The Purana Text of the Dynasties of the Kali Age c2.s1: Prophetic Form of the Account

The Purāṇa Text of the Dynasties of the Kali Age c2.s1

Language: English

Prophetic Form of the Account

c2.s1.1

10. All these accounts profess to be prophetic, yet the standpoints from which these Purāṇas view these genealogies differ somewhat. The Viṣṇu professes to have been narrated by Parāśara to Maitreya, and sets out the Paurava genealogy from the standpoint of the reign of Abhimanyu’s son Parīkṣit, and the Aikṣvāku and Bārhadratha genealogies from the time of the great battle between the Pāṇḍavas and Kauravas⁴. This is absurd, because Parāśara was Vyāsa’s father and was dead long before that battle and Parīkṣit’s birth. All the other Purāṇas profess to have been recited by the Sūta to the ṛṣis in Naimiṣa forest and (except in the Garuḍa) at their twelve-year sacrifice⁵. The Vāyu fixes the time of that sacrifice as the reign of the Paurava king Asīmakṛṣṇa ¹, who is more often called Adhisīmakṛṣṇa ², and who was fourth in descent from Parīkṣit; and the Matsya and Vāyu say the same in nearly the same words when mentioning that king in this account of the Kali age ³. These two Purāṇas thus deal with these genealogies from the standpoint of his reign, and the Brahmāṇḍa, Bhāgavata, and Garuḍa constructively profess to do the same.

c2.s1.2

⁴ Cvṣ iv, 20, 12-13, and 21, 1 : also 22, 1 and 23, 1.

c2.s1.3

⁵ Mt 1, 4; AVā 1, 13-15; Bd i, 1, 17, 18, 35, 36; CGr 1, 3-11. They differ in the Sūta’s name.

c2.s1.4

¹ In its verse, 1, 12— Asīmakṛṣṇe vikrānte rājany an-upama-tviṣi praśāsatimām dharmena bhūmim bhūmipasattame. ² See p. 4, note ¹⁰. ³ A Mt 50, 66, 67; AVā 99, 258, 259. ⁴ See p. 4, l. 6. ⁵ See p. 10, l. 5, and p. 15, l. 13. ⁶ In equating these kings some 20 years must be prefixed to the Paurava list on account of Yudhiṣṭhira's reign after the great battle, before Parīkṣit came to the throne, see § 14. ⁷ GBh ix, 1, 6. ⁸ CGr 140, 40. ⁹ *E.g.* p. 10, note ²³; p. 11, l. 18. ¹⁰ *E.g.* p. 5, l. 11; p. 11, ll. 14, 21. ¹¹ *E.g.* *bhavet* for *abhavat*, p. 10, note ²³; p. 11, note ⁵¹. ¹² P. 12, l.

c2.s1.5

11. The Matsya and Vāyu carry out that view. They bring the Paurava genealogy from Abhimanyu and his son Parīkṣit down to Adhisīmakṛṣṇa as already past, and name Adhisīmakṛṣṇa as the reigning king ⁴; the ṛṣis then inquire about the Kali age, and the Sūta, declaring his intention to set out all the future kings, begins the list of future Pauravas from that monarch. Similarly, in the contemporary Aikṣvāku and Bārhadratha genealogies, these two Purāṇas name Divākara as reigning then in Ayodhyā and Senājit in Magadha ⁵, and mention their predecessors as past and their successors as future. Hence they virtually declare that these three kings were contemporary ⁶. The position taken in the Brahmāṇḍa is the same, though it is obscured by a large lacuna in which all the Paurava and Aikṣvāku kings are lost, and its account begins with line 23 on page 12. Thenceforward it agrees with the Matsya and Vāyu and mentions Senājit as the reigning Bārhadratha king. The Bhāgavata and Garuḍa, though professing to have been recited in Adhisīmakṛṣṇa's reign, take the former the standpoint of Parīkṣit's reign ⁷, and the latter that of his son Janamejaya ⁸; and both treat all the successors and also all the Aikṣvāku and Bārhadratha kings after the great battle as future. The Viṣṇu agrees with the Bhāgavata in this attitude, as already mentioned.

c2.s1.6

12. Accordingly the texts are framed for the most part in prophetic shape, but this character is not maintained completely because past expressions occur here and there, such as *abhavat* ⁹, *smṛta* ¹⁰, &c. Some MSS have tried to be more consistent by modifying such words ¹¹. One line found in three MSS frankly states that the whole Aikṣvāku dynasty was ancient, and naturally does not appear in any of the other MSS ¹². There can be no doubt therefore that the accounts have been steadily though slowly revised in details, so as to improve their prophetic character.

c2.s1.7

²⁶ It is no doubt genuine, for no one would be likely to fabricate and interpolate it to mar the prophecy.

c2.s1.8

13. Though the account is said to have been narrated to Paurava kings or to ṛishis in Naimiṣa forest, yet the ground from which the historic changes are viewed is Magadha. The Paurava and Aikṣvāku dynasties are dealt with briefly, with two kings generally to a line and with no mention of the lengths of the reigns, but the Bārhadratha dynasty of Magadha is set out with one line to each king and the length of his reign is stated¹. After those three ancient kingdoms disappeared, the dynasties treated of are those which reigned in or dominated Magadha. All other dynasties in North India are noticed only in the aggregate, with the exception of the dynasty of Vidiśā, and even that is described but cursorily (p. 49).

c2.s1.9

¹ The *Early Contemporary Dynasties* summarize all except the Māgadhas (p. 23).

c2.s1.10

14. The beginning of the Kali age has been discussed by Dr. Fleet, and he has pointed out that it began on the day on which Kṛṣṇa died, which the chronology of the Mahābhārata places, as he shows, some twenty years after the great battle, and that it was then that Yudhiṣṭhira abdicated and Parīkṣit began to reign². But, as shown above, these Purāṇas virtually begin the Kali age dynasties immediately after the battle, and that position is the most convenient to adopt for the present purpose. The text of the Matsya and Vāyu³ can be brought into harmony therewith by merely altering the order of a few verses without tampering with them, namely, by transposing the four verses containing the ṛishis' questions and the prefatory verses of the Sūta's reply from their position in Adhiśīmakṛṣṇa's reign to the commencement of the account; and, so treated, those verses form a fitting preface to the whole: but it is unnecessary to print the questions here, and those prefatory verses are alone introduced as a sufficient preface (see p. 1).

