Pramāṇa · Search · About · API

The Purana Text of the Dynasties of the Kali Age c6: Appendix I. The Account was originally in Prakrit

The Purāṇa Text of the Dynasties of the Kali Age c6

Language
English

Appendix I. The Account was originally in Prakrit

c6.2

Proof is offered here of the statement made in the Introduction, § 15, that the Sanskrit account as it stands in the Matsya, Vāyu, and Brahmāṇḍa is a Sanskritized version of older Prakrit ślokas, as indicated by these peculiarities: *first*, certain passages violate the śloka metre, whereas in Prakrit form they would satisfy the metre; *secondly*, certain Prakrit forms 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 Sanskritization of names and words; *fifthly*, the copious use of expletive particles; *sixthly*, irregular sandhi. Those three Purāṇas will be dealt with first, and along with them such portions also of the Bhāgavata and Viṣṇu as have preserved the old ślokas uncondensed; but the main portions of these two Purāṇas consist almost entirely of a condensed redaction, and their character will be considered afterwards.

c6.3

i. As an illustration of the first peculiarity, the Mt and Vā¹, when naming the last Paurava king, end the line thus (p. 7)—*Nirāmitrāt tu (or ca) Kṣemakaḥ*, ‘after Nirāmitra was Kṣemaka’; where the *tu* or *ca* in the fifth syllable should be short but is long by position before *kṣ*. No one composing in Skt could end a śloka line with *Kṣemaka*, but its Pkt form *Khemaka* satisfies the metre perfectly. There can be no doubt therefore that this line was composed in Pkt originally, and that the Skt redactor restored the Pkt name to its Skt form and in so doing overlooked the fact that the change violated the metre. The fault was however noticed afterwards, because eVā corrects it by altering the half line to *bhavitā Kṣemakas tathā* (p. 7, note 71). Precisely similar is the mistake in the line that ends with *samā bhokṣyanti trimśatim*², where no difficulty would occur in Pkt since *trimśati* would drop its *r* there³.

c6.4

¹ The Bd has lost this line in a large lacuna.

c6.5

² P. 50 (*Dynasties of the 3rd Cent.*), l. 2; the differences of reading there do not affect this point.

c6.6

³ The phrase *ā-Manu-kṣayāt* in the Vā and Bd at the end of a line (p. 51, l. 8) does not, militate against this view, because the *kṣ* in the middle of this expression would have been *kkh* in Pkt.

c6.7

Again the Mt reads at the end of a śloka line, *aṣṭāvimśatir Haihayāḥ*⁴, where the fifth syllable is long by position contrary to rule; and here the literary Pkt form *vīsati* without a termination would fit the metre. The Vā and Bd read instead *caturvimśat (or -vimśas) tu Haihayāḥ* and avoid the irregularity by reducing *vimśati* to *viṃśat* or *viṃśa* and replacing the lost syllable by a superfluous *tu*, which is the nearest approach to it. This expedient is very common as will be seen in the notes.

c6.8

⁴ P. 23, l. 4: *b*Mt avoids the fault by reading *aṣṭāvimśati*, keeping as near to Pkt as possible.

c6.9

Next may be cited cases where a half line has a syllable too much, which would disappear in Pkt, and the significance of these cases lies in the fact that the superfluity was unnecessary since good Skt equivalents were available, if the verse had been composed directly in Skt. Thus the Bh has a śloka prophesying Viṣṇu's incarnation as Kalki thus—

c6.10

dharma-trāṇāya sattvena Bhagavān avatariṣyati ¹.

c6.11

¹ Bh xii, 2, 16; omitted from p. 57.

c6.12

The second half line has a syllable too much, but the Pkt verb *otarissati* would exactly suit the metre and was no doubt the word used originally, as *dBh* (an old MS of 1407) shows by reading *Bhagavān vatariṣyati*, where *n* and *va* are separate letters. Many two-syllabled equivalents for *bhagavān* were available to suit the metre. Again *eVā* has for the first half of a line, *sapta varṇāṇi Devadharmā* ², where *varṇāṇi* is obviously a misreading of *varṣāṇi*, and there is a syllable too much; but the Pkt form *varṣā* or *vassā* satisfies the metre. To one composing in Skt *samāḥ* would have avoided all difficulty. Similarly *Mt* has the first half of a line, *aṣṭāviṃśati tathā varṣā* with a syllable too much ³, but the Pkt *aṭṭhāvisam* would rectify the metre; whereas one composing in Skt could have written simply *aṣṭāviṃśati-varṣāṇi*, which is indeed the general reading of the *Mt* now. This instance may give us an insight into the process of Sanskritization, if, as the *Mt* reading suggests, the original Pkt was *aṭṭhāvisam tathā vassā*.

