṇa, occurring in the Amarakośa and in various Purāṇas, enumerates creation (sarga), recreation (pratisarga), genealogy (vaṃśa), cosmic cycles (manvantara) and accounts of royal dynasties (vaṃśānucarita) as five characteristics of a Purāṇa, but many of the extant Mahā-purāṇas and almost all the Upapurāṇas do not follow this definition. They have rather become “Codes of Hindu rites and customs by including chapters on varṇāśramadharma, ācāra, śrāddha, prāyaścita, dāna, pūjā, vrata, tīrtha, pratiṣṭhā, dīkṣā, utsarga etc.”
-- The Nilamata Purana: A cultural and Literary study of a Kasmiri Purana, by Dr. Ved Kumari
By that time the idea of a common origin of all puranas had obviously fully taken root, for in the same year A. M. T. Jackson6 [6 1905: 67-77.] wrote an article "to enquire whether we can fix approximately the scope and date of composition of this original." He came to the conclusion that "the original purana may be regarded with some probability as a work of the 4th century B.C." He even decided that it was a Saiva work, and continued:"It is quite possible that the genealogies and the lists of rivers and tribes were originally drawn up in prose. At some date, which is at present unknown, the original purana was re-written in verse, while the original chronology gave place to the system of Kalpas, and the history subsequent to the great war was thrown into prophetic form. This second version was the common source of the extant puranas."
Three years later Blau applied the comparative method to the legend of Saranyu -- and spoke, for the first time, explicitly of an "Ur-Purana" as the common source of the extant puranas "in ihren echtesten Partien." [Google translate: in their most genuine parts]7 [7 August BLAU: Puranische Streifen. I. Der Itihasa von Saranyu in seiner Fortbildung durch die Purana [Google translate: Puranic Stripes. I. The Itihasa of Saranyu in his further training through the Purana], ZDMG 62, 1908, 337-357 at 337. Blau also saw another reason why it is important to determine the oldest parts of the puranas: "Denn so wenig die Massenhaftigkeit der puranischen Produktion und der z. T. hochst unerfreuliche Inhalt dieser Literatur zu naherer Beschaftigung mit ihr einladen mogen, so ist es gerade darum um so wunschenswerter, dass das Ursprungliche und Alte in den einzelnen Purana herausgehoben und miteinander verglichen werde." [Google translate: Because so little the masses of puranic production and the content of this literature is partly extremely unpleasant invite you to take a closer look at it, it is precisely for that reason that it is all the more desirable that the original and the old are emphasized in the individual Purana and compared to each other.]
The first application of the principle on a large scale came from Frederick Eden Pargiter. In an article, in 1913, on "Visvamitra and Vasistha," he defined his methodology as follows: "The texts for each story are cited. They are all obviously based on a common original metrical tradition, and by collating them a revised text may be framed. This I have done, and I give the collated version here with such variant readings only as are material."8 [8 JRAS 1913, 885-904 at 885.] Far more important was the publication, in the same year, of The Purana Text of the Dynasties of the Kali Age. In it Pargiter reconstructed the original puranic account of the dynasties that reigned in India during the Kali era, based on editions and manuscripts of the Matsyaº, Vayuº, Brahmandaº, Visnuº, Bhagavataº, Garudaº, and Bhavisyaº. A more general volume on Indian history, based on the same principles, appeared in 1922: Ancient Indian Historical Tradition. Once again Pargiter clearly stated his methodology:"In examining the genealogies it is of little profit and is likely to be misleading to deal with the accounts of the several Puranas separately. The only trustworthy course is first to collate the texts that generally agree and ascertain as far as possible what original text they indicate, and then construct the genealogy therefrom. By this method individual corruptions and errors can be corrected, losses and omissions remedied, and interpolations and alterations detected with reasonable confidence; and thus a text may be framed which approaches as nearly as is possible to the common original on which all those texts were based."9 [9 PARGITER 1922: 82.]
