Amrita Collected Papers by A M Ghatage

Author(s): Jitendra B Shah
Publisher: Kasturbhai Lalbhai Smarak Nidhi Ahmedabad
Catalog link: https://jainqq.org/explore/006968/1

Page #1

Shresthi Kasturbhai Lalbhai Series-1
AU
COLOX
MILF
VODA
Amrita
The Collected Papers Contributed by
Prof. A. M. Ghatage
Shresthi Kasturbhai Lalbhai Smarak Nidhi
Ahmedabad

Page #2

AMRITA (The Collected Papers Contributed by Prof. A. M. Ghatage)

Page #3

SHRESTHI KASTURBHAI LALBHAI COLLECTED RESEARCH - ARTICLES SERIES
General Editor Jitendra B. Shah
VOLUME-1

Page #4

Amrita
( The Collected Papers Contributed by Prof. A. M. Ghatage )
Shresthi Kasturbhai Lalbhai Smarak Nidhi C/o. Sharadaben Chimanbhai Educational Research Centre
‘Darshan' Opp. Ranakpur Society Shahibaug Ahmedabad-380 004 (Gujarat State) INDIA

Page #5

Amrita (The Collected Papers Contributed by Prof. A. M. Ghatage)
Published by
J. B. Shah Shresthi Kasturbhai Lalbhai Smarak Nidhi C/o. Sharadaben Chimanbhai Educational Research Centre
'Darshan' Opp. Ranakpur Society Shahibaug Ahmedabad-380 004 (Gujarat State) INDIA
Price : Rs. 600/
Printer Chandrika Printery
Mirzapur Road Ahmedabad-380 001 Phone : 5620578

Page #6

Foreword
With this Volume containing the collected papers of Prof. A. M. Ghatage, this Centre inaugurates its Series for publishing the collected research papers and articles by eminent scholars in the field of Indology/ Jainology. We hope that the academics in the research fields in which Prof. Ghatage had worked for long decades (and is still working) will find it a useful book for reference.
Shrenik Kasturbhai

Page #7


Page #8

Prefatory Note
Prof. M. A. Dhaky and I had met Prof. A. M. Ghatage, one of the preëminent and senior Indologists in India, at the Bhandarkar Oriental Research Institute, Pune, in mid May 1997. We then had expressed our view on the desirability of, as well as our willingness to, publishing a volume containing all his important papers, articles, introductions, etc. Most of these are scattered in various research journals and other academic publications, indeed not always easy to lay hands on, spread as they also are over long decades of his scholarly endeavours. And most of these are valuable as reference for further research on the topics concerned. Prof. Ghatage, to our delight, agreed to our suggestion. It took about two years to bring it to the final shape of production at our end. Centre's senior proof-reader Shri Naranbhai Patel had read the proofs just as Prof. Ghatage, too, had read these at his end. Smt. Purvi Shah did the data-entry and Shri Akhilesh Mishra of the Centre effected the laser printing. We are delighted to make the Volume available to the Indologists/Jainologists as well as to the experts on ancient Indian linguistics. The biographical sketch of Prof. Ghatage has been published in the Sambodhi XXII, Ahmedabad 1998, which is planned to be presented at the occasion of the release function for the present Volume, to Prof. Ghatage for his dedicated work in the sphere of Indological studies, spanning as it does over six decades and a half.
Jitendra B. Shah
Director

Page #9

Contents
89
95
149 169 181.
191
207 216 223 233 242
252
Group 1: Prākrit Languages
1. A Basic Tendency of Prākrit Languages 2. The Apabhramsa Language 3. Groups of Two Mutes in Middle Indo-Aryan 4. Repetition in Prakrit Syntax 5. Concord in Prākrit Syntax 6. Instrumental and Locative in Ardha-magadhi 7. A Prākrit Tendency in Rgveda
8. An Unassimilated Group in Apabhrama Group 2: Prākrit Literature
9. The Title Daśavaikālika Sutra 10. Māhārāstri Language and Literature 11. Sauraseni Prākrt 12. The Daśavaikälika-Niryukti 13. The Sūtrakrtānga-Niryukti 14. Narrative Literature in Jain Māhārāstri 15. Didactic Works in Prakrit 16. Hymns in Prākrit 17. A Few Parallels in Jain and Buddhist Works 18. Parallel Passages in the Daśavaikālika and the Ācārānga 19. Kundakunda and his Philosophy 20. A Passage from Haribhadra's Samarādityakathā
21. Pārsva's Historicity Reconsidered Group 3: Sanskrit Language
22. Patañjali on P. VIII. 1.1 23. Panini - VIII. 1.7 24. Idiom-formation and Panini 25. Systematics of Panini's Astādhyāyi 26. Uktarthānäm Aprayogah 27. Pronunciation of Sanskrit 28. Traces of Short ě and > in Rgveda 29. Panini 1. 4. 32 30. Some Etymologies in Manusmrti
31. A Note On Utkalāpaya Group 4: Linguistics and Lexicography
32. The Text of the Tatvārthādhigama-süträni 33. Phonetics And Phonemics in Historical Linguistics-I 34. Phonetics And Phonemics in Historical Linguistics-II 35. General President's Address 36. Indian Linguistics 37. Dictionary of Sanskrit on Historical Principles 38. The Five Mahābhūtas 39. Key-note Address for a seminar on Lexicography 40. A Comprehensive and Critical Dictionary of the
Prakrit Languages
255 259
261
263
266
273 278 289 291 311 316 323 325 327
334
340 347 366 376 457 467
474