c6.13

² P. 29, l. 11, and note 38.

c6.14

³ P. 19, note 26.

c6.15

ii. Actual Pkt forms occur rather often. First may be cited the Vā and Bḍ line ⁴—

c6.16

⁴ P. 52, l. 13 and notes.

c6.17

sthāpayiṣyati rājāno nānā-deśeṣu te janā:

c6.18

where *rājāno* and *te janā* are Pkt accusatives after the verb. They were misunderstood as nominatives, and the verb was altered to the plural in all copies of the Vā except *eVā*, and in the Bḍ. Similarly the Bh introduces the Bārhadratha dynasty with the old line ⁵—

c6.19

⁵ P. 14, note 1.

c6.20

atha Māgadha-rājāno bhavitāro vadāmi te.

c6.21

Here *rājāno* and *bhavitāro* are accusatives, hence the line is not Skt but is actually good Pali. *BrBh* correct the faults by altering *bhavitāro* to *bhavino ye*. Again the *Mt* has a half line *prasahya hy avanīm nṛpaḥ* in many copies, and *prasahya vyasanī nṛpam* in some copies, while the corresponding reading of the Vā and Bḍ is *bālyād vyasaninām nṛpam* ⁶. The *Mt* reading should evidently be *prasahya vyasanīm nṛpam*, and points to a Pkt original something like *pasajja (or pasayha) vasanīm nṛpam* ⁷, but this when Sanskritized became *prasahya vyasaninām nṛpam* with a syllable too much, and so was adjusted in two ways, (1) the half-Pkt form *vyasanīm* was used as an accus. in many copies and became corrupted to *hy avanīm*; or (2) the half line was emended to *prasahya vyasanāturam* in some copies. The Vā and Bḍ may have substituted *bālyād (or balād?)* for *prasahya* to rectify the metre.

c6.22

⁶ P. 33, note 2.

c6.23

⁷ This would be the correct accus. in Pkt, see Pischel's Prakrit Grammar, § 405.

c6.24

Next may be cited a number of actual Pkt or half-Pkt words. All such forms cannot be deemed original, because the copyists, who were not always sufficiently literate, did write Pkt forms sometimes instead of Skt forms, but such deviations are trivial and obvious mistakes¹. It is different however when the Pkt forms violate grammar or sandhi, or suit the metre, and such are these—*uccādayitvā* (p. 53, note 52), *mahāyaśāḥ* as a nomin. plural² (p. 51, note 25) and *varṣā* (see p. 78). There are also instances of the Pkt genit. plural in *āna* (p. 35, note 41), and of its blending with *ante* into *ān-ante*, namely *kulānante* (p. 50, note 2) and *Andhrānante* (p. 59, notes 51, 54; p. 61, note 92). Other words appear to be Pkt survivals and not copyists' errors, such as *aṭṭitriṃśat* (p. 19, note 41), *athóchādya* (p. 34, note 3), *teṣūcchannesu* (p. 48, note 82), *samā* for *samās* before *tasmāt* in Vā genly (p. 34, note 18), and *Asakāḥ* in Vā and Bḍ (p. 24, note 15). In an old verse *Bh* has *papannān uhariṣyati* (p. 26, note 36), which seems more than a mere clerical error. Mistakes precisely like these are found in Buddhist Skt.

c6.25

¹ *E.g.* see p. 2, note 16; p. 43, note 27; and these are found even in Bh MSS, see p. 46, note 27.

c6.26

² This is possible only in Pkt and does actually occur, see Pischel, *op. cit.* § 409.

c6.27

The Bhāgavata has an old verse—

c6.28

yasmin Kṛṣṇo divam yātas tasminn eva tadāhani
pratipannam Kaliyugam iti prāhuḥ purāvidaḥ.

c6.29

The Mt, Vā, Bḍ, and Vṣ all have this verse, but read the last half line *tasya saṅkhyāṃ nibodhata* or in equivalent words³. The Bh reading appears to be the oldest version, because its verse is complete in itself and is obviously an old saying, whereas the last half line in the other authorities was evidently substituted to connect this statement with the following verse when this collective account was drawn up: the reverse is hardly credible. Further, one old Bh MS (*d*Bh, dated 1407) reads *iti-r-āhuḥ purāvidaḥ*, and this with its euphonic Pkt *r* is no doubt the original form, which in the process of Sanskritization was amended to *iti prāhuḥ* as in all the other Bh copies; here also the reverse is hardly credible. *Iti-r-āhuḥ* is the Pkt *iti-r-āhu*⁴. There are one or two other instances of an *r* inserted, which seems to be euphonic⁵; and it may possibly be that the final *r* in the nominatives of numerals is sometimes as much a euphonic Pkt *r* as a Skt *r* by sandhi⁶.