c2.s1.11

² JRAS, 1911, pp. 479, 675, 686; and p. 62, l. 37 *infra*. Hence in equating the Paurava kings with the Aikṣvāku and Bār-hadratha kings, some 20 years must be prefixed to the former.

c2.s1.12

³ This portion in the Brahmāṇḍa is lost in the lacuna, as already mentioned.

c2.s1.13

*Original Language of the Account.*

c2.s1.14

15. There are clear indications that the Sanskrit account as it exists in the Matsya, Vāyu, and Brahmāṇḍa was originally in Prakrit, or, more accurately, that it is a Sanskritized version of older Prakrit ślokas. The indications are these: *first*, certain passages as they stand now in Sanskrit violate the śloka metre, whereas in Prakrit form they would comply with the metre; *secondly*, certain Prakrit words actually occur, especially where they are required by the metre, which the corresponding Sanskrit forms would violate; *thirdly*, Sanskrit words occur at times in defiance of syntax, whereas the corresponding Prakrit forms would make the construction correct; *fourthly*, mistaken Sanskritizations of names; *fifthly*, the copious use of expletive particles; and *sixthly*, irregular sandhi.

c2.s1.15

16. A full examination of these peculiarities would overload this Introduction, and the proof of them has therefore been set out in Appendix I. The above conclusion holds good for the whole of the text of the Matsya, Vāyu, and Brahmāṇḍa;

c2.s1.16

their verses are older Prakrit ślokas Sanskritized. It also holds good for such portions of the Viṣṇu and Bhāgavata as have preserved the old verses; but the main portions of these two Purāṇas are condensed redactions composed directly in Sanskrit. The Garuḍa version is a more concise condensation composed directly in Sanskrit apparently. These conclusions are explained in Appendix I.

c2.s1.17

17. Judging from such specimens of old ślokas and Prakritisms as have survived, it would appear that the Prakrit used in the original ślokas was a literary language not far removed from Sanskrit ¹. The art of writing was introduced into India some seven centuries B.C., and there can be no doubt that it must have been adopted early in the Courts because of its manifest administrative usefulness. Records must have been kept by secretaries and chroniclers in the royal offices, and as those men would not always have been Sanskrit scholars, the language they used would presumably have been as elegant a Prakrit as their courtly surroundings and predilections required. There must have been ample written material concerning the dynasties from the 7th century B.C. from which metrical chronicles could have been composed by bards, minstrels, and reciters ² in the same kind of language, to entertain not only their royal and noble patrons but also all those who found an interest in hearing of former times ³. As Magadha was a great, if not the chief, centre of political activity during those ages, we can perceive how it was that the account grew up with Magadha as its centre (§ 13). The Māgadhas were celebrated as minstrels, and since traditions are most easily remembered, are best handed down, and confer the greatest pleasure, when cast into poetical form, it is easy to understand how this metrical account of the dynasties in literary Prakrit could have developed among them. Hence we may infer that the original ślokas were composed in Māgadhī; or, since the account, much as we have it now, was compiled and edited apparently in North India ⁴, and one verse that the Bhāgavata has preserved is in Pali ⁵, they may have been in Pali, either originally or perhaps more probably by conversion.

c2.s1.18

¹ Pali is such a language, and other specimens are found in the early inscriptions.

c2.s1.19

² Sūtas, māgadhas, and vandins; and other professional singers.

c2.s1.20

³ Such men have existed in India from early times, and a graphic account of them, their methods, popularity, and influence, will be found in Babu Dinesh Chandra Sen's excellent 'History of Bengali Language and Literature', pp. 162-7, 584-5, 588-90. Since the brahmans could and did transmit the Vedic hymns with verbal accuracy for many hundreds of years, there is no improbability in supposing that bards and minstrels could hand down metrical accounts of dynasties with substantial though not with verbal accuracy. As these bards and minstrels existed in all parts of North India, they were a check on one another in the transmission of tradition, and there are indications that the Purāṇic traditions of the dynastic genealogies were compiled with some attempt to ascertain the truth. Moreover there was no objection to the accounts being written down, as soon as writing came into general use; and that would have been also a check on variation.

c2.s1.21

⁴ See § 27.

c2.s1.22

⁵ See Appendix I, § ii. Certain other words mentioned in Appendix I appear to be Pali.

c2.s1.23

18. The account supplies two kinds of internal evidence to fix the time when it was compiled, namely, *first*, the subject matter, and *secondly*, textual peculiarities; and both are important. The latter are dealt with in § 26, and the former is discussed first. The subject matter consists of two parts, the earlier setting out the dynastic details, and the later part describing the unhappy conditions that should prevail and stating certain chronological and astronomical particulars¹. These are treated here separately.

c2.s1.24

¹ The earlier part pp. 1-55, and the later pp. 55 ff.

c2.s1.25

19. The dynastic portion shows two stages of termination. The earlier of these stages is the period following the downfall of the Andhras and the local kingdoms that survived them a while. The Matsya account ends here with the mere mention of the Kilakila kings², and no MS of the Matsya contains anything later. The Andhra kingdom fell about A. D. 236, and it may be said that the Matsya account brings the historical narrative down to about the middle of the third century A. D. and no further.

c2.s1.26

² That is l. 15 on p. 48. The Vṣ says they were Yavanas, see note 82 thereto.