Page #10

Group 1 : Prākrit Languages

Page #11


Page #12

A Basic Tendency of Prakrit Languages
Compared to Sanskrit, the Prakrit languages show a bewildering variety of changes in their phonology, morphology and syntax, and produce an impression of artificiality by the extent and regularity of such changes. On the other hand, we find them used for a vast literary activity with the avowed purpose of coming closer to the speeches of the populace. The way to reconcile these two positions can only be found in an attempt to trace some fundamental principle of linguistic change at the basis of all changes introduced in the Prakrits, which would render them more intelligible.
One such principle we may hope to find in the phonology of the Prakrit which would make us understand how such changes have occurred and made possible. In the whole range of the bewildering changes of sound both vowels and consonants, which one meets in the Prakrits, there appears to be one thing which remains constant and guides all these changes. This principle can be briefly formulated as the tendency to preserve the syllabic quantity of a word. We may now examine the working of this tendency in Prakrit phonology to appreciate the full extent of its operation and effectiveness.
The best illustration of this tendency may be found in the changes of the conjunct consonants. Thus whether the assimilation observable in them is progressive or regressive, the syllabic nature of the word remains in tact. So Sanskrit tapta becomes tatta and yatna may change into jatta, but the syllable values of the words are in no way affected. They continue to have a long syllable followed by a short one. In fact, a change involved in tapta becoming tatta can only be explained by a consideration of the following kind. The usual explanation that the first consonant -p-, being purely implosive, was less audible than the explosive second consonant -t- and so it had a better chance of surviving as it actually does and assimilates the first, is not the whole truth. We find not only the loss of the first mute but also the germination of

Page #13

Amrita
the second and say that it has assimilated the first. But what exactly do we mean by this assimilation ? Obviously it is not that the mind of the speaker regard the first stop as a -t- by mistaken identity, nor does his anticipatory movement of -t satisfies him for the sound of -p-, though this is the starting point of the change. If the speaker had known that his inability to pronounce the sound -p- were harmful to the nature of the word, it would have remained a man mistake and would have been duly corrected. What appears to happen actually is that the mind of the speaker has identified the essence of the word not so much with the acoustic effects of the individual sounds making up the word, a feeling for -p- and that for-t- following a short vowel but more with the sequence of two syllables, a short vowel sound followed by a group consonants so as to render it heavy by position. This essential nature of the word impresses the mind to such an extent that it forms the focus of attention at the time of uttering the word leaving all other constituents vague and unimportant. In other words the syllabic structure of a word is for the speaker a distinctive variant while all others are nondistinctive. Once these others are freed from the essential nature of the word they have less chance of survival and drop out to the extent that they in no way affect the essential nature of the word. Thus after uttering the first syllable which is by nature short, the speaker has a vivid consciousness that it must be followed by a group of consonants or more accurately by a long consonant so as to give it its required syllabic value of length by position. When the stop is uttered, the choice falls naturally on the explosive -t- as clearer of the two and the tendency is to lengthen it to satisfy the demands of the preceding syllable and when once it is satisfied the speaker has no further inclination of reproducing the other elements of the original word.
The same principle appears to explain more accurately the cases of assimilation where the original conjunct is made up of a mute and a fricative, a liquid or a semi-vowel. When a Sanskrit word like putra becomes putta, or a word like tatra becomes tattha or even a word like satya becomes sacca we are accustomed to call the changes as ordinary cases of assimilation and we say that the mute assimilates to itself the other sound. But the actual fact is more complex. In the first place the Sanskrit orthography does not appear to represent the actual sounds of the Sanskrit words in their totality and is to some extent responsible for the usual view of regarding them as cases of assimilation. It is easy to see that a conjunct like -tr- or -ty- is not strictly of the same type as -kt -or-tp- where two mutes are involved. While in the latter type of groups there will be a syllabic division in the body of the

Page #14

A Basic Tendency of Prākrit Languages
conjunct consonant separating the two sounds into two syllables marked by the plosion between the two, no such syllabic division is possible in the earlier type. In fact -tr -is nothing but the sound -t- the explosion of which takes the form of the sound -r- and a conjunct like -pl- has the explosion of the mute -p- in the form of a lateral sound. -ty- would normally be a palatal -t- sound. In all these cases the sound would not be such as can be strictly called a conjunct which can be split into two parts the first of which attaches itself to the preceding syllable and makes it a close one. In other words, if the sounds are exactly what they are written, they would not make position and make the preceding syllable long. That some of the Sanskrit groups were and are of this type is evident from the fact that groups like -tr, .pl- can begin a word in the language as in trāyasva, plavate or tyāga. But when such groups occur in the body of a word they necessarily make position and are always so treated. This can only be possible on the supposition that in actual pronunciation they were real conjunct consonants and involved a long consonant as the first member. In ordinary orthography they were as good as * puttra, *'sattya, *vipplava and so on. Such a sound of these words can also be inferred from the fact that there was no difference of sound in the groups of words like sattra and putra though, following etymology, we write sat-tra (from sad- and tra) but pu- tra(from pu- and tra). In view of the rules of doubling given by the Sanskrit phoneticians, which pertain more to sounds than to writing, it cannot be imagined that the long consonant in sattra was shortened, and we are led to think that the simple -t-in words like putra was germinated. And this is in full agreement with the actual sounds. Once this thing is clear we can see that the change of these groups into Prākrit -tt-, ppor -cc- does not involve a case of assimilation but only that of simplification. The double consonant is found sufficient for keeping the syllabic structure of the word and the following peculiar type of explosion is superfluous for that purpose, with the result that it gives place to the normal type of explosion. In short, it disappears leaving behind a long consonant. Such a supposition alone can explain the change of ty- to cc-where the consonant itself is palatalized.
This principle is best verified on a limited type of change like the development of a glide of a sound like -b- in words of the nature of Sanskrit tāmra, āmra becoming Prākrit tamba, amba etc. That in a group like mr-, mla glide sound like -b- is prone to arise is obvious from the physiological point of view. It is the natural result of the lack of a very fine adjustment of the vocal organs; where the soft palate goes up a moment too soon and gives rise