The merits of Pargiter's methodology in reconstructing the early history of India on the basis of the puranas will be examined later in this volume (see 2.2.4). At this point I merely want to stress the emergence of the concept of reducing parallel accounts in various puranas to one single original. In Pargiter's case this concept was so successful that, from then onward, many historians of India were to base their research and their writings directly on Pargiter's reconstructed text rather than on the more cumbersome editions of individual puranas.10 [10 E.g., K.G. SANKAR (Some Problems of Indian Chronology, ABORI 12, 1931, 301-361) uses PARGITER 1913 "throughout this article." According to Ferdinand BOCK (Die puranas als Geschichtsquelle [Google translate: The puranas as a source of history], WZKM 29, 1931,97-133), "Die neue PARGITERSCHE Ausgabe der Puranas [emphasis added], die ich allen Zitaten zugrunde lege, zeigt dem Philologen auf den erst en Blick so viel Auffallendes, dass eine genaue Prufung des Ganzen unerlasslich erscheint" [Google translate: The new PARGITER edition of the Puranas [emphasis added] on which I base all quotations shows the philologist so much striking at first glance that a thorough examination of the whole seems indispensable.] (p. 98). Occasionally, the puranic lists of dynasties have been emended, starting from Pargiter's text. E.g., V. V. MlRASHI (The puranas on the Successors of the Satavahanas in Vidarbha; Pur 18, 1976, 88-92) proposes to correct the unknown Maunah (at PARGITER 1913: 46) into Maundah.]...
2.1.3 Different Recensions of Puranas
The existing editions of individual puranas exhibit a wide range of discrepancies, from minor variant readings to the inclusion or exclusion of entire chapters or sections. The latter situation applies, for instance, in the case of the Padmaº: "The present Padma, which is the result of several recasts, has come down to us in two distinct recensions -- North Indian (Bengal) and South Indian."1 [1 HAZRA 1940: 107. In the case of the Padmaº this distinction was already made by Luders (Sage von Rsyasrnga, 1897, ref. sub Padmaº), including the statement that the Bengali version is the older one (p. 94 n. 1). Cf. WINTERNITZ 1907: 452.] This statement by Hazra reflects the traditional interpretation of the major discrepancies, especially those between the Vangavasi editions1a [1a Even though, in the particular case of the Padmaº (see there) the Vangavasi ed. too, reproduces the "Western" Padmaº.] on the one hand and the Venkatesvara and/or Anandasrama editions on the other. Haraprasad Shastri formulated the general principle as follows:"They represent the different provincial recensions and that means collation of different classes of manuscripts ... By a cursory view of the two sets, one can at once come to know ... that some khandas of the Puranas are popular in one province and unknown in another and so forth."2 [2 HARAPRASAD SHASTRI: 1928b: 327-328.]
Elsewhere the editions are much closer to one another, as in the case of the Brahmaº: "The AnSS ed. is chapter by chapter the same as the Vanga ed. There are occasional variations in readings and number of verses in the corresponding chapters, but these variations are not many and important."3 [3 HAZRA 1940: 145 n. 163.] We might be tempted to conclude from statements such as this that there are also puranas without major, regional differences, were it not that we know that the Vangavasi editions and the Venkatesvara and/or Anandasrama editions are not always as independent of one another as the geographic distance between Calcutta and Bombay or Pune might make us believe. Thus, in the preface to the Vanga edition of the Skandaº the editor explains that he took the Venkatesvara edition as his basis, and merely added to it chapters and verses he found in the Bengal manuscripts.4 [4 Pp. 10-11, quoted at HAZRA 1940: 157 n. 176.] And with regard to the Vamanaº editions Hazra surmised with good reasons: "The Vanga ed. is the same as the Venk. ed. Both consist of 95 chapters. The variations in readings in these two editions are so small in number that one seems to be a reprint of the other."5 [5 HAZRA 1940: 76 n. 1. Cf. PADOUX (Agniº 1978: 58) on the Agniº: "each new edition seems to make it a point to reproduce faithfully even the most obvious errors of the previous ones."]A good example to show that corresponding chapters in different regional editions of a particular purana do not warrant any conclusion as to the original organization of the text, is provided by the Agniº. Hazra adds the usual note: "The Vanga. ed. is chapter by chapter the same as the AnSS ed. There are, of course, occasional variations in readings and number of verses in the corresponding chapters."6 [6 HAZRA 1940: 134 n. 122.] Gyani agrees, and further specifies: "These editions do not much differ from one another. All the editions contain 382 chapters but the Venkatesvar edition has got one chapter in excess. The Chapter 135 entitled Atha Sangrama Vijaya Vidya cannot be traced in other editions."7 [7 GYANI Agniº 1964: 3 n. 1.] Not only has the additional chapter unostentatiously been inserted in more recent Anandasrama editions, but, more importantly, the apparently uniform division of the Agniº in 382 is, in reality, not a very ancient one. It was the work of its first editor, Rajendralala Mitra [16 February 1822 – 26 July 1891], who used eleven manuscripts, and described their relationship and arrangement as follows: "Of these, eight codices correspond very closely, and give the same number of chapters; two Nos. I and VII are incomplete, wanting several chapters at the end; and one, No. III, has several chapters at the end, and 4 chapters on pilgrimage in the middle the counterparts of which are not to be met with elsewhere. The chapters are not regularly numbered in any of the MSS., and in several no number is to be met with. For the sake of convenience of reference the serial number has been introduced by me, and the total I arrived at from the eight MSS. which correspond is 382".8 [8 Agniº ed., 1873-79, pp. xxxvii-xxxviii.]