c2.s1.27

20. The Vāyu, Brahmāṇḍa, Viṣṇu, and Bhāgavata all carry the narrative on to the rise of the Guptas, which is the later stage. The Guptas are mentioned as reigning over the country comprised within Prayāga, Sāketa (Ayodhyā), and Magadha, that is, exactly the territory which was possessed at his death by Candragupta I who founded the Gupta dynasty in A. D. 319–20 and reigned till 326 or 330 (or even till 335 perhaps), before it was extended by the conquests of his son and successor Samudragupta. With the Guptas are mentioned Nāgas, Maṇidhānyas, and others as reigning contemporaneously over the countries which surrounded the Gupta territory³ and which were subjugated afterwards by Samudragupta⁴. The account takes no notice of his conquests nor of the Gupta empire. These particulars show clearly that this account was closed during the interval which elapsed between the time when Candragupta I established his kingdom from Magadha over Tirhut, Bihar, and Oudh as far as Allahabad⁵, and the beginning of Samudragupta’s reign, for he began his conquests immediately after his accession. That interval is approximately A. D. 320–330 or perhaps 335. It is hardly credible that, if this account was compiled later, it would have omitted to notice Samudragupta’s conquests, or would have mentioned the foregoing kingdoms (which he subdued) in the same terms as his kingdom. The Gupta era was established in A. D. 320, and it may be concluded that this account was closed soon after the commencement of that era, or, if we allow some margin for delay, by the year A. D. 335.

c2.s1.28

³ See pp. 53-5.

c2.s1.29

⁴ V. Smith's History, 2nd edn. pp. 267-9; and JRAS, 1909, p. 342.

c2.s1.30

⁵ V. Smith's History, p. 266.

c2.s1.31

21. Hence it appears that the versified chronicles were first collected about or soon after the middle of the 3rd century ¹ in the shape found in the Matsya, and that they were extended to the rise of the Gupta kingdom before the year 335, which augmented compilation is what the Vāyu and Brahmāṇḍa contain and the Viṣṇu and Bhāgavata have condensed. It has been shown that the Matsya, Vāyu, and Brahmāṇḍa all obtained their accounts from the Bhaviṣya. Hence it would appear that the earlier compilation must have been incorporated in the Bhaviṣya about or soon after the middle of the 3rd century, and that its prophetic account was extended in the later compilation before the year 335. There is nothing improbable in this augmentation, because the Bhaviṣya account has been continually supplemented even up to the present time in order to keep its prophecies up to date, as shown above (§ 6). It follows then that the Bhaviṣya must have been in existence in the middle of the 3rd century ²; and it would appear that the Matsya borrowed what the Bhaviṣya contained before the Gupta era, and that the Vāyu and Brahmāṇḍa borrowed the Bhaviṣya's augmented account about or soon after the year 330 or 335. Further remarks on these dates are offered in §§ 43 ff.

c2.s1.32

¹ There is an apparent indication that a compilation was begun in the latter part of the 2nd century in the Andhra king Yajñaśrī's reign, for 5 MSS of the Matsya (of which three appear to be independent, namely, *b*, *c*, and *l*) speak of him as reigning in his ninth or tenth year; see p. 12, note ⁸. If so, the Bhaviṣya may perhaps have existed in that century.

c2.s1.33

² But not of course in its present condition.

c2.s1.34

22. Further light is thrown on these points by the MS *e*Vāyu, which contains the full account but holds a position intermediate between the general Vāyu version and the Matsya version. The facts to be explained are these. The Matsya has one version which contains only the shorter compilation, the Vāyu generally has a somewhat different version containing the full account, the Brahmāṇḍa has the full compilation in a version resembling the Vāyu closely, *e*Vāyu has a version containing the full compilation in a text intermediate between the Matsya and all other copies of the Vāyu ³, and yet all these Purāṇas declare they borrowed their accounts from the Bhaviṣya.

c2.s1.35

³ The position of *e*Vā is best shown in the account of the Mauryas, pp. 27–9.

c2.s1.36

23. The only theory which appears to me to explain all these facts is this. The Matsya borrowed from the Bhaviṣya the shorter account about (say) the last quarter of the 3rd century. The Bhaviṣya account was then extended down to the time when the Gupta kingdom had acquired the territories assigned to it, and its language was revised ⁴; that would be (say) about 320–325. The Vāyu copied that extended and revised account from the Bhaviṣya almost immediately, and that is the version found in *e*Vāyu. Afterwards, the language of the Bhaviṣya version was revised again, and this must have been done very soon, (say) about 330–335, before the Gupta kingdom had developed into the Gupta empire by Samudragupta's conquests, because it could hardly have failed to notice that immense change if the revision had been later. This second revision was soon adopted by the Vāyu and is the version found now in Vāyu MSS generally. The fact that *e*Vāyu stands unique among all the Vāyu MSS suggests that no long interval could have separated the second revision from the first, and that the first revised version was quickly superseded by the second in the Vāyu. I cannot speak about the Brahmāṇḍa in any detail, because I have not been able to collate any MSS of it: yet two points may be noticed, *first*, it agrees closely with the general Vāyu version ¹ and yet condenses the account sometimes ²; and *secondly*, the probability is that it borrowed the second revised version from the Bhaviṣya not long after the Vāyu adopted that ³. The Bhaviṣya existed in writing when the first revision appeared in it, because *e*Vāyu, as well as all other Vāyu MSS, uses the word *paṭhita* when acknowledging its indebtedness to the Bhaviṣya (see § 7). The Matsya uses the word *kathita* in the corresponding passage, which might imply that it borrowed the account orally at the earlier stage, but that is not probable because of the inferences brought out in Appendix II.

c2.s1.37

⁴ This, as already pointed out, is what has been habitually done to it.

c2.s1.38

¹ The agreement is not only here, but large portions also of the Brahmāṇḍa are almost identical with the Vāyu.

c2.s1.39

² As in p. 22, note 46; p. 35, note 42. In those passages the Bhāgavata partially resembles it, and may have copied from it.

c2.s1.40

³ Unless (what is possible) the Brahmāṇḍa copied its account from the Vāyu (see note 1); and its paraphrase of *Aśoka-vardhanaḥ* as *aśokānāṃ ca tṛpti-daḥ*, if not a late attempted emendation of a text that was unintelligible, suggests that it could not have been composed until Aśoka was wholly forgotten.