c6.30

³ P. 62, ll. 37, 38 and notes.

c6.31

⁴ See Pischel's Prakrit Grammar, §§ 353, 518.

c6.32

⁵ See *hatra-r* in p. 38, note 2.

c6.33

⁶ As in p. 43, l. 36, where the accusative would be proper.

c6.34

Similarly no doubt are to be explained the Bḍ reading of p. 62, l. 40 and the Vā readings in note 31 thereto. The reading in literary Pkt would have been something like *vassāna uccate Kali* or rather *vassāna-r-uccate Kali*. Turned into Skt, *varṣāṇām ucyate Kaliḥ* was good and sufficient, yet notwithstanding, the desire for an expedient to prevent the hiatus persisted in the Sanskritization, for *m*Vā has preserved the euphonic *r*, and *bdfg*Vā inserted *tu* instead. These were no doubt the original forms of the Sanskritizations, but it was perceived that no such expedient was wanted, hence most copies of the Vā dropped it. The reverse is not credible.

c6.35

Most common is the use of numerals with the Pkt freedom from case-terminations, as well as only half Sanskritized, such as—*aṣṭāśīti* and *aṣṭāsīti*⁷, and *viṃśati* often both in the text and in the notes. Some of these instances might be due to the carelessness of copyists in omitting visarga or anusvāra, but that does not account for all such peculiarities, since they are found in carefully written MSS and are sometimes obligatory for the sake of the metre. Thus the Vā and Bḍ read as the last half line of a śloka, *aṣṭāviṃśati Maithilāḥ*⁸, and this was no doubt the original reading because *fyMt* have it also; but the Mt has generally altered *viṃśati* to *viṃśās* (or *°śat* or *°śa*) *tu*. The Skt form *viṃśatir* would violate the metre, and the Mt has avoided the difficulty of Sanskritization by substituting *tu* for the final syllable. This is the converse of the first irregularity noticed above (p. 78), and many similar instances of *tu* substituted for a final *ti* will be found in the notes.

c6.36

⁷ P. 25, l. 5 and note 17.

c6.37

⁸ P. 24, l. 6.

c6.38

iii. Of the third class of peculiarities the following are instances. As the last half line of a *śloka* the Vā and Bd have in one place *varṣāṇi bhavitā trayaḥ*¹, and in another *tasya putraḥ samās trayaḥ*²; and the Mt has in another place *bhaviṣyati samās trayaḥ*³. In all these passages grammatical concord is violated, because (1) these are accus. expressions denoting duration of time, and (2) *varṣāṇi* is neuter, *samās* feminine, and *trayaḥ* masculine and nomin.; but, if the Pkt *tao* be substituted for *trayaḥ*, concord is established, because *tao* is both nomin. and accus. in all three genders⁴, and the metre also is satisfied. Such expressions could not have been composed in Skt originally. There can be no doubt that they were originally in Pkt and that, when the verses were Sanskritized, the exigencies of metre induced the redactor to convert *tao* into *trayaḥ*, because the correct equivalents *trīṇi* and *tisraḥ* would not suit the metre⁵.

c6.39

¹ P. 32, l. 8. The Mt reads correctly *trīṇi varṣāṇi*.

c6.40

² P. 43, l. 32. The Mt reads differently, *samā daśa*.

c6.41

³ P. 40, l. 15. The Vā and Bd omit this, except eVā which alters it to *samā-trayam*.

c6.42

⁴ Pischel's Prakrit Grammar, § 438.

c6.43

⁵ Unless he recast the line, which was obviously not attempted, except by Mt in the first instance, see note 1.