What I would like to suggest is that, even when different editions of a particular purana correspond in their general arrangement, this does not mean that we are in the fortunate position of possessing the text of that purana. On the contrary, there are reasons to believe that, irrespective of whether or not the printed editions exhibit major differences, all titles of puranas have been -- and are -- used to cover a variety of materials which do not appear in our editions. Some of these materials may still be available in -- numerous -- manuscripts which have never been consulted by purana editors; others may have existed in manuscripts which are now lost; others again may have been included in the recitations of individual sutas without ever having been committed to writing.
As early as 1890 Buhler showed that al-Biruni's quotations from "the Visnudharma" are, in fact, from two different versions of the Visnudharmottaraº. One of his conclusions was that "it is evident that in the beginning of the eleventh century two works with the title Vishnudharmottara or Vishnu-Dharma existed, and that both were considered to be canonical by Biruni's Pandits who, one and all, were Vaisnavas."9 [9 Book-notice on Sachau's Alberuni, IA 19, 1890, 381-410 at 407.]
On numerous occasions scholars have taken notice of the existence of manuscripts the contents of which did not correspond to those of any of the published versions. According to Burnell, "eighteen Puranas are mentioned everywhere; but they are often by no means the same works, though under one name."10 [10 1880: 187. Cf. p. 189, on Brahmaº, etc.] In Haraprasad's catalogue of the purana manuscripts at the Asiatic Society in Calcutta one comes across several statements, similar to the one on a Bhavisyaº manuscript: "it does not agree with any of the recensions of the purana known."11 [11 HARAPRASAD SHASTRI 1928a: 424. Cf. p. (Varahaº), p. 647 (Vamanaº), p. 784 (Nilamataº), etc.] R. N. Mehta came to the conclusion that "under the title of Nagarakhanda several works exist."12 [12 Nagarakhanda -- A Study, JUBar 17, 1968, 106.] Bonazzoli refers to the existence of five different Bhavisyaºs.13 [13 BONAZZOLI Bhavisyaº 1979: 26-27.] Even for a purana for which most editions correspond, such as the Devibhagavataº, Lalye saw a manuscript which is totally different.14 [14 LALYE Devibhagavataº 1973: viii.]