c2.s1.41

24. If this explanation be tenable, the Matsya version of these dynastics of the Kali age is older than those of the Vāyu and Brahmāṇḍa ⁴, and *e*Vāyu gives us the earliest text of the Vāyu. The styles of the versions appear to support this explanation, for the Matsya version is somewhat crude at times, and the Vāyu text has been revised more than the Matsya as shown by the story of king Janamejaya’s dispute with the brāhmaṇs ⁵. Though later than the Matsya, the Vāyu account may yet be more accurate at times by reason of the revision which it underwent ⁶. The Vāyu has Prakritisms sometimes where the Matsya has correct Sanskrit ⁷, but this fact is not incompatible with that conclusion, and for either or both of two reasons; (1) the Matsya may have emended such defects at the time of taking the account from the Bhaviṣya, while the Vāyu may have copied them as they stood; and (2) a process of silent emendation has been in continual operation in the MSS ⁸. Further it would seem that the three accounts may have been compared at times, for this would explain certain small variations which appear occasionally between the Vāyu and Brahmāṇḍa in the direction of the Matsya ⁹.

c2.s1.42

⁴ I differ therefore from Sir R. G. Bhandarkar, who estimated (without giving reasons) the Vāyu account to be older than the Matsya; but agree with him that the Viṣṇu is later and the Bhāgavata the latest: *Early History of the Dekhan*, 1895, p. 162. In all this discussion I am dealing only with the time when these accounts of the dynasties of the Kali age were incorporated in these Purāṇas, and not with the age of these Purāṇas themselves such as they were in that early period; see § 28, note.

c2.s1.43

⁵ See Appendix III.

c2.s1.44

⁶ As in the arrangement of verses (see pp. 27, 44), and in many of the readings in the concluding portion (pp. 55 ff).

c2.s1.45

⁷ See Appendix I, § iii, first instance.

c2.s1.46

⁸ *E.g.* p. 18, note 7; see Appendix I, § ii.

c2.s1.47

⁹ These conclusions do not imply that these Purāṇas existed then in their present

c2.s1.48

• 25. The second portion of the account referred to in § 18 consists of (1) an exposition of the evils of the Kali age, and (2) a chronological-astronomical summary of the age, and is found in the Matsya, Vāyu, and Brahmāṇḍa. This second portion therefore existed in the earliest version compiled soon after the middle of the 3rd century, yet with a difference. While the Matsya has a good deal of the exposition, the Vāyu and Brahmāṇḍa version contains some 32 more lines and is nearly twice as long as the Matsya; so that a large addition was made at the revision, and it was made mostly at the first revision, because the account in Vāyu has the full description with the exception of a few verses which may have been omitted by oversight. As regards the chronological-astronomical summary however, all three Purāṇas practically agree, the Matsya wanting only two lines. These particulars therefore were complete in the first compilation and were not added to in the revisions; and this conclusion is corroborated by the fact that this summary in all three Purāṇas brings the reckoning down definitely only to the end of the Andhras, and uses the vague term *Andhr-ānt-ādyās* in referring to future kings¹. No addition was therefore made to it at the revisions to bring it down to the Gupta era. It belongs then to the middle of the 3rd century and must be interpreted accordingly; and it shows that the *Saptarṣi* cycle of 2700 years was known and was in use in India at that time, that is, about three centuries earlier than has been supposed². The treatment of these two subjects, the evils of the Kali age and the chronological-astronomical particulars, affords an excellent illustration of what the revisers did and did not do. They had no knowledge with which to augment or alter those particulars and so left them unmodified; but the deterioration of the Kali age was a subject congenial to pessimistic brahmanic views and they freely availed themselves of the opportunity of dilating upon it.

c2.s1.49

¹ P. 58, ll. 9, 12, and p. 61, l. 23.

c2.s1.50

² See Encycl. Brit., ‘Hindu Chronology’.

c2.s1.51

26. I come now to the subject of textual peculiarities mentioned in § 18. Further information may be discovered by examining the divergent readings of the same passage and especially the corruptions in names. A study of the variations shows that ordinarily the copyists copied what they found in dull good faith to the best of their ability, often writing the same name differently in contiguous lines³. Moreover these dynasties of Śūdras and foreigners offered little inducement to readers to alter the texts. Hence the variations that crept in were mostly due to clerical blunders or to misreadings of the MSS copied; and the mistake might be detected and corrected, or might not. If not detected, the erroneous letter remained; if detected, the correct letter was written or inserted, and the incorrect letter was sometimes cancelled but was not seldom left uncancelled. In that state shape. They have no doubt been freely added to since, see § 28, note.

c2.s1.52

³ Cf. the corruptions in the well-known names, *Kauśāmbī* (p. 5, note 19) and *Kāṇvāyana* (p. 34, note 11). Yet sometimes errors were caused by a droll perversity or would-be cleverness, cf. p. 11, note 30; p. 42, note 1; and p. 17, note 71: and sometimes where the text had become corrupt, it was boldly paraphrased afresh, cf. p. 26, note 42; p. 33, note 52; but the latter was probably the effort of a reader and not of a copyist.

c2.s1.53

the passage was repeated in subsequent copies, and misreadings are important chronologically if we can explain how they arose. If their divergent readings of the same name or passage be written in the ancient scripts, and resemble one another so closely in a particular script that an ordinary copyist might easily misread one for another, it may be inferred that the variation must have arisen out of a MS written in that script, and therefore that the text once existed in that script, that is, it had been written during the time when that script was in use. In this way it may be ascertained which are ancient and which are mediaeval or even modern corruptions. Most of the variations have arisen from misreadings of the Gupta and later scripts¹, but for the present purpose it is unnecessary to consider any that arose from misreading scripts that came into use after A.D. 330, the date when this account was finally compiled, and it is only essential to see whether any variations point to misreadings of Kharoṣṭhī or of Brāhmī.

