Pramāṇa · Search · About · API

On the Aindra School of Sanskrit Grammarians, their Place in the Sanskrit and Subordinate Literatures c8: Appendix A

On the Aindra School of Sanskrit Grammarians, their Place in the Sanskrit and Subordinate Literatures c8

Language
English

Appendix A

c8.1

IN considering the historical development of Sanskrit Literature (which has as yet been done but to a small extent) the only possible way is to compare the contents of books, one with another. But in doing so, no one has thought fit, as yet, to fully enquire whether the texts as we have them can be accepted in their integrity as evidence for particular periods. A general estimate has been formed in many cases of the age of a particular work, but hardly in a single case has there been any attempt made to show under what influences each work has been formed—in short, to ascertain the age of the separate parts.

c8.2

First, it is necessary to remark the singular circumstances under which most of the standard texts have been delivered to the world. These are nearly all according to the recensions established by various commentators who all lived in S. India, during the middle ages, and which recensions or editions early attained a great reputation all over India partly because of the intrinsic merits of these commentaries, partly because they were the means of founding religious sects which soon spread all over India¹.

c8.3

¹ It is only necessary to remark that our chief editions of the most important Vedic books follow Śaṅkarācārya and Sāyaṇa, and that Mallinātha's recension is followed for Kālidāsa's works. (See Stenzler's "Meghadūta", 1874, pp. iii.-iv.)

c8.4

Yet most of our editions—in fact, there are only two or three exceptions—are edited entirely from N. Indian Nāgarī MSS., and most of the editors repudiated with scorn any help from the palm-leaf MSS. of the South, a prejudice which is only now beginning to wear away. South India has, in fact, contributed the texts, but N. India the secondary MSS. from which the editions have been made. Here then is a primary reason for not relying too much on the letter of our orthodox standard editions.

c8.5

Again, it is evident that these commentaries must be taken for what they are worth—very meritorious works in fact, yet sectarian; but a critic who now-a-days would understand any particular passage in the traditional collections known as the Vedas and other early texts, may not rely on sectarian commentators. The works of such are chiefly of value for a history of dogma, an interesting matter in itself, but quite foreign to critical purposes. Such primitive texts as the Vedas are peculiarly liable to misunderstandings, and it will often be found that commentators only pretend to understand them "because they have grown accustomed to misunderstand them". It must be again asked if, besides unintentional perversions, texts may not have been arranged, selected or interpolated wilfully for certain purposes?

c8.6

It appears to me that it is possible to establish some general principles, by which Sanskrit texts may be compared for historical purposes; it was necessary to raise the question in order to show that it is possible to use the grammatical texts in this way as Goldstücker did, and as I have attempted to do in the preceding monograph. If we know some at least, of the laws which regulate the reflection of an object, we may, to some extent, infer as to the nature of that object, though it be not directly accessible.

c8.7

Before dealing with the grammatical treatises, it will be best to see what conclusion other texts point to. The earliest relics of Indian literary culture—the Vedas—are not of direct importance in this enquiry, but it is necessary to notice their present condition and the indigenous views regarding them.

c8.8

The true nature of the Vedic collections is well pointed out by Profr. Roth. He says¹): "Es ist eine bekannte Sache, dass die verschiedenen Sammlungen alter Lieder, Sprüche, Formeln u. s. w. welche jede durch eine besondere Benennung unterschieden unter dem gemeinsamen Namen der Veden überliefert sind, nicht die einzigen ihrer Art waren. Jede dieser Gattungen hatte ihre Arten. Wie in Religionsgemeinden, deren Bestand auf eine Verkündigung oder Lehre zurückgeht, die Feststellung des Inhalts dieser Lehre, da sie nicht bei allen und überall dieselben Wege geht, zu verschiedenen Dogmen und dadurch zu Secten führt, so haben sich in Indien ähnliche Unterscheidungen, wenn auch mehr äusserlicher und darum weniger feindlicher Art, an der heiligen Überlieferung entwickelt und befestigt.

c8.9

¹ "Der Atharvaveda in Kaschmir", pp. 9-10.

c8.10

"Die vier Veden, in welche der gesammte Stoff dieser Überlieferung seit alter Zeit zerlegt wird: Lieder, Gesänge, Opfersprüche, Wünsche und Verwünschungen (carmina, incantamenta, devotiones) sind ursprünglich nicht eben so viel Bücher, sondern Bezeichnungen von vier Gattungen des überlieferten Wortes, welche lange Zeit nur im Gedächtniss von Geschlecht zu Geschlecht fortgepflanzt wurden. Sobald aber aus einer dieser Gattungen eine bestimmte Auswahl getroffen und nach Aufeinanderfolge, Zahl und Wortlaut der einzelnen Stücke festgestellt wurde—was nach meinen Begriffen nur mit Hilfe der Schrift wirksam geschehen konnte—so entstand ein vedisches Buch. Traf ein anderer Kenner und Lehrer aus derselben Gattung des Überlieferten eine im einzelnen abweichende Wahl und setzte dieselbe in seinem Kreise in Wirksamkeit, . . . . . so standen zwei Bücher nebeneinander, welche einen und denselben Veda enthielten, beide aus derselben Tradition erwachsen waren wie zwei Zweige eines Baumes. . . . . . Es lässt sich annehmen, dass unter gegebenen Verhältnissen diese Unterschiede sich zu wirklichen Spaltungen erweitern konnten; in der Regel jedoch sind sie sehr harmlos, ja so gut wie bedeutungslos.”

c8.11

The Indian view of the Vedic texts is remarkable; they are considered to have existed from eternity and to have been uncreated¹). Thus, the real ‘forms’ or types which exist in a higher world (like the Platonic ideas), had representatives, but not necessarily perfect, on this earth. As the Vedic texts²) now exist, they are of little importance for the present matter, but the old Indian doctrine concerning them is of much importance, as it has been extended to other parts of the Sanskrit literature.

c8.12

¹ Singularly enough some of the Muhammadan (Sunni) teachers taught exactly the same doctrine in respect of the Qorān. On the Indian doctrine see Muir's S. Texts, iii.