Equally revealing for the existence of different texts with identical titles are certain statements in the dharmanibandhas. Besides the puranas with these names which he did consult (pp. 2-3) Ballalasena refers in the Danasagara to "another" (apara) Brahmaº, Agniº, Visnuº, and Lingaº which he did not use (p. 7). Inasmuch as the contents of these "other" puranas, as indicated in the Danasagara, do not correspond to those of the extant puranas, the conclusion was drawn that in addition to the latter -- which are spurious by definition (see 1.3) -- there was for each of them at least one other spurious text.15 [15 E. g., HAZRA 1940: 95 n. 40 (Lingaº), 151 n. 168 (Brahmaº).] Similarly, Narasimha Vajapeyin's Nityacarapradipa distinguishes between two different Brahmaºs, one of which is quoted in Laksmidhara's Krtyakalpataru, the other in the works of Hemadri; the text labels the latter an upapurana.16 [16 BI work 160, 1903-28, 1.19.] Ballalasena (p. 7) also refers to a few puranas which were useless for his purpose, since they do not contain rules on gifts (danavidhisunya); one of these is the Brahmandaº. Not only does the extant Brahmandaº contain such rules; Hemadri's Caturvargacintamani, attributes to it many other verses on the subject which do not appear in our editions. This shows "that the text of the 'Brahmanda', used by Hemadri, was in many ways different from that of our present edition as well as from that of the Brahmanda known to Ballalasena."17 [17 HAZRA 1940: 19.]The Kalikaº is a title which has created problems for several scholars. Raghavan was the first one to suggest the existence of at least three versions: the one represented by most manuscripts and editions, comprising from 90 to 93 chapters; L. 370 of Aufrecht's Catalogus Catalogorum, which Rajendralala Mitra himself calls Candiº; and the India Office manuscript, a later and different text.18 [18 RAGHAVAN Kalikaº 1938: 331.] The editor of Laksmidhara's Krtyakalpataru, K. V. Rangaswami Aiyangar, came to the conclusion that "it has been ... difficult to find any of the quotations from Kalikapurana in any of the printed editions of it. The existence of Kalikapurana in more than one recension, and the radical differences between rival versions of it, might justify the suspicion that we do not now possess it in the form in which it existed in the 11th and 12th centuries."19 [19 RANGASWAMI AIYANGAR Nandiº 1941-42.: 159.] On the basis of the India Office catalogue19a [19a EGGELING 1899: 1193-1198, no. 3344.] Hazra accepts the existence of another Kalikaº, also called Kaliº and Satiº, which is quite different from "the present" Kalikaº.20 [20 HAZRA 1963: 259.] Elsewhere Hazra mentions that Raghunandana's Durgapujatattva (pp. 8-9) quotes ten verses which are introduced as dusprapakalikapuranantare 'pi. It shows "that Raghunandana knew another Kalika-p. which was different from the present one profusely drawn upon by him in his Tattvas."21 [21 HAZRA 1963: 236.]Raghunandana was born at Nabadwip, to Harihara Bhattacharya. He was a pupil of Srinatha Acharya Chudamani. His writings mention Rayamukuta (1431 CE), and are mentioned by Viramitrodaya of Mitramisra (early 17th century). Thus, it can be inferred that Raghunandana lived around 16th century CE.
-- Raghunandana [Raghunandan Bhattacharyya] [Raghunandana Bhaṭṭācāryya], by Wikipedia
Notice also that Nilakantha's [Mimamasakabhatta's] Vyavaharamayukha, after quoting three stanzas from the Kalikaº, says that they do not deserve absolute confidence, "for they are absent from two or three manuscripts of the text."22 [22 Ed. P. V. KANE, Bombay: NSP, 1926, p. 114.]The introductory verses in the mss. of all the Mayukhas present a perplexing problem. Hardly any two mss. of the same Mayukha contain the same introductory verses. For example, one of the three mss. of the Samayamayukha in the Bhau Daji collection of the Bombay Royal Asiatic Society has only one introductory verse2 [[x].]; while in the other two that verse does not occur at all. In one of these two latter there are four introductory verses and in the other there are five, the Benares edition agreeing with the last. The Benares edition of the S'antimayukha (of 1879) contains fourteen introductory verses, nine of which (from the second) give the genealogy of the family of Bhagavantadeva; while one ms. of the S'antimayukha in the Bhau Daji collection has only one introductory verse which is not found in the Benares edition; and another ms. of the same in the same collection has three verses, only one of which is found in the Benares edition. In the same way the printed editions of the Prayas'cittamayukha and the Acaramayukha (Benares, 1879) contain fourteen introductory verses each; while mss. of these two Mayukhas in the Gattulalji collection in Bombay have only two and three verses respectively. This perplexing variance in the number of introductory verses cannot be satisfactorily explained by supposing that in all cases of such differences the scribes of the mss. and others introduced unauthorised interpolations. The hypothesis which, after a careful consideration of all the introductions, seems most probable is that Nilakantha himself (or probably his son) from time to time revised his works, recast the introductory verses, added to them and also made slight alterations and additions in the body of the works.