c2.s1.54

¹ Many such may be detected in the notes, n. r and v, c and v, &c. such as mistakes of p and y, n and r, l and

c2.s1.55

27. It would overload this Introduction to examine such particulars here, and in Appendix II are noticed such cases as appear to throw light on this subject. It is shown there that errors are found in the Matsya, Vāyu, and Viṣṇu which point to misreadings of Kharoṣṭhī as their source. Hence it seems there is reasonable ground for inferring that this account of the dynasties was, in its earliest form, written in Sanskrit in Kharoṣṭhī, and, since Kharoṣṭhī was current only in Upper India, that the account was probably put together there: that is, since the earliest account was in the Bhaviṣya, that the Bhaviṣya account was written originally in Kharoṣṭhī and was put together in Upper India. If these conclusions are sound, it would follow that the account could not have been compiled later than about A.D. 330, because Kharoṣṭhī went out of use about that time. Further, judging from the point of view displayed in the portion which was added to the Bhaviṣya to bring it up to date about the year 320², it would seem that the composers of this portion were probably in Madhyadeśa, and more particularly perhaps in the country between Magadha and Mathurā. The Viṣṇu account was probably based on the same original for three reasons: (1) its dynastic matter agrees closely with that in the Vāyu and Brahmāṇḍa, and also the ślokas where it has preserved them; (2) it closes its account where they close theirs; and (3) it is not probable that its account was a new and independent compilation from early chronicles when the compilations in the Bhaviṣya, Matsya, and Vāyu were available. At the same time it was composed early enough for its account to be drawn from Kharoṣṭhī MSS. It seems probable then that the main part of the Viṣṇu which is in prose was composed from those Purāṇas directly in Sanskrit not very long after the Gupta era, (say) perhaps before the end of the 4th century. The Bhāgavata was, as shown in Appendix II, composed afresh in Sanskrit, except in so far as it has incorporated old ślokas; and must have been based on the same materials for the same three reasons mentioned above, yet most probably on the Viṣṇu chiefly, to which it has the closest resemblances ¹; and it was probably not composed till the 8th century or even later. These conclusions strictly refer only to these dynastic accounts.

c2.s1.56

² See §§ 19-21.

c2.s1.57

¹ See p. 18, note ⁷; p. 25, notes ⁶, ¹⁶, ²²; p. 28, note ³⁰; and in its description of the evils of the Kali age, where the Viṣṇu and it have matter peculiar to themselves. It has resemblances to the Brahmāṇḍa in p. 22, note ⁴⁶; p. 35, note ⁴²; p. 41, note ⁸⁰.

c2.s1.58

*Sanskritization of the Account.*

c2.s1.59

28. It has been shown that the account was first compiled for the Bhaviṣya Purāṇa about the middle of the 3rd century A.D., and there are reasons why that was appropriate. Since royal genealogies constituted one of the subjects which every Purāṇa should treat of, the Bhaviṣya, as a work professing to deal with the future, could hardly ignore the dynasties that reigned after his time; and the dynasties of the Kali age would hold the same position in it that the ancient genealogies held in the Purāṇas which dealt with ancient stories ². The account of these dynasties would then naturally have been required for the Bhaviṣya, and all that was necessary was to collect the Prakrit metrical chronicles and convert them into Sanskrit prophecies uttered by Vyāsa ¹. That was done as shown in Appendix I, and then the Matsya first, and the Vāyu and Brahmāṇḍa afterwards, borrowed the account from the Bhaviṣya. The original Bhaviṣya account has been lost, but these three Purāṇas have preserved and reveal what its contents were; otherwise it would have been impossible to know what it contained at that time. A comparison of their accounts with the present condition of the Bhaviṣya shows to what bold lengths pious fraud has gone.

c2.s1.60

² The title Purāṇa indicates that such works narrated ancient stories, but the Bhaviṣya professed by its name to treat of the future, and the title Bhaviṣya Purāṇa is a contradiction in terms. Such a name could hardly have been possible, until the title Purāṇa had become so thoroughly specialized as to have lost its old meaning and become the designation of the kind of works now known by this title. The name Bhaviṣya Purāṇa therefore proves that the kind of composition that passed under the title Purāṇa had become stereotyped before the title could have been assumed by the Bhaviṣya; that is, that genuine Purāṇas must have preceded it so long before as to have specialized the title Purāṇa. It has been shown above that the Bhaviṣya existed in the middle of the 3rd century, hence some at least, if not many, of the true Purāṇas must be considerably older. This inference does not, of course, mean that the Purāṇas contained at their beginning all that they contain now, because there can be no doubt that they have been freely added to since. It is highly probable that they consisted at first mainly of ancient stories, genealogies, ballads, &c., which formed the popular side of ancient literature, and were quite probably in Prakrit originally. In fact, it seems to me that they were largely in an old literary Prakrit used by the higher classes, but that, as the spoken languages diverged in time more and more from Sanskrit through political vicissitudes, that literary Prakrit became unintelligible, while Sanskrit remained the only polished language of brahmanic Hinduism. Hence it was natural that this literature should be Sanskritized, if it was to be preserved, a process that was not difficult because the old literary Prakrit was not far removed from Sanskrit, yet it was not always effected completely, especially in poetry where the necessity of preserving the metre sometimes qualified that process, and hence Prakrit forms might survive embedded in good Sanskrit as *prarartayitrā* in p. 88, l. 14. It was the brahmans probably who saved and improved the status of those old compositions by converting them into Sanskrit, and afterwards, perceiving what an excellent means they provided for reaching popular thought, made use of them to propagate their own views and doctrines by freely augmenting them with brahmanical fables, philosophical discussions, and ceremonial expositions which were enforced with the authority of Vyāsa. I should say therefore, speaking generally, that what may be called the kṣatriya, or better perhaps the popular, matter of the Purāṇas constituted the really old and genuine *purāṇa*, and that the brahmanical and ritual matters now found in them were later additions and interpolations made from time to time. This inference is based on the fact that it is in the former portion of the Purāṇas that peculiarities occur such as are

c2.s1.61

¹ This was, as has been pointed out above, the beginning of a pious fraud, whereby the prophetic matter has been continually revised and brought up to date in the Bhaviṣya. To be able to point to such prophetic accounts in the literature would have been a valuable weapon, moreover, in the hands of the brahmans against adversaries of other creeds; and it may be noted in this connexion, that the Veṅkaṭeśvara edition of the Bhaviṣya has incorporated a summary of the Biblical account from Adam to Abraham in the early chapters of Genesis (Bhav. iii, 4, 17-19, 29-60; 5, 1-20). There can hardly be any doubt that this interpolation has been made very recently in view of Christianity.