c8.13

² As we have the Vedic texts at present, they are not free from errors which must have been made at the time of their last revision, or in the pre-historic times of India. So also some stanzas are irrelevant interpolations where they occur, and the proper order has not always been preserved in hymns which are otherwise intact. See Roth “Der Atharvaveda in Kaschmīr,” p. 19; and “Siebenzig Lieder des Ṛigveda” passim. See also, Benfey—“Einleitung in die Grammatik der Vedischen Sprache” p. 8 ffg.; and his treatise “Die Quantitätsverschiedenheiten in den Saṃhitā- und Pada-Texten der Veden,” i., pp. 12, 25. Special researches of the most interesting character are those by Profr. de Gubernatis (in “Annuario della Società Italiana per gli studi Orientali” I. pp. 41-49); Bollensen (in “Orient und Occident,” ii., pp. 457 ffg.) But the whole tradition of the Vedic texts as handed down is, beyond doubt, surprisingly accurate [see especially Profr. Max Müller’s Ṛigvedasaṃhitā (translation), i. pp. lxvii. ffg.]; and this is not only the case with the Saṃhitās but also in many cases with the later Brāhmaṇas. Some, however, have undergone revision (see my “Sāmavidhāna Br.” i., pp. x., ffg. in which M. Barth concurs—“Revue Critique” 1873, No. 44, p. 282). I have in another place (“Devatādhyāya Br.” Preface, p. xi.) attempted to show that some Brāhmaṇas were put in their present order by Sāyaṇa, and that he must be considered as the final author of the Vedic canon; on this see M. Barth’s article in the “Revue Critique” No. 12 of 1874, p. 179.

c8.14

Thus Patañjali says:¹) “Bṛihaspatir Indrāya divyavarshasahasram pratipadoktānām śabdānām śabdapārāyaṇam provāca nā ’ntam jagāma | Bṛihaspati ca pravaktā Indraç ca ’dhyetā, divyam varshasahasram adhyayanakālaḥ | na cā ’ntam jagāma” etc. This is the ‘ideal’ science; to meet actual wants of mortals we have ‘grammar’. A little reflection will show that this view must necessarily hold good in respect of all ‘çāstras’, and hence, that with treatises which profess to come originally by revelation it is not possible to expect an exact reproduction of a primitive text, for there probably never was one; rather, the oldest treatises we have, are to be looked on as the joint work of many generations of teachers and pupils, and the result of much revision. The Indian view is given by Colebrooke²): “It is a received and well grounded opinion of the learned in India that no book is altogether safe from changes and interpolations until it have been commented: but when once a gloss has been published, no fabrication could afterwards succeed; because the perpetual commentary notices every passage, and, in general, explains every word.” But there are comparatively few such commentaries, and very few of a remote date, and there is often reason to believe that these have caused in effect, the loss of different recensions³), which not being protected by such a useful aid, have fallen into neglect. As I have already remarked, such commentaries⁴) often originated new errors in addition to those already existing in the texts.

c8.15

¹ Benares lithog. ed. i., ff. 10 b.-11 a. See in the magnificent photo-zincograph edition issued by the India Office, p. 14, of vol. i.

c8.16

² “Essays” i., pp. 98-9.

c8.17

³ See my “Sāmavidhāna-brāhmaṇa” i., pp. xxxiii.-xxxiv.

c8.18

⁴ Burnouf [“Bhāgavatapurāṇa”, ii. (folio ed.) pp. vi.-vii.] already in 1844, protested against a servile adherence to the Commentaries. He says: “J’ai continué à accorder une assez grande confiance au commentaire de Çrīdhara Svāmin, qui est en général ample et exact. Je l’ai suivi principalement toutes les fois qu’il a été question d’un point de fait ou de doctrine; dans les passages purement poétiques, je me suis permis plus de liberté. Ce n’est pas que je pense que nous devions abjurer le sens commun et la connaissance que nous avons acquise de la langue sanscrite, pour nous soumettre en aveugles aux explications souvent mesquines et erronées des commentateurs indigènes; mais je suis d’avis que leurs opinions mériteront toujours une attention particulière . . . . Enfin, la condescendance qu’on peut avoir pour leurs opinions n’exercera jamais sur le progrès européen des études indiennes une très-fâcheuse influence; car, de deux choses l’une: ou les explications brahmaniques sont vraies, et alors elles (vii.) se justifieront plus tard d’elles-mêmes; ou elles sont fausses, et alors la critique ne tardera pas à posséder les moyens d’en faire justice. Qui aurait le courage de reprocher au digne et à jamais regrettable Frédéric Rosen d’avoir suivi, un peu servilement peut-être, les sentiments des commentateurs indiens? et qui, d’un autre côté, pourrait être blâmé d’opposer à ces sentiments quelques-unes de ces interprétations simples et fécondes, qui sortent si naturellement des textes expliqués pur les seuls secours de la philologie?” This is Profr. Max Müller’s view.

c8.19

Thus, it is evident, it will not be safe to start with an assumption that the later Sanskrit texts have been preserved from interpolations, or that we can look back in hopes of arriving at a text as originally written by its author; in this they form a striking contrast to the Vedic collections.

c8.20

1. The current texts of the epic poems fully support this view.

c8.21

a. Rāmāyaṇa. It has long been known that two chief recensions at least of this epic are in existence; the texts of which agree generally in substance but very little in language and expression. The existence of the second recension¹) which is generally accepted as the oldest and best, was first established by A. W. von Schlegel in his edition published in 1838; but the many persons who have used the Rāmāyaṇa as an authority for early Indian history and the history of Indian ideas, have (except Dr. Muir) entirely neglected this fact. The most remarkable Essay by Profr. Weber—‘On the Rāmāyaṇa’—has excited so much attention that better hopes may be entertained for the future.

c8.22

As regards the first recension (commonly called Gauḍa) which prevails in Bengal, there is comparatively little information as to its history and eventual origin. It has been edited in full by Gorresio, and in the most splendid style; the first edition (of part only) was by Carey and Marshman, from 1806 to 1810²). But, like all the earlier editions, this last was eclectic, and does not always follow the same MS. or any consistent rule. Gorresio’s estimate of this recension is as follows: “Il mio giudizio è; che la recensione Gaudana non può essere in nessun modo nata d’una rimutazione, d’un rifacimento della boreale; che essa è al tutto indipendente da quella, non rilavorata, rimpastata, ma originale, autentica, e degna di fede quanto l’altra; che ella attinse da sorgenti sue proprie, schiette, ed autorevoli, quanto quelle della boreale e rappresenta fedelmente un’altra tradizione del poema—” etc³). Of the history of this recension I cannot find any certain traces, but it must be of some age, as Gorresio shows that there have been several commentaries on it⁴). Profr. F. E. Hall, however, considers this recension to be a modern depravation⁵).