c6.44

The same fault occurs in places where metre was not at stake. Thus all three Purāṇas read *catvāriṃśat trayaḥ* *caira* as the first half of a line⁶, where *samās* or *varṣāṇi* is implied and *trayas* is wrong as regards both gender and case. CVā attempts to rectify the discord by reading *trayam*. Similarly in another passage the Mt has *samās trīṇy evam*, while the Vā and Bd read *samās tisra eva*⁷. It is impossible to suppose that these wrong expressions were composed originally in Skt, and they are intelligible as perfunctory Sanskritizations of Pkt expressions containing the numeral *tao*, or *tiṇṇi* which also is of all three genders⁸. Similarly we find the phrase *ṣaṣty-uttara-śata-trayam* used with *varṣāṇi* in the Bd and with *samāḥ* in the Bh⁹. Other instances are *saptaṣaṣṭis tu varṣāṇi*¹⁰, and *aṣṭāśītis tu varṣāṇi*¹¹, where the case is wrong; *ye cānye Mleccha-jātayaḥ*¹² which eVā has corrected to *yāś cānyā°*; and perhaps *divyābdāni*¹³ where the correct *divyābdās* was as easy as in the Bd.

c6.45

⁶ P. 22, l. 14 and notes.

c6.46

⁷ P. 32, l. 7; but *dfyjm*Mt alter it to *tisro vai*.

c6.47

⁸ Pischel's Prakrit Grammar, § 438.

c6.48

⁹ P. 22, note 46.

c6.49

¹⁰ P. 46, l. 7.

c6.50

¹¹ P. 25, note 17.

c6.51

¹² P. 3, l. 11 and note 37.

c6.52

¹³ P. 60, l. 16 and note 70.

c6.53

iv. Some forms of names look strange as Skt but are readily intelligible if they are mistaken Sanskritizations of Pkt forms. Thus the name *Śiśunāga* as found in the Bd, Vṣ, and Bh appears as *Śiśunāka* in the Mt and Vā¹⁴. *Śiśunāga* as Pkt might naturally be Sanskritized as *Śiśunāka*, because a Pkt *g* often represents a Skt *k*: otherwise it is difficult to see how the form *Śiśunāka* could have arisen. Similarly eVā has *Śuṅka* and *Śaṅka* for *Śuṅga*¹⁵; *eka-kṣatro* appears instead of *eka-cchatro*, and *eka-kṣaṭrām* instead of *eka-cchattrām*¹⁶.

c6.54

¹⁴ P. 21, ll. 1, 3; p. 22, ll. 15, 17; and notes thereto.

c6.55

¹⁵ P. 30, note 50; p. 32, note 47; p. 49, note 17.

c6.56

¹⁶ P. 25, l. 4 and notes 14, 15.

c6.57

In this class may be mentioned certain incorrect forms: thus the Vā generally reads *caturas* instead of *catvāras* in p. 34, l. 7 (note 20), where the Pkt *caüro* may have been used as a nomin. though it is strictly accus.¹ So the Mt generally has *catvāriṃśad* instead of *catvāraś ca* (or *tu*), which would be an intelligible mistake if the Pkt was *cattāri ca*, for *cattāri* though neuter was often used as masculine². The plural verb *bhokṣyanti* instead of the dual in p. 50 (*Dynasties of the 3rd Cent.*), l. 2, would be correct in Pkt but not in Skt.

c6.58

¹ Pischel's Prakrit Grammar, § 439.

c6.59

² P. 35, note 29. Pischel, § 439.

c6.60

Vernacular names had to be Sanskritized and so developed strange forms; compare for instance *Simuka* in p. 38, note 17, and other Andhra names.

c6.61

Attention may also be drawn to p. 59, l. 11, where all the divergent readings are obviously attempts to Sanskritize one and the same original Pkt statement that was puzzling.

c6.62

v. The fifth class of peculiarities is a very noticeable feature of these texts, namely, the copious use of particles as mere expletives, such as *tu*, *hi*, *ca*, *vai*, &c., and especially *tu*. The lines in which two such particles occur are too numerous to be mentioned, but three and even four are sometimes found in a single line, and the following lines are cited as most illustrative:—

c6.63

bhavitā cāpi Sujyeṣṭhaḥ sapta varṣāṇi vai tataḥ³
Svātiś ca bhavitā rājā samās tv aṣṭādaśaiva tu⁴
Śivaśrīr vai Pulomā tu saptaiva bhavitā nṛpaḥ⁵
sapta Gardabhinaś cāpi tato 'tha daśa vai Sakāḥ⁶
trayodaśa Muruṇḍāś ca Mauṇā hy ekādaśaiva tu⁷
saptaṣaṣṭis tu varṣāṇi daśābhīrās tathaiva ca⁸
śatāni trīṇy aśītiṃ ca Śakā hy aṣṭādaśaiva tu⁹
Pulomās tu tathāndhrās tu Mahāpadmāntare punaḥ¹⁰

c6.64

³ P. 40, l. 13, Mt.