Page #15

Amrita
to the glide. This same glide is observable in the change of Anglo-Saxon slumerian into English slumber and Latin numerum into French nombre. Though the two cases are quite parallel as regards the development of the glide the change illustrated by English and French shows a vital difference from the change in Prākrit. In both the European languages the two members of the group (for we must assume an intermediate step where the groups arose) are kept along with the glide in the words slumber and nombre, but the Prākrit words drop one of the members of the original groups. The preservation was possible in the earlier case because the English word has developed a new syllabic sound (a) while in French the nasal sound has only nasalized the preceding vowel. On the contrary in the absence of both these possibilities, in the Prākrits the group of two consonants was all that was needed to preserve the syllabic nature of the word and quite naturally the additional sound of -s- or -- was lost. From this it is but an obvious deduction that such a conception of the essential nature of a word would not allow a group of three consonants, which is actually the case in Prākrit. Similarly a conjunct at the beginning of a word served no useful purpose for the syllabic structure of the word and was uniformly lost.
This very principle would explain that striking change of dropping most of the intervocalic consonants which gives these languages their distinctive appearance. As in other languages, notably in "French, the intervocalic consonants became voiced, turned into spirants, and as spirants were rare in Indo-Aryan, were finally lost. But their loss was in no way detrimental to the conception of the word as viewed by the speakers who stressed above all the number and sequence of the syllables which were kept in tact in spite of the loss of the consonants. As compared to the vast number of words where the syllables are preserved, the cases showing the loss of syllables are quite negligible and most of them are simple cases of contraction. It is only when the Middle Indo-Aryan period is over and the New Indo-Aryan period has begun that we find a change in the idea of the nature of the word and the consequent loss of syllables accompanied by other changes.
Most of the vowel changes follow the same principle. It is obvious that the diphthongs -ai- and -au- and the long vowels -e- and -o- are not different in their metrical length and following a primitive Sanskrit tendency the former are reduced to the latter in Prākrits. But more interesting is the
tion of two new sounds, the short -e- and -o- under the pressure of the same tendency. In Sanskrit the long -7- and -7- when followed by groups of

Page #16

A Basic Tendency of Prakrit Languages
consonants preserved their long quantity. Thus in a word like vestana or ostha the vowels of the first syllables were necessarily long and were so pronounced. But these words were in no way different as regards quantity from words with the vowels -- and -u- with a following group of consonants as in mitra or ustra. The qualities of the vowels were no doubt different but the metrical value of the words was the same in spite of the difference of quantity of the vowels in the first syllables. Now the difference between the two can only be about the length of the following conjunct consonant. If -eand -- were pronounced long in the first two words and -- and -u- were pronounced short in the other two and yet the words had the same metrical scheme, it follows that in the first group the first member of the group was of a shorter duration than in the second set of words. This was possible because the first member of the group in Sanskrit was able to show some amount of variation in its length to preserve the quantity of the preceding vowels distinct. With the assimilation of the groups in the Prākrit stage no such possibility existed and the natural result was that the preceding long vowels were shortened. They, however, preserved their distinctive quality and resulted in short -ě- and -Õ-.
A host of other changes will be found to confirm this principle. Thus cases of anaptyxis like śrī = siri, sūksma = suhuma, gemination of consonants like taila = tella, khāta = khatta, duküla = dugulla, simplification of groups like varsa = vāsa, gātra = gāya and most of the Sandhi rules can be taken to illustrate this principle.
One such change based on this principle is of greater importance in the explanation of the morphology and syntax of the Prākrit languages. It is the regular alternance between a long vowel and a short vowel with an anusvāra after it. Both have the same metrical value and both appear to alternate with each other as a purely phonetic variant. Cases of spontaneous nasalization like vayasya - vayassa, aśru = aṁsu and changes like vimsati = vīsā, simha = sīha fall under this alternance.
This alternance would explain a number of individual words which are otherwise obscure. So in Pāli akamsu for akārsuh, bādh- in the sense of bandh- 'to bind', siyam for siyā as potential third person singular of as-, sirimsapa for sarīsspa, niramkatvā for nirakrtvā, the proper name vaṁgisa which may be the same as vāgūša, the form caṁki probably Sanskrit cakrī, khaluñka from Sanskrit khaloksa, bhimsanakam for bhisanaka, sanamtano for sanātano and in Ardha-Māgadhī vikanthayai for vikatthate, sambali for salmali,

Page #17

Amrita
ghimsu for grisme, samdāsa for samdama and a number of other words.
More interesting is the fact that this type of alternance makes a number of anomalous constructions in the Prākrit syntax quite understandable and regular. Thus chamā for the regular chamam of the Acc. sing. would explain Pāli sentences like tattheva nipatim chama 'she fell there on the ground' bijāni pavapaṁ chama 'sowing seeds in the ground', pāsādatale chamā patitam 'fallen on the ground or the floor of the palace' which is comparable to the sentence occurring soon after disvāna chamam nisinne which would make the equation of chamā and chamam quite apparent. Similarly often a form of the Acc. appears to be replaced by the form of the Nom. because of this alternance. So Pāli : imā girā abbhudīresum 'they uttered these words', Ardha-Māgadhi itthi pumam pavvaiyam gihim vā l. On the other hand the long vowel of the Nom. is replaced by the short vowel with an anusvāra which produces the semblance of a form of Acc. Thus Pāli : tam bhūmim rāmaneyyakaṁ | for să bhūmi, etc. Ardha-Māgadhi : nāsanti appana param ca natthā | or tāran appāna param ca tinna
Two very frequent constructions are best explained by this alternance. Thus the use of sakkā as a predicate when the subject is Neuter or an Infinitive is only a phonetic variant for the regular form sakkam. So in Pāli : na sakki balimuddhattum | dubbacanaṁ kim sakkā kātuye na ca sakkā aghatamānena | AMg. sakkā saheum āsāi kantagā | Equally frequent is the use of attha for attham in AMg. to express the purpose of an act. Thus we read appanatthā paratthā vā | or annassa atthā ihamāgao mi . This is probably the real explanation of the apparent use of the Acc., where we should expect an Abl. which ends in 2. Thus Pāli : kāla kālam bhavā bhavam akatam dukkatam seyye or Amg. jai param maranam siya.
A further investigation would reveal many such facts both in the morphology and syntax of the Prākrit languages which would find their explanation in such phonetic alternances based on the fundamental principle of syllabic quantity. That this principle would also shed some light on the problem of derivation can be seen in the explanation of the plural forms of the Neuter nouns like phalai or mahūi which correspond to Sanskrit phalāni and madhni. The equation is often denied on the phonetic ground that a loss of a name should nasalize the preceding vowel and not the following one as it does in these cases. One can compare the development of the French nasal vowel in cases like chanter from Latin cantare or vent from Latin ventus. That in the present case the nasal can nasalize the following vowel can be