c8.23

1) Schlegel’s edition i., pp. xxii. ffg. A remarkable MS. of this recension is described by Profr. Aufrecht, “Catalogus” pp. 344-5. Gorresio calls it ‘boreale’, but it is current all over India, except Bengal, and with slight variations.
2) See p. xvi. of Vol. 1, for Schlegel’s justly severe criticisms on this edition; Gorresio (i., p. xix.) concurs.
3) ed. Gorresio, i., pp. xl. ffg. The matter is discussed at length.
4) Do. pp. xxii-xxiii.
5) See his edition of Wilson’s Viṣṇupurāṇa, ii., p. 190 note.

c8.24

Of the second chief recension which Schlegel brought into notice it is possible to show that it has not been left unaltered. This important information is due to the anonymous author of the oldest commentary that we possess as regards this recension, and which is referred to everywhere as an authority—the Kaṭakaṭīkā. The name given to this work is ominous, the Kaṭaka being the berry of the poisonous strychnine tree which is used in India (so people say) to clear muddy water¹; but the author has stated, in more than one place, the principles on which he worked. These were first pointed out by Dr. Muir²; they are as follows:

c8.25

¹ See Wilson's Dy. *s. v.* The author says:

c8.26

² "Sanskrit Texts" iv., pp. 180-1.

c8.27

1. That a passage is to be rejected which is not in the old copies. And apparently:

c8.28

2. That a passage is to be rejected, if inconsistent with the rest of the text. The subsequent commentators quote these principles and the precedent of the author of the Kaṭakaṭīkā in justification of their treatment of the text. It is, therefore, obvious that it is impossible to trust very far the text of the commentators, for they clearly state that they have revised the text on principles which are not admitted to be trustworthy by modern critics. Besides these two chief recensions, the MSS. vary greatly in minor details as Profr. Weber has proved.

c8.29

There is no way of determining the date of the author of the Kaṭakaṭīkā, but there is nothing to indicate that this commentary is older than the revival of Sanskrit studies in the Deccan which dates from the 12th century. As the author invokes Kālahastiça, he must have been a native of the Telugu country.

c8.30

It is unfortunate that we have none of the older commentaries; for these would throw much light on the actual results of the work of their authors, but it is not difficult to see that as all commentators started with a knowledge of grammar and rhetoric as their means, they would gradually eliminate what they deemed ungrammatical or inconsistent with the rules of the Alaṅkāraçāstra. There is yet another cause which

c8.31

Asamgatavyākṛitipāmsupaṅkilam
Rāmāyaṇatīrthasamuddhr̥itāmṛitam |
Yogīndravāṇīkatakād vipaṅkilam
Sarvopakāraxamam astu sarvadā ||

c8.32

I quote from the Tanjore MS. The name should rather be "Rāmāyaṇāmṛitakaṭaka", but the abridged form is also used.

c8.33

Profr. Weber first pointed out¹)—the influence of a prevailing riti or ‘style’. Daṇḍin, the author of the Kāvyādarça, and who probably belongs to the 6th century A. D., recognized two such, the Gauḍī (or Northern) and Vaidarbhī (or Southern) rīti. Later writers such as Mammaṭa, Vāmana and others add four—Pāñcālī, Lāṭī, Āvantikā and Māgadhī—all of which came between the two original styles. The descriptions of these various kinds of rīti are often rhetorical and fanciful, but any observant reader of the two recensions of the Rāmāyaṇa and (e.g.) a Jain treatise in Sanskrit, will at once notice a marked difference in the style of composition and in the vocabulary of all three. Daṇḍin lays most stress on the rhetorical peculiarities, less on the linguistic; he says²):

c8.34

¹ “Über das Rāmāyaṇa”, p. 76.

c8.35

² i., 40-43 of the Bibl. Indica edition (pp. 37-9).

c8.36

Asty aneko girām mārgaḥ sūxmabhedaḥ parasparam |
Tatra Vaidarbha-Gauḍīyau varṇyete prasphuṭāntarau ||
Çleshaḥ prasādaḥ samatā mādhuryaṃ sukumāratā |
Arthavyaktir udāratvam ojaḥ kāntisamādhayaḥ ||
Iti Vaidarbhamārgasya prāṇā daçaguṇāḥ smṛitāḥ |
Eṣhām viparyayaḥ prāyo dṛiçyate³) Gauḍavartmani ||

c8.37

³ Tanjore MS. 10,511b. reads laxyate.

c8.38

Having thus distinguished the two styles, he proceeds to illustrate the peculiarities, but the first only is appreciable:

c8.39

Çliṣhṭam aspṛiṣhṭaçaithilyam alpaprāṇāxarottaram |
Çithilam⁴); “mālatīmālā lolālikalilā” yathā ||

c8.40

⁴ Do. “çithilā mālatī” etc.

c8.41

It is remarkable that the phonetic character here assigned to the Southern style would reduce the alphabet to very nearly the Dravidian phonetic system, and the alleged contrary preference of the aspirates and sibilants (which is a feature so noticeable in the N. Indian vernaculars) indicate that Daṇḍin’s remark has a solid foundation⁵), though I have not been able to prove it in spite of a careful analysis of the letters in several different passages. I have, however, noticed that the S. recension has, e.g., krūra instead of the Northern ghora. The other alleged peculiarities of the Southern style are not capable of analysis in this way, but they show accurate observation. That the influence of local style must have been powerful is a necessary consequence of the circumstances under which the later Sanskrit books were composed. The language was then dead, and cultivated for literary purposes by people of very different races and varied habits. It is difficult to imagine two languages the construction and phonetics of which differ more completely than is the case with Hindī and Tamil, and it is easy to see in the perfectly correct and grammatical Sanskrit compositions of writers of these two races that there is a perceptible difference of style, and, the Sanskrit vocabulary being amazingly copious, difference of race has led to the preference of peculiar words in different parts of India, as more harmonious¹. The authors of N. India have (except in verse) preserved a good deal of the primitive terseness of the oldest Sanskrit prose; the writers of the South (e.g. Śaṅkarācārya and Sāyaṇa) are, on the other hand, remarkable for excessively involved and long sentences chiefly constructed by the aid of ‘iti’ which is there used just like the Tamil ‘eṉṟu’. The effect of these influences is very remarkably evident in the dialects now spoken by tribes which have emigrated from N. to S. India. The Dakhnī differs from Urdū not only by a softer pronunciation but also in construction and vocabulary, and this is still more the case with the dialect of Mahrāṭhī spoken by the silk-weavers of Tanjore who (they say) originally came from Devagiri in the Deccan.