Some of the Mayukhas such as the printed editions of the S'anti, Prayas'citta, S'raddha and Acara Mayukhas contain the genealogy of the family of Bhagavantadeva. The genealogy is more or less mythical, but there are no weighty reasons to suppose that the verses are spurious and not from the pen of Nilakantha himself1 [The verses are: -- [x]. Vide also Aufrecht's Bod. Cat., p. 280. No. 656 and I.O. Cat. part III, p. 429, No. 1444 and Mandlik's Introduction LXXVII.]. The genealogy is: from Brahma was born Kas'yapa, whose son was Vibhandaka, whose son was Rsyas'rnga. In the family of the latter was born S'rngivara, after whom the family came to be known as Sengara. King Karna was born in that family. Then follows a line of eighteen kings, the last being Bhagavantadeva.
-- The Vyavaharamayukha or Bhatta Nilakantha, With an Introduction, Notes and Appendices, by P.V. Kane
A most interesting case is presented by the Agniº. Gyani23 [23 GYANI Agniº 1964: 1 n. l.] lists five editions which "do not much differ from one another," and he therefore restricts himself to using one edition: Venkatesvara. Yet, ever since the publication of the catalogue of Sanskrit manuscripts of the India Office Library23a [23a EGGELING 1899: 1294-1298, no. 3582.] the existence of a Vahniº -- "different from the work usually designated by that title" -- was known.24 [24 Cf. WINTERNITZ 1907: 473 n. 2; at 1963: 496 n. 2 it is called an upapurana. In fact, Buhler's Report for 1872-73 to the Director of Public Instruction mentions: "Among the puranas the Vahnipurana is new to me. It is not identical with the Agnipurana" (extract at IA 2, 1873, 304). Cf. WILSON 1840 = 1961: xxxviii.] Only at the end of his treatment of the Agniº Hazra25 [25 HAZRA 1940: 139-140.] briefly refers to this manuscript: "Besides the extant Agni-p., Mss. have been found of another work called 'Vahni-purana.'" A little over ten years later Hazra had an opportunity to study the manuscripts of the Vahniº; he concluded: "With the spread of Tantricism this spurious work [= the present AgniP] attained great popularity, and the genuine Agneya-purana [= the earlier Agni-p.] had to save itself from extinction by assuming a different title, viz., 'Vahni-purana.'"26 [26 Discovery of the Genuine Agneya-purana, JOIB 5, 1956, 411-416 at 411; also, Studies in the Genuine Agneya-purana alias Vahni-purana, OH 1, 1953, 209-245 [contents at 218-224]; 2, 1954, 77-110.] He rightly points out that what makes a mahapurana is its being well known; that what is less well known becomes an upapurana. "As modern scholars did not know the real nature of this 'Vahni-purana' occurring in Manuscripts, they took it to be an Upapurana of minor importance." Although Hazra twice uses the expression that it is the Vahniº rather than the Agniº that "is identical with" the genuine Agneyaº, the absence in it of the Isanakalpa -- here too, even as in the extant Agniº! -- and various other arguments make him admit that, once again, this Vahniº has lost many chapters of the original Agneyaº; they have been replaced by passages drawn from other sources, and the whole assumed a "Vaisnava form." In short, this is a unique case in which one "extant" purana has actually been replaced with another, at least by one scholar. Yet, in 1964 Gyani continued to analyze the Venkatesvara edition; his sole reference to Hazra's article appears in six lines of the conclusions.Vaishnavism is one of the major Hindu denominations along with Shaivism, Shaktism, and Smartism. According to a 2010 estimate by Johnson and Grim, the Vaishnava tradition is the largest group within Hinduism constituting about 641 million or 67.6% of Hindus. It is also called Vishnuism since it considers Vishnu as the Supreme Being. Its followers are called Vaishnavas or Vaishnavites (derived from IAST: Vaiṣṇava), and it also includes some other sub-traditions like Krishnaism and Ramaism, which consider Krishna and Rama as the Supreme Being respectively...
Key texts in Vaishnavism include the Vedas, the Upanishads, the Bhagavad Gita, the Pancaratra (Agama) texts, Naalayira Divya Prabhandham and the Bhagavata Purana.
-- Vaishnavism, by Wikipedia
Modern scholarship noticed all these facts. It recognized "that the extent of the genuine Agneya-purana was not the same at all times and in all places and that it varied with the difference in time and locality."27 [27 HAZRA, JOIB 5, 1956, 414-415.] It realized that of the hundreds of verses attributed to the Deviº, which are not traceable in our text, some were quoted by certain nibandhakaras, some by others. "This shows that the text of the Devi-p. was not the same everywhere but differed considerably in different provinces."28 [28 HAZRA 1963: 193-194.] Yet, one failed to draw the logical conclusion: besides the version or versions of puranas that appear in our manuscripts, and fewer still in our editions, there have been numerous other versions, under the same titles, but which either have remained unnoticed or have been irreparably lost.