c6.65

⁴ P. 42, l. 29, Mt and eVā.

c6.66

⁶ P. 45, note 12, Vā and Bḍ.

c6.67

⁷ P. 46, l. 5, Vā and Bḍ.

c6.68

⁸ P. 46, l. 7, Mt.

c6.69

⁹ P. 46, l. 9, Mt.

c6.70

¹⁰ P. 58, l. 7, Mt.

c6.71

One cannot imagine that these verses were composed originally either in Skt or in Pkt with so many expletives, when the authors could easily have improved their verses by employing appropriate words denoting 'reign' or 'exist' or 'relationship'. No one composing in Skt would mar his verse and proclaim his literary poverty by such shifts; but these blemishes are readily intelligible, if the verses were originally in Pkt as chronicles of the past and were converted into Skt prophecies. Future tenses are longer than past tenses, and if they could not be fitted into the place of the past tenses, it would have been natural to substitute expletives. Thus it may be conjectured that the second, fifth, sixth, and seventh lines ended originally with a past verb corresponding to *abhavat* or *abhavan*. Again, Pkt forms are sometimes longer than their Skt equivalents, and the substitution of the latter would have been compensated for by adding an expletive; thus in the third line *Śivaśrīr vai* no doubt stands for the Pkt *Sivasiri*, and in the eighth line *Pulomās tu tathāndhrās tu* probably mean the ablat. case and stood originally something like *Pulomādo tathāndhrādo*, or *Pulomamhā tathāndhramhā*.

c6.72

It has been noticed above (pp. 78, 80) that the particle *tu* is used sometimes to compensate for the loss of the final syllable of *viṃśati* and *triṃśati*. When the full forms of these words vitiated the metre, they were reduced sometimes to *viṃśat* or *viṃśa*, and *triṃśat* or *triṃśa* and the lost syllable was replaced by an expletive *tu*. This expedient is very common and many instances of it will be found in the notes. Indeed it is hardly too much to say that the occurrence of *tu* throughout the account, if not required by euphony (see next para.), almost certainly indicates a lost syllable, and in many cases *tu* in the Mt and *e Vā* has been altered to *ca* in the Vā and Bd as an improvement. Other instances of compensatory expletives may be surmised in the notes, such as these—*Suṣeṇaś ca-Antarikṣāc ca* (p. 10, l. 13) is hardly explainable unless the second *ca* has replaced the lost syllable of the Pkt ablative; and *Dharmiṇaḥ sa* (p. 11, l. 15) no doubt stands instead of the Pkt genitive *Dharminassa*.

c6.73

The use of expletives was however carried beyond necessary requirements, and they are often inserted merely to prevent two vowels from coming together, as *tv* in the second of the above-cited lines, and *hy* in the fifth and seventh lines. Skt sandhi did not require this device, but it is intelligible in Pkt. This superfluity is found in the Bh also, where it has not condensed the older ślokas, as in *śūdra-prāyās tv adhārmikāḥ* (p. 25, l. 3).

c6.74

vi. The instances of irregular sandhi may be divided into two classes; *first*, those in which the form it takes resembles Pkt sandhi and is unnecessary, because regular Skt sandhi would have been proper and sufficient; and *secondly*, those in which it consists of double sandhi in order to contract the words for the metre.

c6.75

Of the first class may be cited *varṣāṇi ’kārayat* instead of *varṣāṇy akārayat* (p. 15, note 29); *triṇi ’śitīś* for *triṇy aśitīś* (p. 46, note 48); *Daśarathāṣṭau* instead of *Daśaratho ’ṣṭau* (p. 28, note 5); and *Agnimitrāṣṭau* for *Agnimitro ’ṣṭau* (p. 31, note 10). Such sandhi can be explained through Pkt, and it is difficult to understand how any one composing in Skt could have adopted it; nor is it probable as a copyist’s error.