c8.42

⁵ This would tell much against the S. Indian MSS. were they not supported by early commentaries, and were it not an ascertained fact that the copyists of the Peninsula have always been more painfully scrupulous than those of the North.

c8.43

¹ Lassen noticed long ago that the Gauḍa and Southern recensions of the Rāmāyaṇa differed in unessential forms of expression. (Gorresio, i., p. xxxviii.) Schlegel had previously noticed the irregularities of form which are frequent in the Gauḍa recension.

c8.44

Lastly, come the alterations and interpolations which result from extensive religious movements. Of the earliest period of this nature we know but little; the resulting influences appear to have been eclectic and anti-buddhistic. The most important period, and, in fact, the only one deserving serious notice is that of the reforms originated by Śaṅkarācārya whose Vedānta doctrines are so well known.

c8.45

The question of such modifications as these, and the results produced by them was first treated by Burnouf in the prefaces² to the several volumes of his edition and translation of the Bhāgavata Purāṇa; he there remarked the strong Vaiṣṇava tendency not only of the Mahābhārata but also of the Bhāgavatapurāṇa, and especially the assimilation of Viṣṇu with Nārāyaṇa³. The Rāmāyaṇa also displays a strong Vaiṣṇava tendency⁴. It is evident that this is the natural consequence of the vedāntism of Çaṅkarācārya which identifies Viṣṇu-Nārāyaṇa with the Supreme Being¹); this system is by far the most popular in the Peninsula, and has had the most influence in literature. The subsequent Vaiṣṇava reforms (of Rāmānujācārya and Madhvācārya) came comparatively late, and certainly did not lead to any alterations of texts. The literature to which they gave rise consists chiefly of commentaries on original works the text of which was already settled.

c8.46

² He, however, considers that the Mahābhārata and Rāmāyaṇa have escaped to a great extent (vol. iii., p. xxix.)

c8.47

³ Do. p. xliv. The whole preface to vol. iii. refers to this subject.

c8.48

⁴ Weber, “Über das Rāmāyaṇa”, pp. 6-7. 59.

c8.49

¹ Hall, "Rational Refutation", p. 199 note, and 212 note.

c8.50

Profr. Weber's researches have brought down the earliest limit of the actual redactions of the Rāmāyaṇa to about the 3rd or 4th century A. D. and considering all the circumstances that he has discussed, it is difficult to refuse to accept this conclusion. The real point of his treatise seems to me, as yet, unanswered; for, though, no doubt, the Rāmāyaṇa has old matter as a foundation, and has incorporated many of the old verses which form the débris of the older Sanskrit literature and which are often met with elsewhere²), the latest period of redaction can only be considered so long as the text has not been critically separated into its elements.

c8.51

² Burnouf's remarks (Bhāgavata P. i., p. cx.) are worth quoting still. He says: "On commenta les Vedas; on en développa les opinions spéculatives; on rassembla les légendes relatives aux sages dont ces anciens livres faisaient connaître les noms. En un mot, on reproduisit dans un idiome plus facile et plus épuré, les opinions et les croyances dont ces livres, incontestablement antiques, avaient gardé le dépôt. Voilà pourquoi les productions des derniers âges de la littérature sanscrite ont encore un caractère si manifestement ancien. Ces productions sont, pour le fond du moins, de beaucoup antérieures à la date qu'elles portent. Leur forme seule est moderne . . . . . J'en excepte les modifications qu'apportèrent au fonds des croyances vediques les inventions des sectes qui se les partagèrent, pour les développer chacune à sa manière."

c8.52

The causes of interpolation and alteration which I have mentioned above, together with the results of the clerical errors of copyists during several hundreds of years, are quite sufficient to account for the rise of different recensions such as we now find, though originally the Rāmāyaṇa was the work of a real author. It is probable that before it was reduced into the form of a regular book, parts were current in the form of ballads³).

c8.53

³ Cfr. Weber, "Literaturgeschichte", pp. 182-3.

c8.54

*b.* Mahābhārata. Comparatively little attention has been paid to this epic, and (except by a few scholars) the uncritical Calcutta edition is generally quoted. It has been supposed⁴) that there is only one recension of the Mahābhārata; but this, I shall show, is a mistake, though it will be impossible to give more than general results here, owing to the enormous extent of all texts of the Mahābhārata, and also because I have not a complete copy of the new (or Southern) recension to refer to. The other most marked recension is that which prevails in N. India, and which appears to have been finally settled about the 16th century A. D. by Nīlakaṇṭha's commentary. All the MSS. of the Harivaṃça to which I have been able to refer, are so nearly alike, that it is pretty certain that this is a comparatively modern work, as has indeed been generally supposed. The Calcutta edition of it represents the current text very fairly.

c8.55

⁴ Monier Williams, "Indian Wisdom", p. 388. "It should be noticed that the purity of its (i. e. the Rāmāyaṇa's) text has been exposed to risks which the longer epic has escaped."

c8.56

To give briefly the chief results of such superficial comparison of the two recensions of the Mahābhārata as I have been able hitherto to make, I shall separate the text into three parts according to the evidence in their favour—1) the chief episodes, 2) the shorter, final books, and³) the longer books which are mostly at the beginning.

c8.57

1. The Episodes agree narrowly in all the MSS. As regards the Bhagavadgītā, the very complete commentary of Çaṅkarācārya, the text of which is supported even to the minutest details by subsequent glosses, amply proves that this famous poem is in exactly the same condition now¹) as it was in the 7th century A. D. There were commentaries before Çaṅkara's, but they are not now in existence²). Another well known episode—the Nalopākhyāna—presents a few various readings in the grantha MSS., but not one of these is of any importance. The ‘Samudra-manthana’ in the S. recension contains some verses which are not in the Nāgarī MSS.

c8.58

¹ Mr. C. P. Brown had a most careful collation made of Çaṅkarācārya's commentary with Lassen's text; if I recollect rightly, the result was a single v. l. of no importance! I am not able, at present, to refer to Mr. Brown's edition (in the Telugu character) which was published at Madras in 1852. This agrees with the results of my own collations.

c8.59

² See the passage quoted by Aufrecht—"Catalogus", p. 8. Çaṅkarācārya gives 700 as the number of verses, which is actually the case. Bruce and Grasberger's 'Nala' are eclectic compilations.