c2.s1.62

29. Since the chronicles existed in the form of ślokas in literary Prakrit, all that was necessary was (1) to convert the Prakrit words into Sanskrit, and (2) substitute futures for past tenses, while maintaining the śloka metre. The first process appears to have been made word by word as nearly as possible ², and the Sanskritization was crude as the many Prakritisms noticed in Appendix I indicate, for they must have existed in the Bhaviṣya account, otherwise it is difficult to see how they could appear in the Matsya, Vāyu, and Brahmāṇḍa. Indeed it would almost seem that the Bhaviṣya account may have been composed in a literary Prakrit rather than in true Sanskrit. Both processes of conversion would have upset the metre, since Prakrit words are sometimes a syllable longer or shorter than their Sanskrit equivalents, and future tenses are generally longer than past tenses; hence three correctives were adopted; (1) words were dropped which might be omitted without impairing the sense, such as ‘reigned’, ‘years’, &c.; (2) compensatory expletives were inserted; and (3) the sentence was occasionally recast ³. Still the Sanskritization was imperfect and sometimes grammar or metre was sacrificed, and these blemishes have persisted, as pointed out in Appendix I, in spite of attempts to rectify them afterwards.

c2.s1.63

² See the phrase aṣṭaviṃśati tathā varṣā in Appendix I, § i.

c2.s1.64

³ Cf. for instance the lines in the Andhras where the two versions are given, and the notes thereto.

c2.s1.65

noticed in Appendix I. It seems highly probable too that it was largely through the Purāṇic literature, that brahmanism re-established itself over the people and secured the revival of Hinduism and the downfall of Buddhism. That was what actually happened in Bengal and has been called by Babu Dinesh Chandra Sen the ‘Pauranik Renaissance’, which he has described very clearly in his excellent work ‘The History of Bengali Language and Literature’ (ch. iv).

c2.s1.66

*Errors, omissions, and rare verses.*

c2.s1.67

30. Though there was originally one text common (but qualified by the revisions suggested in § 23) to the Matsya, Vāyu, and Brahmāṇḍa down to the end of the Andhras, yet present MSS show many errors and omissions and some misplacements. Such defects easily occurred through the carelessness of copyists¹, damage to² or loss of³ leaves, or disarrangement of leaves⁴. The blemishes in the text appear to have been generally accidental. The brahmans who compiled the Sanskrit account could and did fabricate passages portraying the evils of the Kali age, but had neither inclination nor incentive to invent particular dynasties or kings of foreign or base origin. The chief changes that can be placed under the head of fabrications are various attempts by later readers to improve the text in details in which it appeared to be corrupt or inelegant⁵, or to remove inconsistencies⁶. Among the latter some alterations, though made apparently in good faith, involved tampering with the text, as in the Śiśunāga dynasty, where the Matsya, by mistakenly introducing the first two Kāṇvāyana kings, names twelve kings instead of ten as all the other authorities declare; so that some copies of the Matsya have boldly altered the total to twelve, while others more cautiously have made the passage indefinite⁷. Misreadings have also produced incorrect statements and there are many errors in names and numbers⁸; but of deliberate falsification I have found no instance except in the story of the dispute between Janamejaya and the brahmans⁹.

c2.s1.68

¹ As for instance the mistaken introduction of the first two Kāṇvāyana kings among the Śiśunāgas in the Matsya (see p. 21 and note ^24).

c2.s1.69

² Damage probably explains the frequent loss of verses here and there in different MSS.

c2.s1.70

³ Hence no doubt the absence of all the first part in the Brahmāṇḍa (see pp. 1, 3, 8).

c2.s1.71

⁴ As for instance the displacement in eVāyu of the last half of the Early Contemporary Dynasties, all the Nandas, Mauryas, Śuṅgas, and Kāṇvāyanas and the first twelve lines of the Andhras after Viśvasphāṇi (see pp. 23, 24, 27, 30, 33, 35, 50).

c2.s1.72

⁵ See p. 26, note ^42; p. 33, note ^52; p. 52, notes ^37, ^38, ^42.

c2.s1.73

⁶ As in p. 29, note ^32.

c2.s1.74

⁷ See p. 22, note ^43.

c2.s1.75

⁸ As where the Bh misread *trayodaśa* as *bhūyo daśa*, p. 46, note ^22. As regards the readings *Tuṣāra* and *Tukhāra* in pp. 45, 47, it may be noted that *ṣ* has often been pronounced *kh* for centuries in North India, and that the letter *ṣ* was used at times for *kh*; hence these two letters are often confused: cf. p. 6, notes ^32, ^43; p. 19, note ^29; p. 41, note ^50; p. 51, note ^24; &c.

c2.s1.76

⁹ See Appendix III.

c2.s1.77

31. It is reasonably certain, then, that in the main these versions have suffered from nothing but carelessness and accident, and considering what little interest this account could have for educated readers, especially those brahmanically-minded, the text has been fairly well preserved. Much may have been lost altogether, for some passages have almost disappeared. One Purāṇa, or even one MS only, has preserved a passage or verse sometimes which is wanting in all the rest: thus eVāyu, alone of all the Matsya, Vāyu, and Brahmāṇḍa MSS, contains the verse about Śāliśūka, and his existence might be doubted if it depended on that alone, but it is testified to by the Viṣṇu and Bhāgavata¹. Again in the Bhāgavata only one copy has preserved the verse about Suśarman². Such being the conditions, no verse should be discarded even if it is found in only one MS. Thus line 26 of the Aikṣvākus appears only in three MSS, and lines 12–14 of the Preface only in *Vāyu*; yet it is not credible that they were fabricated, and they might easily have been regarded as valueless in the other MSS, for the former contradicts the alleged prophetic standpoint, and the latter merely name sundry and some unknown dynasties. Such rare passages appear to be relics of genuine tradition; and it is possible that lines 30, 31 of the Bārhadrathas found only in *Matsya*, and line 28 of the Andhras found only in *Vāyu*, may be genuine. Other peculiar verses will be found in the notes³.