The danger I want to point out here is that those readings and arrangements of a particular purana which happen to have been included in the printed edition or editions, have automatically been considered as representing "the" purana -- be it only the less valuable extant one --, whereas all variant readings and arrangements exhibited in manuscripts which were not used by the editors, have been generally overlooked or neglected. These latter materials never -- or very rarely -- come to the notice of those doing research on the puranas; and if they do, they are invariably treated as less valuable or negligible deviations from the standard text.
One scholarly publication in which, besides editions of puranas, manuscripts have been used quite extensively, is Pargiter's Purana Text of the Dynasties of the Kali Age. The author made a conscious search for as many manuscripts as he could possibly find (see his account, pp. xxix-xxxiv), and the variant readings found in them are carefully noted in the critical apparatus. Occasionally, however, even Pargiter cannot help giving more weight to the printed editions than to handwritten materials. For instance, in the Bhavisyaº the text on the dynasties of the Kali age occurs in the Venkatesvara edition, whereas "I have examined the following MSS, but none of them contain anything about these dynasties" (p. xxx); these manuscripts are ten in number, and Pargiter therefore concludes that the passage in the edition is "not genuine." A similar situation presents itself in the Brahmandaº: none of the five manuscripts which Pargiter consulted contains the passage as it occurs in the Venkatesvara edition which "professes to be based on several MSS, yet gives variant readings only rarely, and leaves on my mind the impression that it has been silently emended at times" (p. xxx). Yet, in this case Pargiter failed to draw the same conclusion he drew in connection with the Bhavisyaº.
No one has probably consulted as many purana manuscripts as R.C. [Rajendra Chandra] Hazra; no one was more acutely aware of the problems they raise than he was. Yet, although he points to the existence of a manuscript of the Vamanaº which "seems to differ much from our printed editions," he too, cannot help basing his discussion of the nature and contents of "the extant Purana" on the printed editions only.
Other scholars did not use manuscripts to the extent Pargiter or Hazra did.29 [29 HAZRA 1940: 76-77.] The manuscripts were, in most cases, inaccessible to them, or too voluminous to handle. In fact, most scholars who wrote on puranas did not even have access to all existing editions.30 [30 The editions in Telugu script are, in many cases, the earliest ones; yet they have remained practically unknown and unused. The editions in Bengali script, especially those of the Vangavasi Press, have been used more often, but in general only by Bengalis. Most purana studies, by Indians all over the subcontinent and by non-Indians, are based on the devanagari editions, especially the Venkatesvara and Anandasrama editions, the latter being even more widely available than the former. E. g., "Reference is made here only to the texts that have appeared in Devanagari" (PUSALKAR 1968: 692); "Unless otherwise indicated, the texts are those of the Anandasrama Series" (W. JAHN: Uber die kosmogonischen Grundanschauungen im Manava-Dharma-Sastram [Google translate: About the cosmogonic Basic views in the Manava-Dharma-Sastram], Leipzig: Drugulin, 1904, p. 20).] Their conclusions are based on those editions which happened to be available to them; in many cases this just meant one single edition. Kirfel's Purana Pancalaksana, for example, uses a single manuscript of the Brahmandaº, in addition to its Venkatesvara edition; for the Visnuº it uses one edition from Calcutta and one from Bombay; for the Padmaº it uses two editions from -- at that time -- Bombay Presidency; and for all other puranas it relies on single editions. Kirfel was aware of the problem; he confessed facing the alternative of making purana studies progress with incomplete materials, or postponing publication of his research indefinitely, and he decided in favor of the former.31 [ 31 "Wenn im Vorwort zum Puranalaksana gesagt wurde, dass zur Aufhellung mancher textlicher Verderbnis oder Abweichungen noch eineAnzahl alter und guter Handschriften hatte herangezogen werden mussen, so gilt dieses auch fur den vorliegenden Band. Es besteht aber auf absehbare Zeit nicht die geringste Aussicht, an jene heranzukommen. Aber selbst wenn dies der Fall ware, wurden sich doch nicht alle Textverderbnisse berichtigen lassen, wie die in Poona erscheinende kritische Ausgabe des Mahabharata beweist. Zudem wurde es sich bei den