c8.60

2. The shorter final books. All the MSS. agree fairly well in these books (xvi., xvii., xviii.), though various readings are numerous. The differences in entire Çlokas do not amount to more than five per cent, and these are generally omitted in the N. recension.

c8.61

3. The longer books. These differ to as great an extent as the two chief recensions of the Rāmāyaṇa, they also often differ in the number of their chapters as follows:

c8.62

| | Nāgarī recension | Grantha recension |
| I. Ādiparvan | 250 | 248 |
| II. Sabhā— | 111 | 120 |
| III. Vana— | 306 | 302 |
| IV. Virāṭa— | 77 | 77 |
| V. Udyoga— | 200 | 200 |
| VI. Bhīṣma— | 118 | 118 |
| VII. Droṇa— | 198 | 198 |
| VIII. Karṇa— | 119 | 111 |
| IX. Ṣalya— | 65 | 67 |
| X. Sauptika— | 26 | 28 |
| XI. Strī— | 18 | 19 |
| XII. Ṣānti— | 364 | 363 |
| XIII. Anuṣāsana— | 252 | ? |
| XIV. Aṣvamedha— | 105 | 119 |
| XV. Āṣrama— | 46 | 45 |

c8.63

The conventional division of the Mahābhārata into 18 books is very unsatisfactory for critical purposes, as it prevents an easy separation of the several parts, and the S. Indian division into 24 is preferable. By this the Ādip. is divided into three (Ādip. = 8 ch.; Āstikap. = 40 ch.; Sambhavap. = 200 ch.), the Ṣalyap. into two (Ṣalyap. = 56 ch.; Gadāp. = 11 ch.), the Sauptikap. into three (Sauptikap. = 9 ch.; Aihikap. = 11 ch.; Viṣokap. = 8 ch.), and the Ṣāntip. into two (Rājadharma = 172 ch.; Moxadharma = 191 ch.). Other divisions can be traced.

c8.64

The following will show the differences between the two recensions such as ordinarily occur. I have chosen a fairly average instance where the texts fairly agree; to show how interpolations occur in the Nāgarī text would need more space than I can give here.

c8.65

Nāgarī. Ādip. *ch.* xi. 1-7, with which the *ch.* ends:

c8.66

Ruruḥ | *Katham himsītavānt sarpān ca rājā Janamejayaḥ* |

c8.67

*Sarpā vā himsītās tena kimartham? dvijasattama!* || 1 ||
*Kimartham moxitāc cai 'va pannagās tena dhīmatā* |
*Āstikena tad ācaxva ṣrotum icchāmi tatvataḥ.* || 2 ||

c8.68

Ṛiṣir uvāca | *Ṣroṣyasi tvam Ruro sarvam āstikacaritam mahat* |

c8.69

*Brāhmaṇānām kathayatām ity uktvā 'ntaradhīyata* || 3 ||

c8.70

Sautīr uvāca | Ity uktvā ’ntarhite tasmin yogād ṛṣivare prabho |

c8.71

Sambhramāviṣṭahṛdayo Rurur mene tad adbhutam || 4 ||
Ruruç cā ’pi vanam sarvam paryadhāvat samantataḥ |
Tam ṛṣim naṣṭam anvicchan sa çrānto nyapatad bhuvi || 5 ||
Sa moham paramam gatvā naṣṭasamjña ivā ’bhavat |
Tad ṛsher vacanam tathyam cintayānaḥ punaḥ punaḥ || 6 ||
Labdhasamjño Ruruç cā ’sthāt tac cā ’cakhyau pitus tadā |
Pitā cā ’sya tad ākhyānam pṛṣṭaḥ sarvam nyavedayat || 7 ||

c8.72

Southern. Ādip. ch. viii., end:

c8.73

Ruruḥ | Katham himsitavānt sarpān xattriyo Janamejayaḥ |

c8.74

Sarpā vā himsitās tena kimartham? ṛṣisattama! || 1 ||
Kimartham moxitās tena pannagāç cai ’va çamsa me |
Āstīkena tad ācaxva çrotum icchāmy açeṣataḥ || 2 ||

c8.75

Ṛṣiḥ | Çroṣyase tvam Ruro vatsa! āstīkam caritam mahat (sic) |

c8.76

Brāhmaṇānām kathayatām ity uktvā ’ntaradhīyata || 3 ||

c8.77

Sūtaḥ | Rurus tv atha vanam sarvam paryadhāvat samantataḥ |

c8.78

Tam ṛṣim naṣṭam anvicchan sambhrānto nyapatad bhuvi || 4 ||
Labdhasamjño Rurur utthasthau tac cā ’cakhyau pitus tadā |
Pitre tu sarvam ākhyāya duṇḍubhasya vaco ’rthavat || 5 ||
Apricchat pitaram bhūyaḥ so ’sti kasya vacas tathā |
Yat tad ākhyānam akhilam duṇḍubhenā ’tha kīrtitam || 6 ||
Tat kīrtyamānam bhagavan çrotum icchāmi tatvataḥ |
Pitā cā ’sya tad ākhyānam pṛṣṭaḥ sarvam nyavedayat || 7 ||

c8.79

Nāgarī. Ādip. ch. xii., 1:

c8.80

Çaunaka uvāca | Kimartham rājāçārdūlaḥ sa rājā Janamejayaḥ |

c8.81

Sarpasattreṇa sarpāṇām gato ’ntam tad vadasva me! |
Nikhilena tathā tatvam Saute! sarvam açeṣataḥ! || 1 ||
Āstīkaç ca dvijaçreṣṭaḥ kimartham japatām varaḥ |
Moxayām āsa bhujayān pradīptād vasuretasaḥ || 2 ||
Kasya putraḥ sa rājā ’sīt sarpasattram ya āharat |
So ca dvijātipravaraḥ kasya putro? ’bhidhatsva me! || 3 ||

c8.82

Sūtaḥ | Mahad āstīkam ākhyānam yad etāt procyate budhaiḥ |

c8.83

Sarvam etad açeṣeṇa çṛṇu me vadatām vara! || 4 ||

c8.84

Çaunakaḥ | Çrotum icchāmy açeṣeṇa kathām etām manoramām |
Āstīkasya purāṇarsher brāhmaṇasya yaçasvinaḥ || 5 ||
Sūta uvāca | Itihāsam imam viprāḥ purāṇam paricaxate |
Kṛṣṇadvaipāyanaproktam Naimiṣāraṇyavāsishu || 6 ||

c8.85

Southern recension. Ādip. ch. ix., 1:

c8.86

(Benediction).