c6.76

The second class is commoner, and we find—*bhavisyāṣṭau* for *bhaviṣyāḥ aṣṭau* (p. 5, l. 10); *bhaviṣyōdayanas* for *bhaviṣyaḥ Udayanas* (p. 7, l. 23); *Yavanāṣṭau* for *Yavanāḥ aṣṭau* (p. 45, l. 4; p. 47, l. 10); and *bhāvyānyāḥ* for *bhāvyāḥ anyāḥ* (p. 47, l. 13). Here ordinary sandhi would have given a superfluous syllable, and the double sandhi rectifies the metre; but the significance of it is that it was easily avoidable in Skt, because the first two phrases might have been written *bhāvino ’ṣṭau* and *bhavitōdayanas*. The simplest explanation seems to be, that the conversion of the Pkt past tense into the Skt future was made perfunctorily, and overloaded the verse with a superfluous syllable which was adjusted by the double sandhi. The third phrase would have been *Yonā aṭṭha* in Pkt, and the Sanskritization of *Yonā* into *Yavanāḥ* produced the difficulty of the extra syllable. Attempts at improvement were made; see p. 45, note 16. There are many similar instances, such as *tatōtsādya* and *tatōtpātya* (p. 34, notes 3, 5); *atōddhṛtya* and *tatōdhṛtya* (p. 38, note 4).

c6.77

Crasis of this kind is ordinarily explained as *ārṣa-sandhi*, but this explanation is manifestly untenable here 1. All these irregularities are readily intelligible on the two suppositions, that Pkt words were converted into their Skt equivalents, and that past tenses were changed to futures, with the metrical difficulties that naturally ensued.

c6.78

vii. All these peculiarities are found in the Mt, Vā, and Bd throughout, and show that their version must have been composed originally in Pkt ślokas and that the ślokas were Sanskritized for incorporation in the Bhaviṣya, from which the Mt and Vā confessedly, and the Bḍ impliedly, borrowed their accounts (see Introdn. § 7). The Prakritisms which have been cited are not mere casual variations, for such might be due to the ignorance or carelessness of copyists, but have an important *raison d'être* in the verse in many cases. The same conclusion holds good for the Vṣ and Bh in the passages where they have preserved the old śloka form.

c6.79

¹ In the Purāṇas what is called *ārṣa-sandhi* is really Prakrit sandhi; see p. 20, note 2.

c6.80

viii. The main part of the Viṣṇu account is in prose and, not being affected by the exigencies of metre, runs in ordinary Skt, and displays no verbal peculiarities. It contains the same matter found in the Vā and Bḍ but in a condensed shape, and closes its account where they end, so that it must have been composed directly in Skt from them or their original, the revised version in the Bhaviṣya, for it is not probable that its account was a new and independent compilation, when the compilations in those Purāṇas were available. A difference may be noticed in its account to this extent that the dynastic matter is generally narrated in curt sentences, often without regard for sandhi¹, and that the subsequent matter of the evils of the Kali age is in ordinary good prose Skt with a predilection for compound phrases. Hence it would seem that the dynastic portion was an earlier and somewhat crude condensation, and that the latter portion was an addition made with regard to the canons of good prose.

c6.81

¹ As in p. 18, note 7; p. 30, note 45; and in these curt sentences *tasyāpi Aśokavardhanaḥ, tataś ca Ariṣṭakarmā, and tasmāt Yajñaśrīḥ*. *abGr*, which may be a faulty Sanskritization of the Pkt *Dadhasena + ka*, though it might also be the form of that name in one kind of Pkt; see p. 16, note 75.

c6.82

ix. The Bhāgavata account, which is mainly a condensation, is evidently a later redaction. Peculiarities of the kinds noticed above do not appear therein, but it is in good Sanskrit, and phrases occur in it which indicate that it must have been composed directly in Skt. Two are especially significant. A śloka line ends with the words *ekādaśa kṣitim* (p. 48, note 77), where the *śa* is long by position before *kṣ* as it should be, but would not have been long in Pkt in which *kṣ* would have become *kh*; so that this line must have been composed in Skt and not in Pkt. Similarly another line ends *iti śrutaḥ* (p. 32, note 45), where the second *i* is long by position in Skt but would not have been so in Pkt.

c6.83

x. The Garuḍa has no Prakritisms except in some of the names, and these are too uncertain a basis on which to argue, for those Prakritisms might be original or might be due to the carelessness of copyists, yet one name certainly seems somewhat suggestive². All that is clear is that its account is the last and concisest redaction, that it was probably composed afresh in Skt, and that it makes frequent use of the termination *ka* for the sake of the metre. Its treatment of the name Adhisīmakṛṣṇa suggests that it was composed from a bare list of kings, for it divides the name into two, *Adhisīma + ka* (ending one line) and *Kṛṣṇa* (beginning the next line)³—which seems inexplicable unless it had only a prose list and chopped the names into groups for each line.

c6.84

² *Dṛḍhasenaka* appears as *Dathasenaka* in

c6.85

³ See p. 4, note 10.

Linked texts (7)