Page #18

A Basic Tendency of Prākrit Languages
explained on the ground that it was always kept in its own syllable, and the persistence of all the three syllables and their separateness may explain the nasalization of the final syllable and not the preceding one. In the parallel cases from French, one can see that the nasal nasalizes the preceding vowel only when it is followed by another consonant and thus properly belongs to the preceding syllable. In other cases we find that Latin amare gives rise to French aimer.
The investigation of this one principle underlying a number of phonetic changes would suggest that in the apparent welter of linguistic changes of Prākrits there does run as an undercurrent some well-marked principle which cannot be possible in a group of artificial changes produced by grammarians and literary men. On the contrary it suggests strongly that it was the result of the unconscious tendency of the speaker to value the syllabic structure of the word more than anything else. This is probably the strongest proof in favour of regarding the Prākrit languages as having a natural origin. Because they have been preserved to us only in literary documents they are bound to show some deviations from the actual spoken forms on which they are based.
000
A Basic Tendency of Prākrit Languages B.C. Law Vol.2 1946
5

Page #19

The Apabhramśa Language
At the beginning of this century our knowledge of the Apabhramsa language was confined to its description found in the Prākrit grammars of Hemacandra and others, stray remarks about it in Sanskrit works on dramaturgy and poetics, a few illustrative examples cited by them, and the verses in the fourth Act of Kālidāsa's Vikramorvašīya. Since 1918, however, a large number of works in Apabhramsa was brought to light due to the labours of a number of Sanskrit and Prākrit scholars and today we possess a sizable literature in it along with numourous attempts to describe the language both historically and descriptively. Yet much remains to be done in explaining its linguistic structure and its place in tracing the origins of the New Indo-Aryan languages. In fact, a number of problems connected with it, like its exact relation to the Prākrit dialects of the earlier stage, its connection with the socalled Avahattā and the Deśībhāsās, its different dialects as noted by the Prākrit grammarians and revealed by the writers of Apabhramśa works and its precise relation to the New Indo-Aryan languages must be resolved in order to follow accurately the development of the Modern Indian languages as a whole. Before any of these problems can be usefully discussed it is necessary to have a general sketch of the language as far as it can be ascertained with the material available to us. For this purpose, the material found in the grammar of Hemacandra is far more useful than the larger Apabhraíša works, because of its variety both in contents and form and ampler traces of dialectal variations and freedom from the influence of the Prākrit dialects from which much of the Apabhramśa literature suffers.
The vowel system of Apabhramsa can be stated as follows :
e
0

Page #20

The Apabhramsa Language
This gives us a system of five vowels with a threefold distinction of tongue-height and a twofold distinction of place. In some dialects and according to some grammarians the presence of a vowel r is also attested but all the evidence of orthography, metrical considerations and rhymes suggest that in all such cases the actual sound system showed a consonantal phoneme r followed by a front high vowel i. Thus a word like priya is written as prya and words like śrī and strī are written as srya and trya. Sometimes the same word is written in both ways as in case of priu and pru, priyahu and pryahu. Kriya rhymes with srya and prya with piu and piya with srya. When we further note that an initial cluster with r as the second member was also often kept or assimilated in the same Apabhramśa dialect we can easily understand such pairs as piya and srya or piu and prya. What is beyond doubt is the vocalic element involved which was invariably a high front vowel.
The feature of length in vowels gives rise to a complex problem in this dialect. Phonetically speaking all vowels occur with and without length and in this regard the situation is the same as in the older Prākrits including Pāli and different from the Old Indo-Aryan dialects. But the phonemic status of length is not quite clear. In two types of environment length is not distinctive and some kind of neutralisation is to be seen. Thus before long or germinated consonants all vowels are short, no long vowel being allowed here. In nonfinal open syllables the mid vowels e and o are always long, which is an invariable situation in case of the first syllable of a word but allows some exceptions in the penultimate syllable if it is not a part of the stem. The exceptions pertain only to the front vowel e and not to the back vowel o.
In other environments length is phonemic. The vowels a, i and u contrast with their long counterparts ā, ī. ū regularly in the non-final syllable, but only to a limited extent in the final syllable. Compare manu with mānu, tiha with diha and kuru with cūru. The contrast in the final syllable can be seen in cases like dhana but sāmalā, hiai but gamihi. Cases of contrasting u are rare : tanu but sāhū (Par. 2.10). This contrast in monosyllabic words is confined to a and a but not to i or u. We have mä, sä, jä but also ma, na, ya, but only ki, vi, ji, or ju, tu, su.
Vowel length in case of e and o cannot be shown to be phonemic. The bulk of Apabhraíśa literature is metrical and hence the final syllables of the lines are anceps where e and o occur most frequently. They can be read as