c8.87

Çaunakaḥ | Kimartham rājaçārdūlaḥ sa rājā Janamejayaḥ |
Sarpasattreṇa sarpāṇām gato 'ntam tad vadasva me! || 1 ||
Āstīkas tu dvijaçreṣṭaḥ kimartham japatām varaḥ |
Moxayām āsa bhujaṅgān dīptāt tasmād dhutāçanāt || 2 ||
Kasya putraḥ sa rājā 'sīt sarpasattram yathā 'harat |
Sa ca dvijātipravaraḥ kasya putro? vadasva tat! || 3 ||
Çrotum icchāmy açeṣeṇa kathām etām manoharām |
Āstīkasya purāṇasya brāhmaṇasya tapasvinaḥ || 4 ||
Sūtaḥ | Mahad ākhyānam āstīkam yathai 'tad procyate budhaiḥ |
Sarvam etad açeṣeṇa çṛṇu me vadatām vara! || 5 ||
Itihāsam imam vṛddhāḥ purāṇam paricaxate |
Kṛṣṇadvaipāyanaproktam Naimiṣāraṇyavāsinaḥ || 6 ||

c8.88

In general, the result of a collation of the two recensions of the Ādiparvan is, that the Nāgarī recension has about ten per cent more çlokas than the S. recension; these çlokas generally form passages wanting in the last. Of the rest of the text, a considerable portion (numerous vv. ll. apart) is the same in both; the rest of the text presents çlokas found in the Dev. recension, but with many vv. ll. and in a totally different order. The short chapters agree generally in both recensions.

c8.89

It is obvious that I can do no more here than give the principal results of a long (though, as yet, incomplete) examination of the Tanjore MSS. I hope to give more details in my Catalogue of those MSS. The causes of this present condition of the text of the Mahābhārata appear to be the same that I have already mentioned in respect of the Rāmāyaṇa, and like it, the different recensions of the Mahābhārata appear to have sprung from a text based on a number of ballads, and then subjected to an eclectic revision, thus:

c8.90

Original collection of ballads. Eclectic text with episodes added to it. A. N. Recension (Nāgarī), a revision of the longer books, settled finally by Nīlakaṇṭha’s Commentary. B. S. Recension (Grantha), an independent revision of the longer books.

c8.91

(For Lassen’s views, see I. A-K. ii., 499 ffg.)

c8.92

How long shall we have to wait before we have critical editions of even parts of both these recensions? Even the late Profr. Goldstücker did no more than think of such a task as possible with respect to the Nāgarī recension; had he known of the existence of the S. recension, I doubt if even he would have ventured to think of such a stupendous labour¹).

c8.93

¹ There are about 836 MSS. of the whole or parts of the M. Bh. in the Tanjore Library.

c8.94

It is impossible for me here to enter on the question of the condition of the texts of the Purāṇas, because with the exception of two (Bhāgavata and Vaiṣṇava) they have been hardly studied at all. Both these Purāṇas are preserved in a single recension only, but it must be remarked that the Bhāgavata has been long suspected²) to be the work of Vopadeva (12th century), and it is thus impossible to expect a variety of recensions especially as the text has been well supported by commentaries. Both strongly support the Vedānta doctrine³).

c8.95

² See Burnouf’s edition (in folio) Vol. i., pp. lii. ffg., where the Indian tracts on its origin are translated in full. Burnouf’s conclusions appear to render this attribution to Vopadeva most probable, and this view had already been approved by Colebrooke.

c8.96

³ As regards the text of the Purāṇas, Sāyaṇa’s ‘Çankaravilāsa’ will give much information; for he there quotes whole chapters from a number of Purāṇas of which he gives the names; he also gives the numbers of the chapters.

c8.97

But the existence of several recensions of a text is not confined to the epic poetry of India; we find it to be the case with purely literary compositions by historic personages, *e.g.* Kālidāsa; the existence of two recensions of one play, the Çākuntala, has long been known, and they have been discussed by Dr. Pischel in an elaborate essay⁴). He inclines to consider the Bengālī recension to be the best. The existence of different recensions of other dramatic poems is also established. The causes of this are obscure; errors of copyists and the usual causes of various readings are insufficient to account for such differences as we find in the Çakuntalā, even though (as in this case) they have been at work for an unusually long period. In all probability these causes are to be sought in the difference of the æsthetic (or rather rhetorical) ideals of the different parts of India, and which have already been mentioned above. Dr. Pischel¹) has severely criticized Profr. Monier Williams' statement that "the bold and nervous phraseology of Kālidāsa has been either emasculated or weakened" which is certainly not what one would expect from the Gauḍa style as defined by Indian rhetoricians²); but the long passages in the third act of the Bengālī recension where (says Profr. Monier Williams)—"the love-scene between the king and Çakuntalā has been expanded to five times the length it occupies in the Devanāgarī recension, and the additions are just what an indelicate imagination might be expected to supply³)"—are certainly interpolations. Dr. Pischel has shown the origin of several of the readings of the Devanāgarī recension, and proved that they are inferior to those of the Bengālī or Gauḍa recension; yet, it is evident that the last-mentioned recension does not present an immaculate text. But there is a third recension, that of the Grantha MSS., which seems preferable to either the Devanāgarī or Gauḍa recensions; if those be taken to represent the result of the action of the Vaidarbha and Gauḍa æsthetical notions upon the original text, it is certain that the Grantha MSS. have preserved many readings preferable to either, and which do not in any way show traces of the Vaidarbha style, though the Grantha text is nearest to the Devanāgarī.

c8.98

⁴ “De Kālidāsae Çākuntali recensionibus”, particula prima, 1870.

c8.99

¹ u. s. p. 29.

c8.100

² See above, p. 73.

c8.101

³ "Çakuntalā", 1853. p. viii.

c8.102

The condition of other Sanskrit books such as the Pañcatantra is yet more remarkable, as identity of form is seldom preserved in the several recensions we have of these texts, and additions to the matter are frequent. At least four texts of the Pañcatantra are known, and Profr. Benfey has discussed these with the greatest care and minuteness⁴). The fourth (which I found in S. India) he considers, is the nearest to the original form⁵). But the Pañcatantra is not the only popular collection of stories which appears in various forms; the Vetālapañcaviṃçati, Çukasaptati and others are in the same condition.

c8.103

⁴ Benfey, "Pañcatantra, übersetzt mit Einleitung und Anmerkungen", 2 vol. 8°—1859.