c2.s1.78

¹ P. 29, l. 10. He is also mentioned in the Gārgīsaṃhitā according to Max Müller in *India: what can it teach us?* (ed. 1883) p. 298; but the passage is spurious, see JRAS, 1912, pp. 792–3.

c2.s1.79

² See p. 34, note 19. As p. 40, l. 13; p. 42, note 8.

c2.s1.80

³ [No text provided for this footnote in the source image]

c2.s1.81

*Formation of this Text.*

c2.s1.82

32. The Bhaviṣya account having been the common source of the Matsya, Vāyu, and Brahmāṇḍa versions, the various readings are often equivalent or not materially different, so that real divergencies are far fewer than the places where the readings vary. The text now offered has been prepared according to the printed editions and the MSS collated. The Matsya and Vāyu versions are of far greater value than the Brahmāṇḍa, because they have been printed at Calcutta and in the Ānandāśrama series from a number of MSS, and I have collated besides 13 MSS of the Matsya and 11 of the Vāyu; whereas of the Brahmāṇḍa only the Veṅkaṭeśvara edition has been available⁴, and I have seen no MSS containing this account. Where variations occur I have endeavoured to choose the most weighty, it being remembered (1) that the Matsya gives us the oldest version, *Vāyu* the next, and all other copies of the Vāyu and the Brahmāṇḍa the third recension; and (2) that the Matsya is at times a somewhat crude Sanskritization of the old Prakrit ślokas, and the later versions may be more accurate. Their general agreement must be understood, but variations and omissions are always noticed, so that where no notes are given, the copies all agree.

c2.s1.83

⁴ See *List of Authorities*, Brahmāṇḍa Purāṇa.

c2.s1.84

33. The Viṣṇu and Bhāgavata cannot elucidate that common version except in the occasional passages where they adhere to it; and there they have been used to frame the text. Otherwise they can only help towards determining the correct names of the kings and the duration of the dynasties, and are so utilized in the notes. The Bhāgavata is also useful in determining the order of the kings, because, while the single lines devoted to individual kings might be and have been displaced at times in the Matsya, Vāyu, and Brahmāṇḍa, its versified lists preclude the shifting of names in a verse, and the disarrangement of lines would produce manifest disarrangement of groups of kings. Besides the Calcutta edition of the Viṣṇu and the Gaṇpat edition of the Bhāgavata, I have collated 10 MSS of the former and 18 of the latter, and also the French edition of the Bhāgavata so far as its Sanskrit text goes.

c2.s1.85

34. The Garuḍa is of use only for the names of the kings in the three earliest dynasties, and I have been able to collate only the Calcutta edition and two MSS. The Veṅkaṭeśvara edition of the Bhaviṣya is of no value as already explained.

c2.s1.86

35. As regards variations in words, these when small, such as errors in sandhi¹, or optional ways of writing², or obvious clerical mistakes³, or mere trivial differences⁴, are generally disregarded or corrected unless there is something noteworthy in them⁵, for many of the MSS are carelessly written and abound in such blemishes; yet the notes will show that I have erred probably rather on the side of inclusion than of exclusion. Various letters are often written so much alike in the MSS as to be easily confused, such as *b* and *v*, *p* and *y*, *c* and *v*, *n* and *l*, *n* and *r*, subscript *r* and *u*, and the MSS often contain superfluous letters written by mistake and not cancelled. These flaws, when obviously purely clerical, have been disregarded in some cases, but otherwise, and especially where these particulars may prove significant in the matter of Prakritisms and scripts, have been cited in the notes as they stand, the superfluous letters being enclosed in square brackets. *B* and *v* when not distinguished in the MSS have been generally transcribed as they should be correctly, unless the actual letter seemed worthy of notice. Since the account is only a Sanskritized version of Prakrit ślokas, Prakrit forms have been admitted into the text if they are supported by the best authority, as more truly representing the original words especially in numerals. Variations of readings and corruptions of names have been arranged in the order of modification, so as to elucidate as far as possible the process of the changes, and when so placed, readings that are corrupt often prove to be highly instructive as regards both language and script⁶.

c2.s1.87

¹ *E.g.*, in p. 60, note 63 *dītvā* have *śatāt śatam* actually.

c2.s1.88

² As where conjunct nasals are written for convenience as anusvāra, or where consonants conjoined with *r* are optionally doubled.

c2.s1.89

³ Thus the Calc. edition of the Vāyu has *Sneccha* sometimes instead of *Mleccha* by an obvious printer's error: see p. 47, note 78.

c2.s1.90

⁴ As the insertion or omission of final anusvāra or visarga through mere carelessness.

c2.s1.91

⁵ To have noticed such minutiae would have swollen the notes beyond all reason and usefulness.

c2.s1.92

⁶ *E.g.* p. 39, note 45; p. 40, note 69; p. 47, note 72; p. 49, note 14.

c2.s1.93

36. Though I am not an advocate of the use of Roman characters in lieu of Devanāgarī, yet, as this work is intended for the use of all interested in Indian archaeology whether Sanskrit scholars or not, practical usefulness should be the chief consideration in this presentation of the Purāṇic accounts of the dynasties of the Kali age. Hence the Roman character has been used throughout, because it offers several advantages over Devanāgarī, namely, (1) words can be separated which would be all run together when written properly in Devanāgarī; (2) compound words and words that have fused together by sandhi can be divided by hyphens and so displayed distinctly; and (3) by so treating words capitals can be introduced for names, and names can be exhibited unmistakably, even when initial vowels have been modified by sandhi. The system of transliteration is that adopted by the Royal Asiatic Society and most other Oriental Societies. Where vowels are blended by sandhi, the resultant vowel has been marked with a circumflex, except *ai* and *au* where a circumflex is inconvenient and hardly necessary. Changes in sandhi, which are required by the variant readings, are treated as necessarily consequential and are not mentioned. It has been necessary to introduce the double hyphen (used in transliterating inscriptions and MSS) in order to distinguish separate words that have become fused by sandhi¹, and I trust this sign may be pardoned here, especially as this Purāṇic account is not literature but only patchwork Sanskritization.