Page #21

12
Amrita
either short or long according to the demands of the metres used and do not help us in deciding the phonemic length of these vowels. In the interior of a metrical line we find these phonemes either in the body of a word or at its end. Again at the end of a word these vowels can be nasalised or not, giving rise to the following type of distribution as regards length.
Monosyllables with the vowels e and o are always treated as long. Such are words like jo, to, je, te etc. If they are treated as short they are uniformly written as ji, ti etc. and will be considered as having the phonemes i or u. In the interior of a word they are always long and usually without nasalization. At the end of a word but inside the metrical line we find them both short and long but with the following restrictions. In case of e it is long if there is a nasalization accompanying it but short when without it. Only in a case of a few forms where e occurs in the penultimate syllable and without nasalization, it is found to be short. The vowel o is without nasalization and is always long in the interior of a word but short at its end. It does not occur in the penultimate syllable in a comparable situation with e and is not found short. Scanning the 123 dohās of Paramappapayāsu constituting its first chapter we find that in open syllables there are no cases of short ő and ē is always long. A fine case of contrast is seen in the two lines of the same dohā 1.34
short ě : dehě vasantu vi navi chivai long ē : dehẽ chippai jo vi ņavi.
From the evidence adduced above it is difficult to prove the phonemic length in case of e and o. If we look into the Apabhramśa verses quoted by Hemacandra in his grammar we find such cases as vankehi, payārēhi, tumhēhi, amhèhi which show a contrast of — with long e and tarunahỏ, loyahð ehỏ with short o at the end of words in the interior of a metrical line.
Some interesting historical conclusions can be drawn from the structural analysis of vocalic length in Apabhramśa. Length in a and a has been distinctive throughout and is protected by their basic qualitative difference. In case of i and u the final contrast is due to a later stage of development caused by a contraction in the last two syllables of the loss of the final syllable with a compensatory lengthening of the penultimate syllable." In monosyllabic words the basic Apabhramśa phonology shortened all vowels and forms like ma, sa are borrowals from the earlier Prākrit stage. In case of monosyllabic words in e and o the Apabhraṁsa phonology merged them with

Page #22

The Apabhraísa Language
13
final i or u and long forms like jo to or je, te are again Prākritic borrowings. In forms like amhēhi, payārēnā etc. the phoneme in the penultimate syllable was either i or a and the writing is a mere imitation of the Prākrit forms. If we thus sort out the earlier survivals and the later developments we find the basic Apabhramśa phonology following the rule that vowel length was significant in the non-final syllables in case of a, i and u and nowhere else. This is in remarkable agreement with the Old Indo-Aryan system except that final length has disappeared. An unexpected result of this position is the chance to judge correctly the material which is handed down to us in scribal tradition and to reinterpret it in a linguistic sense. Thus in Paramappapayāsu chapter II we must interpret niyamim in 18 and mohim in 53, 55 as really standing for niyamē and mohē which has the further advantage of getting rid of final anusvāra which is so rare in Apabhramśa. So also in Hemacandra 343 we should read aggim in to tē aggim kajju as aggē. We are now in a better position to postulate three stages in the growth as regards the vowel length; a pre-Apábhramsa stage (Prākrit), the Apabhramśa stage proper and a postApabhramba or early New Indo-Aryan stage.
Nasalization in vowels is phonemic, but is confined to ā, i, ū and ě, among which ě is always long. Compare forms like tanahā, sokkhahā, mai, tahi, rayanai, hau, tuhū, duhū and daivě, daiē, kajjă etc.
It is much easier to set up the consonantal system of Apabhramśa. We have the following phonemes :
ctt. P ch th
kh
gh
jh
y
The following examples will prove their presence and status :
kuru, campa, tarkara, taņu, para, khagga, cholijjantu, thavai, thali, phedai, goradi, jehau, dongara, dadavada, bappiki, ghodā, jhaivi, dhollā, dhana, bhuvana, nāi, navi, mānu, raha, lacchi, hasiu, sāmalā, sāyaraho, viraha, jīva.

Page #23

Amrita
The limitations on their occurrences are also clear. Initial ț and y are rare or do not exist. There is free variation between n and ņ. Both ń and ñ have no phonemic status and can be easily assigned to n.
The germinate consonants can be considered as clusters along with clusters of unaspirated and aspirated stops and clusters of nasals and the homorganic consonants. Their examples are : (1) CC :
pakka, khagga, bhiccu, jajjariāu, phutta, boddia, rattadī, niddae, uppari, nibbalu, suvanna, annu, sammānei, dhollā, parassu, nivvahai. There are no germinates of h, r and y. (2) CCH
lakkhu, nigghina, vacchahe, majjhahe, ditthi, daddhā, patthari, muddha, tadapphadai, abbhuddharanu. (3) NC :
vanka, sangara, sankhaha, 'sandha, ghanghalai, kañcana; kuñjaru, ghuntehī, ganthi, khandai, pavasanteņa, panthia, sundara, sandhihi, campaya, vilambu, kumbhai.
As there is no contrast between a cluster of an aspirated stop and an unaspirated stop after a simple stop it is possible to consider them as germinates of aspirated stops, which is also the practice of some writing systems current in the Mss. In fact a difference of writing in kkh and khkh has no linguistic significance. Phonologically speaking it may lead to greater economy to favour the second solution and the same may be true of using the solution of a consonantal cluster of a nasal stop and a homorganic stop following it in place of using the phoneme of anusvāra in place of the nasal. Both these linguistically inferior methods are, however, favoured by the Middle Indo-Aryan philologists.
The case of the other aspirated consonants like the nasal n and m or the lateral l is interesting in another direction. The initial clusters like lh or mh or nh do not contrast with the medial Ilh or mmh or nnh and can be considered as the same phonemes and in view of their clustering habits may be considered as aspirated unit phonemes and should be added to the list above as
nh mh lh.
This will make them exactly parallel to the other aspirates like initial