c8.104

⁵ The same in the Academy, iii., p. 139 (1st April, 1872) and also in the Allg. Zeitung.

c8.105

An interesting example of the way in which these texts were modified is shown by the Bṛihat Kathāmañjarī of Xemendra, from which Somadeva adapted his Kathāsaritsāgara. Besides improving the bald style of the original, and putting the stories in a better form, Somadeva has done little but change the order of the sections as originally arranged¹; the matter and frame-work of the collection remain the same. In this case both authors appear to have been natives of N. India; Somadeva lived in Cashmere.

c8.106

¹ The first 5 books follow the same order in both texts; the others correspond as follows: vi. (Bṛihatk.)=8 (Kathāsarits.); vii.=6; viii.=11; ix.=12; x.=18; xi.=13; xii.=17; xiii.=14; xiv.=7; xv.=9; xvi.=10. The short passage I have given above (pp. 2. 3. notes) will show the difference of style. See also Dr. Bühler's paper in the Indian Antiquary, i. pp. 302 ffg.

c8.107

Collections of stanzas (as might be expected) differ in an endless number of ways, but yet the number of verses not common to the various texts is comparatively small. Examples of this are furnished by Hāla's Saptaçataka which has been so carefully examined by Profr. Weber², and also by the collection attributed to a Bhartṛihari³.

c8.108

² "Über das Saptaçatakam des Hāla" 1870. "Über d. Saptaç. Hāla" (additions) in Z. d. M. G. xxvi. pp. 735-745. "Zum Saptaç. des Hāla" in the same, xxviii. pp. 345-436.

c8.109

³ See the 'Bombay Sanskrit Series' edition, pp. 3 ffg.

c8.110

Even the Buddhist texts exhibit different recensions, as Profr. Féer has shown to be the case with the Daharasūtra⁴.

c8.111

⁴ "Journal Asiatique" viie. série, iv., pp. 297-368.

c8.112

Thus, alike, books invested with a sacred character as well as ordinary literary works, have not escaped alteration, revision, and modification in many ways.

c8.113

To multiply examples (as I could easily do) would be useless. The result, so far, of this brief summary of the enquiries into the actual condition of the texts of some of the more important of the different classes of Sanskrit works may be fairly stated thus: they exist in different recensions which have been altered under various influences—religious and eclectic, critical and the like⁵. The necessary conclusion is that it is indispensable to examine closely the texts of the grammatical treatises to see if they have been also modified in any way, or if they form an exception; for before this is determined they cannot be safely used for historical purposes.

c8.114

⁵ I think that it is obvious that the religious influences to be traced in the Sanskrit texts noticed above—viz., 1) the eclectic anti-buddhistic, and 2) the earlier Vedānta movements—are of little consequence compared with the rhetorical revisions to which nearly all Indian texts (except the Vedas) have been gradually subjected.

c8.115

It is obvious that grammatical treatises are little likely to be influenced by some of the causes already mentioned; but owing to their intricate arrangement, it is evident that they are very likely to have been amended and completed almost unconsciously; I shall, therefore, briefly consider what evidence there is as to the state of the texts of these grammatical treatises, and the alterations that may have been made in them.

c8.116

1. As regards the Çikṣās, it is only necessary to remark that the Pāṇinīyaçikṣā exists in two widely different recensions, and that it is probable that the Nāradaçikṣā is in the same state, and perhaps also the Çikṣā attributed to Bhāradvāja. The others are known only by name as yet. The brevity of these little treatises leaves small room for criticism, for comparison is hardly possible where there is so little to analyse.

c8.117

2. The case of the Prātiçākhyas is very different, and the state of these texts has been already ascertained with considerable exactness.

c8.118

a. Ṛgvedaprātiçākhya. Of all the grammatical treatises in Sanskrit that we possess, this remarkable work bears every sign of being a primitive treatise. Written in verse, especially in çlokas—a metre well suited to assist the memory, and certainly used in very early times in India—the construction of the rules is simple, and shows no trace of the ingenuity so remarkable in Pāṇini's sūtras. But with all this, the language is essentially the modern literary Sanskrit, and M. Regnier has assigned very good grounds for believing that parts are accretions as Profr. Whitney appropriately terms them.

c8.119

Thus as regards the second chapter he says: "Ce qui dans ce chapitre et dans tout le reste de l'ouvrage, laisse le plus à désirer, c'est l'ordre et la méthode, surtout pour les énumérations d'irrégularités. Elles sont, en beaucoup d'endroits, entassées comme au hasard, au lieu d'être rangées soit d'après leur ressemblance ou leur identité, soit dans l'ordre ou les règles ont été données. En outre, quelques-uns des derniers çlokas ne se rapportent pas directement au sujet traité dans le chapitre. C'est une trace de compilation, comme nous en rencontrerons, et en bien plus grand nombre, dans d'autres parties de l'ouvrage."

c8.120

Again he says of a. xiii. (iii., 1): "Le chapitre xiii. est désigné particulièrement par ce nom (Çikṣāpaṭala) à la fin de la formule qui le clôt . . . C'est, entre tous peut-être, celui qui a le plus visiblement le caractère d'une compilation. Les fragments dont il se compose y semblent jetés, vers le milieu surtout, pêle-mêle et sans ordre. On n'a pris aucun soin pour déguiser la diversité d'origine des sūtras et les concilier entre eux; mais, malgré, ce désordre, on pourrait même dire, à certains égards, à cause de ce désordre même, ce chapitre est fort curieux; il expose . . .

c8.121

des opinions diverses, et emprunte des axiomes à plusieurs écoles. Il nous permet de juger, par un exemple caractéristique, de la terminologie des maîtres antérieurs à Çaunaka, et des progrès qu'avaient faits jusqu'à ce dernier, et peut-être grâce à lui, la précision et la netteté.”

c8.122

Again of the three last chapters he says: “Le chapitre xvii a beaucoup moins d'unité et offre, ainsi que le suivant, de nombreuses traces de compilation.... Par leur sujet, comme par leur style et leur forme, ces trois paṭalas se détachent du reste de l'ouvrage, et ressemblent a un traité spécial et complet en lui-même, qu'on¹⁾ aurait rattaché après coup.”