vielen hier in Betracht kommenden Purana's um eine so grosse Anzahl von Handschriften handeln, dass die Drucklegung des durch so umfangreiche Kollationierungsarbeiten ubermassig angeschwollenen kritischen Apparates schon aus praktischen Grunden scheitern wurde. Dies hiesse zugleich, die mit der riesigen purana Literatur verknupften Probleme auf unabsehbare Zeit zuruckzustellen; dies ware ein Verfahren, das mit den Prinzipien von Forschung und Wissenschaft nicht mehr recht vereinbar ist" [Google translate: If in the foreword to the Puranalaksana it was said that for enlightening some textual corruption or discrepancies still a number of old and good manuscripts had to be used, this also applies to the present volume. But there is no prospect of getting hold of them in the foreseeable future. But even if that were the case, not all text corruptions would be corrected as the critical edition of the Mahabharata published in Poona shows. In addition, it became one of the many Puranas under consideration here. Large number of manuscripts act that the printing of the by so extensive Collation work on excessively swollen critical apparatus already done practical reasons would fail. This would also mean the one with the huge purana literature postpone linked problems for an indefinite period of time; this would be a procedure that is no longer quite compatible with the principles of research and science." ] (KIRFEL 1954: XI-XII).] Abegg's Pretakalpa is based solely on Jibananda's edition of the Garuda -- the editions of which are very different indeed.32 [32 "Neben dieser in der Einleitung S. 2 allein angefuhrten Ausgabe ist noch eine von Pancanana Tarkaratna, Calcutta 1891 erschienen, sowie eine in Bombay; beide waren mir nicht zuganglich" (Pretakalpa des Garuda-Purana, 1921, p. VIII n. 1). [Google translate: In addition to this edition, which is only cited in the introduction on p. 2, there is also one of Pancanana Tarkaratna, Calcutta 1891, and one in Bombay; both were not accessible.] A study such as Hohenberger's "Das Vamanapurana"33 [33 HOHENBERGER Vamanaº 1963.] which uses the Venkatesvara edition of the text only, cannot claim to be more than a study of that particular version of the Vamanaº; it is not an analysis of the Vamanaº as a whole. Surabhi H. Trivedi34 [34 TRIVEDI Brahmaº 1968-69.] lists four editions of the Brahmaº (p. 75), but uses nothing else than the Anandasrama edition. V. S. Agrawala based his study of the Markandeyaº, in Hindi, on Jibananda's edition and Pargiter's English translation.35 [35 AGRAWALA Markandeyaº 1961.]
In short, the existing editions, useful as they may have been in making the puranas available in print, have done a definite disservice to scholarly research, in that they have accidentally raised one or two versions of each purana to the rank of the purana. By doing so they have obliterated all other versions which might be equally or, eventually, more important than the published ones.36 [36 E. g., Sylvain LEVI (1905: 210) noted the existence of not less than five versions of the Svayambhuº: Svayambhumahaº (12 chapters), Svayambhu-utpattikatha or Madhyamasvayambhuº (10 ch.), Brhatsvayambhuº (printed in the BI, 3000 lines in ms.), Mahatsvayambhuº (2000 lines), and Svayambhucaityabhattarakoddesa (250 lines). He considered the first one to be the best; as opposed to the printed one with its "style barbare et metrique abominable." "Il est regrettable que la Bibliotheca Indica ait imprime de preference cette derniere recension, et que l'editeur du texte ait cru devoir farcir a plaisir de barbarismes et de solecismes le sanscrit macaronique de son auteur; il n'est pas conforme au 'fair play" meme entre brahmanes et bouddhistes, de choisir, comme de parti pris, les lecons les plus incorrectes et d'eliminer les autres" [Google translate: "barbarian style and abominable metric." "It is unfortunate that the Bibliotheca Indica has preferably printed this last review, and that the editor of the text thought it his duty to stuff barbarisms and solecisms the macaronic Sanskrit of its author; he does not comply to 'fair play' even between Brahmans and Buddhists, to choose, as if by bias, the most incorrect lessons and eliminate the others."] (p. 212 n. 1).] As a result they have made purana research based on them one-sided and, therefore, inaccurate.37 [37 Cf. Albert B. LORD (The Singer of Tales, Cambridge: Harvard UP, 1960, pp. 124-125): even though one or more versions of oral poetry become eventually fixed, this does not in any way affect the singer or his audience. The singer "continued, as did his confreres, to compose and sing as he always had and they always had"; the audience "thought in his terms, in the terms of multiplicity."]