Page #24

The Apabhramśa Language
15
kh but medial kkh etc.
There remain a few more cases of nasal sounds which are neither nasal vowels nor nasal consonants. Such cases are amsu, daṁsana, samvarevi etc. where the nasal resonance is followed by eithers or v. In a strict synchronic sense we must set up a phoneme of anusvāra for them with a limited distribution. But we can as well conclude that this also is a case of archaism from the earlier Prākrit stage or possibly a mode of writing of a possible nasalised long vowel which was its later development in the New Indo-Aryan stage. A nasalised ✓ phoneme is well attested and is graphically represented as my or an anusvāra on the preceding syllable.
In addition, Apabhramsa shows a limited number of consonant clusters with r like priu, ghrum, broppinu, prassadi, vratu etc. which is certainly a dialectal feature known to the Prākrit grammarians and met with in some specific books or families of Mss. A similar dialectal feature is seen in the few examples.cited by Hemacandra in which the medial voiced consonants like g, gh, d, dh, b, bh are kept and to which the name saurasenism is given.
The other parts of Apabhramśa grammar are very briefly summarised here. The general effect has been to reduce the morphology of the language to the minimum. The declensions of nouns tend to merge into each other and we have at most only three types with the endings a, i and u. The distinction between Masculine and Neuter is disappearing and the formal difference between these and the Feminine is also slight. Thus the Ablative singular of a number of different words would run formally alike : narahe, phalahe, girihe, muddhahe, devihe etc. The distinction among the cases is also breaking down and we can at best set up such syncratic groups as NominativeAccusative, Instrumental-Locative and Ablative-Genitive. This is very close to the distinction between the Direct and Oblique of the NIA. What is new is a kind of novel pattern between the singular and the plural of the same case which is marked by the mere absence or presence of the nasalisation. Compare Gen. naraha and naraha, Abl. narahu and narahü, Inst. narahi and narahi, Loc. girihi and girih i. There is also a tendency towards vowel harmony between the stem vowel and the vowel of the inflection. Compare girihi, guruh ū and narahã. Many of the pronominal forms are assimilated to the nominal type and most of the deviations, like amhe, amhi, amhāsu, tena, tāo are either borrowals from Prākrits or less likely archaisms.

Page #25

16
Amrita
In the verbal system there is really a single conjugation and the difference between the present and future, primitive and causal or active and passive is indicated by a difference in the stem : karahi, karesahi, karāvahi, karijjai. Among the participles the potential passive has already become a verbal noun and the gerund and the infinitive of purpose have merged together. The syntax is greatly simplified and postpositions begin to appear in conjunction with the remnants of the case forms. The so-called Deśī words occur with greater frequency and the absence of a conjunction like ya is peculiar. The sentence-types are scarcely different from those in use in the early stages of the NIA languages.
ODD
The Apabhramsa Language
Seminar in Prakrit Studies, Pune 1969
'.6

Page #26

Groups of Two Mutes in Middle Indo-Aryan
THOUGH it is natural to suppose that the assimilation of various consonant-groups was a slow process, showing different stages in its growth, before it reached its maximum extent and scope, which it did in the period of Classical Prākrits, it is not always possible to note all such intermediate steps in the history of the MIA. languages themselves. Like all linguistic changes, assimilation has its own period of growth, and it cannot be rigidly confined to any one given stage which may be conveniently postulated in the growth of a family of languages like the Indo-Aryan. However careful we may become in choosing a number of features as marking the MIA. stage as distinguished from the OIA., it is bound to result into some kind of arbitrariness, and these distinctions are to be accepted more for the sake of convenience than for their strict accuracy.
Usually, assimilation of all kinds of consonant-groups along with a few other changes has been regarded as the hall-mark of the MIA. stage, and this is found convenient in readily distinguishing the MIA. dialects from the OIA. But this cannot be taken in a very strict sense. The process of assimilation has an earlier beginning which goes beyond the limits of the MIA. and reaches back into the OIA. On the other hand this process has not attained its completion even by the end of the MIA. period. Even if we try to define more accurately the distinguishing feature of the MIA. to consist in a complete more than a partial assimilation, the historical limits of even this restricted phenomenon do not coincide with the accepted limits of the MIA. dialects.
Probably the nearest approach to such a demarcating feature can be found in the process of assimilation of two plosives, which is almost universally operative in all the stages of the MIA. Even then the beginnings of this change are to be found in the OIA. period. Classical Sanskrit itself shows various traces of this type of assimilation. When two stops come into

Page #27

18
Amrita
contact, we find them assimilated as regards their voiced or unvoiced nature. In other words, there occurs the assimilation of voice to breath or of breath to voice, though other phonetic features of the two consonants may remain
nchanged. Thus ad- 'to eat' gives attur (ad-+tum) and atsi from ad-+si; vid'to know forms vettha with the inflection -tha. In all such cases the voice is assimilated to the following breath. On the other hand we have cases where the breath is assimilated to the following voice, as in sak- 'to be able' sagdhi. In both these types of assimilation, the important fact to be noted is the greater value attached to the second consonant, which prevails over the first. Considering the fact that in conjuncts of this type, the first consonant was merely implosive while the explosion belonged only to the second, the change is quite natural. This physiological fact is the guiding principle throughout the history of this assimilation and is fully operative in the MIA. stage.
There is, however, one apparent case, in which the tendency of assimilation appears to be a contrary one. Thus labh - 'to get' forms its past participle labdha with the termination -ta. Here the voice of the first consonant has assimilated the following unvoiced -ta. But the reason for this change is somewhat different. When the consonant-group was formed, the voiced aspirate h of the first member, following a general tendency of the Indo-Aryan languages, went to the end of the group. Thus the breath consonant was placed between two voiced sounds. This naturally led to the voicing of the middle consonant and the result was a group fully voiced from beginning to end. Thus the change can be formulated as : labh-+-ta = lab--ta+h = labdha.
That this assimilation of two stops was the first to be effected is seen from the Greek transcription of the name Candragupta as Sandrakottos,' where, while the group of the plosive and the liquid is kept, the conjunct of the two plosives is already assimilated. From the assimilation of voice and breath, the change further extended to the place of articulation of the consonants, at first under favourable circumstances. Thus Sk. ītte comes from id-+te. Here the particularly strong articulation of the cerebral has assimilated the dental of the inflection.
The nature of this assimilation of two plosives in Sanskrit, therefore, reveals the following facts. Firstly, the assimilation is operative between the rootending and the termination, while it has no scope in the body of the word. Secondly, no unassimilated groups of two plosives can stand at the beginning of a word, and naturally a geminated plosive is also impossible in such a position.
The MIA. stage differs from the OIA. in making this assimilation complete and not partial as in Sanskrit and carrying it over with equal