c8.123

These admirable criticisms will show that the Ṛigvedaprātiçākhya is a work gradually built up¹⁾ in its present form; in it, we find that it is said to be a redaction by Çaunaka of a treatise by Çākalya; but though this would (in accordance with the contents) point out that the material is primitive, there is nothing (except the technical style), that would justify the attribution of this Prātiçākhya, as it stands, to a period before Pāṇini. So far as the more important part of the materials goes, this may well be the case, but the language and interpolations render much uncertain. With these reservations it entirely deserves the title given it by Profr. Max Müller “the oldest text book of Vedic phonetics.”

c8.124

That the Ṛigvedaprātiçākhya really does represent a treatise older than Pāṇini has been proved in a remarkable way. Pāṇini quotes Çākalya four times, and Profr. Max Müller shows that these rules are really the same as rules in the Ṛigvedaprātiçākhya. He has also pointed out in it other primitive matter common both to Yāska and Pāṇini. In such a case, and considering Pāṇini’s style, it is obviously impossible to expect more than identity of matter. Profr. Goldstücker took a different view, and held that this Prātiçākhya is posterior to Pāṇini. His chief reason is that the Prātiçākhyas are “more complete, and deal much more satisfactorily with their subject-matter than Pāṇini”²⁾, and that they must, therefore, be subsequent to the latter whose work they “complete and amend”. He, however, admitted that there is but little than can be compared³⁾ in the two, and this little consists of the matter referring to the changes of n into ṇ and s into sh (Ṛigvedapr. ch. v.) and the prolongation of the vowels a, i, u, (chapters vii., viii., and ix.), and a few analogous rules in Pāṇini’s Grammar. No doubt, here the Prātiçākhya is more complete, but it simply gives all the cases in a particular text of the Ṛigveda, most of which could hardly enter into Pāṇini’s scheme of a general grammar of current Sanskrit; for he appears to have simply included so much of the Vedic Grammar as was required for general purposes, and never to have intended a complete grammar of the Vedic dialects, or how is it that we meet with the sūtra ‘bahulam chandasi’ more than once¹) and in connection with different topics? Pāṇini must have known by heart the Vedic texts then current, and could have had little difficulty, if he had wished to do so, in writing as complete a Vedic Grammar as the one he has given of Sanskrit. The real cause of Profr. Goldstücker’s difficulty is that he assumed that Pāṇini’s Grammar is the Vedāṅga vyākaraṇa, whereas in fact the science of grammar in the abstract was intended by this term.

c8.125

1) As even Profr. Goldstücker admitted (Academy, July 9th 1870. p. 272.) There are also a few probable interpolations, e. g., where Çaunaka's opinion is discussed.

c8.126

2) Do. p. 271; “Pāṇini's Place,” p. 199.

c8.127

3) “Pāṇini's Place,” pp. 198-9.

c8.128

¹ It occurs eleven times! Cfr. Benfey’s “Einleitung” p. 3.

c8.129

It appears to me that, for these reasons, Profr. Goldstücker’s arguments that the Prātiçākhyas are more recent than Pāṇini, and are intended to supplement his Grammar, do not apply to the Ṛigvedaprātiçākhya; of it, we probably have a revision subsequent to Pāṇini, but, there can be little doubt, the greater part of the text and the matter contained in this work are far older²). His arguments have much greater force in respect of the three other treatises which belong to the same class³).

c8.130

² Lassen (“Indische Alterthumskunde” 2nd ed. ii., p. 478) unhesitatingly attributes the Ṛigvedapr. to Çaunaka whose date he puts at 460 B. C.

c8.131

³ Westergaard (Über den ältesten Zeitraum der Indischen Geschichte p. 67) considers that the Ṛigvedapr. is later than Pāṇini; he grounds this opinion on the character of the versification. Pischel (“De Grammaticis Prācriticis” Theses) upholds this view.

c8.132

b. Taittirīyaprātiçākhya. Profr. Whitney has already⁴) pointed out a number of accretions and interpolations in this treatise, such accretions are chapters xvii., xviii., xxii., xxiii., and xxiv. On the other hand the unusually numerous references to discordant opinions of the authorities are considered by Profr. Whitney to be interpolations made in the text after it “ceased to be a mere body of practical rules for the guidance of a school, and in virtue of its thoroughness and comprehensiveness gained more the character of a phonetic ‘treatise’ on the Black Yajur Veda, and was used in other schools than that which originated it”. Profr. Whitney considers that the sūtras i., 25-7, 60; vii., 13-14;

c8.133

⁴ See p. 432 of Profr. Whitney’s edition of the Taitt. pr.

c8.134

xiv., 12-13; xv., 6-9; xvi., 24 are interpolations, as they express rules not indispensable to the work, and evidently in wrong places where they interrupt the natural connection. He also considers that the parts referring to prosody are interpolations; this may be so here, but I have already given reasons for believing that prosody was an essential part of the old Sanskrit Grammar.

c8.135

The Taittirīyaprātiçākhya is composed in a highly developed sūtra style; this is evidence of a revision in comparatively recent times.

c8.136

c. Kātyāyanaprātiçākhya. In addition to the reasons already given for believing that the Ṛgveda and Taittirīya prātiçākhyas are not preserved in their original form, and which apply equally to the Kātyāyanaprātiçākhya, there is, in the case of this last and of the Atharvaprātiçākhya, positive proof that they have been brought into their present form at a period later than Pāṇini; for they show an acquaintance with the anubandhas he uses, and which, Patañjali says, were invented by him.

c8.137

Thus we find tiṅ (i., 27), āṅ (vi., 24 ?), luk (iii., 12), lup (i., 114), et and ot (i., 114; iv., 58) as indicated by Profr. Weber¹, and the first alone of these, tiṅ, is sufficient to place this Prātiçākhya after Pāṇini, as was pointed out by Goldstücker².

c8.138

¹ “Ind. Studien”, iv., p. 83.

c8.139

² “Pāṇini’s Place”, p. 207.

c8.140

Profr. Weber also pointed out³ the close connection between the wording of the sūtras in this Prātiçākhya as compared with Pāṇini's, and Profr. Goldstücker attempted⁴ by a minute criticism to show not only that the Kātyāyana to whom this Prātiçākhya is attributed is the same person as the author of the criticisms on Pāṇini's sūtras, but also that he wrote the former before the latter. He assigns two reasons for this view: 1) that the Vārttikas merely include criticisms on sūtras of Pāṇini which were not noticed in the Prātiçākhya, and 2) that some of the Vārttikas are obvious improvements on sūtras in the Prātiçākhya. But an enquiry of this nature is hardly possible at present, nor will it be so for a long time to come.

First 140 of 326 passages. Read on

Linked texts (45)