c2.s1.94

¹ Chiefly where names have fused with other words by single or double sandhi, as *bhaviṣyodayanas* (p. 7, l. 23; p. 82) and *bhavitāśoka* (p. 27, l. 2; p. 28, l. 4 in eVā); these are printed as *bhaviṣy-odayanas* and *bhavit-āśoka* in order to bring out the name clearly. It could hardly be dispensed with in such cases of double sandhi as *Yavaṅgśau* and *bhāryānyāḥ* (see p. 82); and as no line could well be drawn regarding its use, the simplest course was to adopt it throughout, except in the Appendixes and Introduction.

c2.s1.95

37. In interpreting the account the fact must be borne in mind that it was written in Prakrit originally, and this will throw light on many points, especially the variations in names and the meaning of numbers. It will explain how corruptions in names have sometimes occurred², it will help to elucidate doubtful passages³, and will be the best guide in solving difficulties in readings which appear corrupt⁴. The best course in such cases is to convert the different readings into literary Prakrit, write the Prakrit forms in the various old scripts, compare them, and see whether one can divine what was probably the original Prakrit statement. These remarks apply especially to the Matsya, Vāyu, and Brahmāṇḍa versions.

c2.s1.96

² *E.g.* p. 40, note 69; p. 41, note 74.

c2.s1.97

³ *E.g.* p. 52, notes 37, 38, 42.

c2.s1.98

⁴ P. 59, line 11 is an excellent crux for such solution.

c2.s1.99

38. There is often great variation in names. In some cases the correct form can be selected by reference to other books or to inscriptions, but where there is no such agreement I have not ventured to emend the Purāṇic forms from other sources, because it is my duty simply to edit the text and not to attempt to make it square with our present scanty knowledge of ancient Indian history--which is a separate matter. In such cases I have confined myself to estimating what form of the name is best attested by the MSS, and often the only feasible course is to adopt the most central form from which the other forms may be considered to diverge¹; but this is a measure more of convenience than of accuracy, because it happens sometimes that the correct form is what would appear to be an aberrant form²; and in such cases what is, or would seem to be, the correct form is suggested sometimes in the notes³.

c2.s1.100

¹ See p. 39, note 45 for an instance.

c2.s1.101

² E.g. Vindusāra, whose name is given correctly only by the Viṣṇu (p. 28, notes 23, 24).

c2.s1.102

³ As in p. 6, note 29; p. 42, note 98.

c2.s1.103

39. The numbers present much difficulty. Those that occur oftenest are *viṃśati* and *triṃśati*, and their abbreviated forms *viṃśat* and *triṃśat*, *viṃśa* and *triṃśa*⁴; and the difficulty arises because *tr* and *r*, if written carelessly or if partially frayed, are hardly distinguishable in the later script⁵, and *t* and *v* in the Prakrit forms of these words might have been confused from the first in Kharosthī. Hence in many cases either may be read as other data may indicate, irrespective of the weight of the MSS.

c2.s1.104

⁴ In nMt *ṣaḍviṃśati* looks like *ṣaḍyimiśati* generally, and *ṣaṭtriṃśati* like *ṣaḍimśati*.

c2.s1.105

⁵ For a clear instance see p. 57, note 4.

c2.s1.106

40. Various groups of misreadings will appear on an examination of the notes, and the most important may be mentioned here. First, *abda*, if the loop of the *b* be carelessly written so as to touch the top bar (as I have found it sometimes), may easily be misread as *aṣṭa*, and there can be no doubt that *abda* and *aṣṭa* have often been confused. Thus, where most MSS read *aṣṭapañcāśataṃ cābdaṃ*⁶, one has *°cābdā*, two *°cāṣṭān*, and one *°cāṣṭā*; and here *aṣṭa* is plainly a corruption of *abda* because it is impossible after *aṣṭapañcāśataṃ*. Again, one set of readings is *so smā daśa*, *so 'smāl daśa* and *tasmād daśa*, and another set is *aṣṭām daśa*, *aṣṭādaśa* and *so 'ṣṭādaśa*⁷: the latter suggest the reading *abdān daśa*, which (with the frequent use of anusvāra for nasals) would be often written *abdāṃ daśa* and might be misread as *aṣṭām daśa* and so pass to *aṣṭādaśa*: thus *abdān daśa* would reconcile all the readings as regards the number and would seem to have been the original reading. The same confusion occurs in other places⁸. This liability of *abda* and *aṣṭa* to be confused may harmonize other passages where the numbers 10 and 18 are in conflict. Moreover, *abda*, if the initial *a* is elided by Sanskrit or Prakrit sandhi, becomes *bda*; and *bda* may be mistaken for *dva* (= *dvau*)⁹; hence *drā* and *dvau* become a third alternative, and this possibility may harmonize other passages¹⁰. Secondly, *samā* and *sapta* have been confused sometimes, for it is not always easy to distinguish *m* and *pt* where written carelessly in the more modern scripts, as I have found. Thus two readings occur *saptāśītiṃ* and *samāśītiṃ*¹¹, and either might be derived from the other¹².

Cites (8)

Matsyapurāṇa — Vyāsa, Sanskrit
Vāyupurāṇa — Vyāsa, Sanskrit
Bhaviṣyapurāṇa — Vyāsa, Sanskrit
Bhāgavatapurāṇa — Vyāsa, Sanskrit
Brahmāṇḍapurāṇa — Vyāsa, Sanskrit
Garuḍapurāṇa — Vyāsa, Sanskrit
Viṣṇupurāṇa — Sanskrit
Viṣṇupurāṇa — Vyāsa, Sanskrit