Page #28

Groups of Two Mutes in Middle Indo-Aryan
consistency to the middle of the word and not restrict it to the elementum formans. The exact and clear-cut pronunciation of Sanskrit conjuncts in all places is at the basis of all these changes. The MIA. has also preserved the importance of the second member of the group which is already visible in OIA. with the result that assimilation is everywhere regressive in case of two plosives. In spite of this wide-spread and nearly regular nature of this change, we do find a few cases where even the MIA. appears to keep groups of two plosives. That these groups are not simply inherited from the OIA. but are a new development will be evident from the cases themselves. Their origin is to be sought in some obscure tendency in the OIA.
Such groups of two plosives are most frequent in the Girnar version of Asokan inscriptions, while rare in other places. In Girnar we find the termination of the gerund -tvā of Sanskrit represented by the group -tpā : alocetpā (kālocayitvā), ārabhitpā (*ārabhitva), dasayitpā (*darśayitvā), paricajitpa, ("parityajitvā). The same group-tva- in a different context gives us -tpa- : catparo (catvārah). Here we find the tendency of assimilating the second member to the first both as regards breath and plosion; so that the semi-vowel -v- becomes -b- and then p-. A similar change at the initial position is found in dbādasa (dvādaśa), where the voice of the first consonant has preserved the voice of the second. This group of two plosives at the beginning of a word, which will not be tolerated even in the OIA. shows that either we have here an unusual tendency foreign to IA. languages, or that we have to do with some graphic peculiarity, not exactly reproducing the intended sound. Besides the semi-vowel, the nasal-m- has also been assimilated to the preceding stop, which has given rise to other groups of plosives. Asoka, Girnar : ātpa-(ātmā), ātpapāsamdapūjā; tadātpano (tadātmanah); Sd. mahātpā (mahātmā) but br. mahatpā. This change is presupposed by the other versions of Asokan inscriptions also, as they carry the change a step further and assimilate this group of two plosives. Thus S. badasa (dvādaśa). In the Niya Prākrit we have the form badaśa.
A number of considerations make it difficult to believe that here is a mere graphic peculiarity. Thus there are the other versions of Asokan inscriptions which show the further stage of the assimilation of the group - d+b-; the Māhārāstrī shows the form appa- where -pp- is the regular assimilation of -tpa- from tma-; Pāli has bidala for dvidala; the classical Präkrits show baravai for dväravati, bära (dvāra), (bedve); The Greek transcription of Dvārakā is Barake. In all these cases the first consonant has exerted a strong influence over the second in assimilating it as regards

Page #29

20
Amrita
contact, voice and denasalisation. This tendency is evidently contrary to the normal procedure of the Präkrits.
In order to explain the nature of this group of two plosives resulting from a peculiar process of assimilation, we must examine more carefully the phonetic structure of the Sanskrit groups from which they have developed. One thing which requires particular notice in this connection is the fact that in the original Sanskrit groups the second member is either a semi-vowel -vor a nasal stop-m-, both of which lack the complete closure of the regular plosive which forms the first member. Phonetically, it is possible to pronounce a consonant with a semi-vowel like -y- or -V- without making a conjunct capable of making position. In other words, instead of forming ordinary conjuncts of the regular type, we may pronounce them as simple consonants which are either labialised or palatalised. Normally, however, in the medial position, the preceding consonant was doubled in Sanskrit and catvārah was pronounced as cattvārah, with the first syllable long by position. This usual practice where the syllabic division was cat / tvārah has resulted into the normal MIA. form cattāro. With the preservation of the syllabic structure of the word, -tva- became -tta- But in some locality or in some dialects, there appears to have existed another kind of syllabic division and a different pronunciation of such groups, as can be seen from the statement of the Taittirīya Prātiśākhya XXI. 7 according to which the first consonant of such groups went over to the following syllable, giving rise to a division like ga tvā. Dr. Varma rightly objects to this division on the ground that such a syllabication cannot be valid if the first syllable is heavy by position. But it is equally likely that what the statement means is the fact that 'this syllabic division also implied a different syllabic value of the word and that -tva was not a conjunct making position. It was a simple labialised plosive. That this is not a mere theoretical supposition but an actual fact is seen from the important feature revealed by the illustrations of the Aśokan inscriptions (ātpa-mahātpā, etc.) where the first syllable preceding the seeming conjunct, preserves its normal length, which is abnormal in MIA. phonology. In these cases, we must admit an open syllable in the original words from which they have developed and thus admit a syllabic division like ā/tman exactly similar to ga/tva. With such a division current in some locality or dialect, the abnormal development of these groups can be easily explained.
These considerations lead us to the supposition that a dialectal difference existed in the pronunciation of the Sanskrit conjuncts giving rise to two divergent developments in the MIA. one of which is revealed by the

Page #30

Groups of Two Mutes in Middle Indo-Aryan
Girnar inscriptions of Asoka, while the other is shown by his other versions and the Classical Prākrits. The treatment of the Girnar version is further continued in the Apabhramśa dialect where we find the gerundial termination -ppi (-tvi) -ppinu (-tvīnam) and its further phonetic development into -vi, -evi. It is equally clear that this distinction of syllabication cannot exist with the same precision when the group occurs initially and the usage must have been unsettled. Exactly in such cases as dvādaśa = bāras, dväravati = bāravai, we find this typical change spread over most of the Prākrits, the intermediate step of which development is seen in the Aśokan form dbādasa. Similarly Apabhramśa shows alternative forms like tai, pai for tvayā. A very good parallel to all these changes is furnished by the changes in the formative period of Latin. Here initial dv- has become b- in bellum, bonus, bis (earlier dvis) and bimus (from *dvi-him-us). The earlier forms like dvellum, dvonorum are preserved in inscriptions. Further Latin also shows the same alternative change in words like dīmus, dēs in dialectal forms and sometimes in the classical form dirus (for *dvei-ros). And we know that Latin also kept a syllabic division like pa/tris.
We are thus led to the following conclusion. The conjunct of the type, mute followed by -v- or -m- in the OIA. showed a different development according as it stood initially or medially. In the initial position, no dialectal difference was observable and the Prakrit dialects show both treatments in different words or even in the same word side by side. In the middle of a word, however, two distinct and mutually exclusive lines of development are observable. The more frequent line is represented by the Classical Prākrits where the plosive assimilated the following semi-vowel or the nasal. The other, though less extensive, is found in the Girnar inscriptions of Asoka and the Apabhramsa morphology. These two developments are to be traced to a difference of pronunciation and the consequent mode of syllabication of the OIA. groups from which they develop.
BIBLIOGRAPHICAL
For earlier discussion of the problem cf. Pischel : Grammatic der PrākritSprachen (1900) Sect. 277, 300; Geiger : Pāli Literatur und Sprache (1916) Sect. 53; Bloch : La Formation de la Langue Marathe (1920) Sect. 129, 130; Varma : Critical Studies in the Phonetic Observations of Indian Grammarians (1929) pp. 75-77; Jain : A Phonology of Punjābi (1934) Sect. 150 (iv); Bloch; L'Indo-Aryen (1934) pp. 84-85.
Groups of Two Mutes in Middle Indo-Aryan,
JBU.XVI.1945
13