These facts have been recognized. Kane discusses the unfortunate situation that most editions of puranas are based on one manuscript or on a few manuscripts selected at random. "Many conclusions, therefore, drawn from the current printed editions of the Puranas or from mss. of the Puranas, must be regarded. as merely tentative and as likely to turn out to be wrong."38 [38 KANE 1962: 838.] Haraprasad Shastri used a presidential address at the All-India Oriental Conference to stress the fact that "in the matter of the Puranas every manuscript has a peculiar feature, and so, all manuscripts are important from the point of view of a collector and a scholar."39 [39 AIOC 5 (1928), 80. Cf. DAS (1924: 119): "One might almost say that no two manuscripts of any Purana are exactly the same."].
Yet, scholars continue to quote and draw conclusions from the Visnuº, the Matsyaº, etc. Sten Konow,40 [40 Note on the Andhra King Candasata, ZDMG 62, 1908, 591-592.] for example, while discussing the name of the last but one Andhra king, relies heavily on the puranas to defend the form Canda rather than Vincent Smith's Candra."The forms of the name given in the puranas are not at all in favour of this supposition. Candra, it is true, occurs in THE Visnu and Bhagavata Purana, but THE Matsya has Canda and THE Vayu Danda, and this last form cannot well be explained as a corruption of Candra. I therefore think that Canda is the correct form of the word."41 [41 Ibid., p. 591 (emphasis added). Konow's argument is even less convincing, since he draws these forms, second hand, from R. G. Bhandarkar's Early History of the Dekkan (2nd ed., 1895, p. 164), where the sources used for the puranas are not even specified.]
In connection with the different versions of puranas reference should also be made to titles of puranas to which adjectives such as Brhad-, Laghu-, Vrddha-, etc., have been prefixed. They have laid to confusion and misinterpretation.42 [42 The confusion existed already in the Sanskrit texts. For instance, the Ekamraº distinguishes between a principal Narasimhaº and an upapurana called Brhannarasimhaº. Gopalabhatta's Haribhaktivilasa also quotes verses separately from Narashimhaº (ca. 100) and Brhannarasimhaº (63). The Caturvargacintamani, on the other hand, ascribes the latter to Narasimhaº. HAZRA (1958: 356) concludes that either Hemadri considered the two works to be identical, or the verses occurred in both. They do not, however, appear in our printed Nrsimhaº.] Winternitz,43 [43 WINTERNITZ 1907: 466-467.] for example, confused the Naradaº and Brhannaradaº in the original, German Geschichte; the confusion has only been partly eliminated in its English translation.44 [44 WINTERNITZ 1963: 489-490.] Both in the German and English editions the Brhannaradaº is called "the great Purana of Narada," whereas the much longer Naradaº, the edition of which is not even mentioned, is called an upapurana. Similarly, when Winternitz45 [45 Ibid., p. 478; cf. HAZRA 1958: 353-356.] assigns an early date to the Visnuº because few mahatmyas [Māhātmya can be translated as “glory” or “greatness”, and is also a term for a text genre.] claim to be part of it, he adds the following note: "Nevertheless it is noteworthy that Matsya- and Bhagavata-Purana give the number of slokas of the Visnu-purana as 23,000, while in reality it has not quite 7,000 verses, and that also a 'Great Visnu-Purana' (Brhadvisnupurana, Aufrecht CC I, 591) is quoted." He thus seems to equate the larger version of 23,000 verses with the Brhadvisnuº. In reality, here as elsewhere in Sanskrit literature46 [46 Compare the many smrtis to which the same adjectives have been prefixed. On the status of such texts, see J.D.M. DERRETI: Dharmasastra and Juridical Literature, [part of vol. IV in this series,] Wiesbaden: Harrassowitz, 1973, p. 40.] we should expect Brhad-, Laghu-, Vrddha-, etc., to refer to compositions which are secondary as compared to the corresponding titles to which these adjectives have not been prefixed....