Page #31

Repetition in Prākrit Syntax
The device of repeating a word or a grammatical form in close succession to express an idea of greater emotional intensity or one of frequent occurrence is used in nearly all the languages and is a survival of the early devices used by the speaker. Whenever the emotional colouring of the individual's experience is strong enough, it finds expression in language by this device of repeating the word expressing the idea. Even though it is, in this manner, primarily a means of giving expression to one's emotions, in course of time it came to acquire some intellectual meaning, as well such as the ideas of totality and continuity.
The repetition of forms pertains to nearly all the grammatical categories, to words of all kinds and in few cases even to parts of words. Thus we find substantives, adjectives, verbs, adverbs, pronouns and others repeated to express different ideas connected with them
Historical grammar regards INTERJECTIONS and forms of IMPERATIVE to be perhaps the oldest elements of the language. Naturally this device of repetition which is intimately connected with the early beginnings of a language finds greater scope with them. Nearly all the interjections are used with a repetition, while forms of the imperative of the second and the third person usually express some kind of command and concession, ideas which are often required to be stated with emphasis and this is accomplished by repeating the words. AMg; hantā hantā bahave kandimsu / Ay. I. 9. 1. 5. “Look hear, Look hear, so cried many people.” JM : avvo avvo tti vāharanti hasanti sisā / Vas. 127. 16. "They cry, Alas, Alas, and the pupils laugh." eha eha siggham dacchaha acchariyaṁ / Vas. 134. 28. "Come, do come quickly, see the wonder.” Sometimes when the expression is not very strong, the forms of the imperative are separated from each other, one of it standing at the beginning as a place of emphasis, and the other at the end of the sentence, which is the normal place of the verb. AMg. : uppayāhi khalu bho paumavarapondariya

Page #32

Repetition in Prākrit Syntax
uppayāhi / Sut. II. 1. 6. "Fly up, Oh you beautiful lotus, fly up." gaccha nam * tumam devānuppiyā sālādavim corapalliṁ vilumpāhi 2 abhaggasenam
corasenāvaim jīvaggāham ginhāhi 2 mamaṁ uvanehi / Vip. 75. "Go, you beloved of the gods, to the camp of the thieves called Sālādavi and plunder it. Take hold of the chief of the thieves Abhaggasena alive and bring him to me." JM : tato tie bhaniyam ehi ehi tti / Vas. 53. 4. “Then she said, come, come." tā pahiya turiyaturiyam vaha vaha ullavai kalakanthi / Vaj. 651. "Therefore, O traveller, proceed quickly, so says the cuckoo.” In all such cases we find the original meaning of this device, to give greater emphasis to the idea expressed, in its pure form and obviously there is no scope for any change in the meaning itself.
The remaining finite forms of the verb are not often repeated. The VERBAL, DERIVATIVES, however, are repeated in order to give expression to different meanings, the idea of frequency being prominent. When the form refers to the same subject it serves to point out that the action is done by him repeatedly. P : puttho puttho cāham tesař vyākareyyam / MN. I. 13. "Being repeatedly asked I will explain to them." AMg.: kasappahārehim tālemānā 2 kalunaṁ kāganimaṁsāiỉ khāventi / Vip. 63. "Beating him repeatedly, they make him eat his own flesh in a pitiable manner.” chippatūrenam vajjamānenaṁ 2 / Vip. 68. "While the trumpets were blown." evam dubbuddhi kiccānam vutto vutto pakuvvai / Das. IX. 2. 19. "Thus a wicked pupil does when repeatedly asked by the teacher.” JM : te ya bhayavanto
panamanto panamanto aikkamai / Vas/ 74. 14. “Saluting those venerable .. sages repeatedly he goes forward."
The repetition of these forms also points out the fact that the action is not repeated but performed continuously. It must, however, be admitted that the distinction between the continuous and the repeated action is mainly due to the primary meaning of the verb and is not the direct result of the device of repeating. When the meaning of the verb does not admit of the idea of repeated performance it naturally develops the idea of continuity. AMg : pagadhijjamāne 2 uvāgae / Vip. 11. "He approached dragging it continuously.” tae nam sā miyadevi .... padijāgaramāni 2 viharai / Vip. "Then that queen Miyā continued to keep awake." taṁ katthasagadiyam aņukaddhamani 2 uvägacchai / Vip. 16. “She approaches dragging the wooden cart after her.” Sometimes the sense of totality is also expressed by repeating the verbal derivatives. P.:. laddham laddham vināseti / Sn. 106. He destroys all that he gets."
Slightly different meanings are found expressed by repetition of the

Page #33

Amrita
verbal forms in the following illustrations. P: so tathāgate cakkhupatham vijahante vijahante yeva pasannamano kālaṁ katvā suttapabuddho viya devaloke......nibbatti / Dh. Co. I. 23. "While the Tathāgata was passing out of the sight, he, having died with a delighted mind, was born in the world of the gods, as if awakened after sleep." imassa dassanatthāya agatāgatā anto gehe săpateyyam passissanti / Dh. Co. I. 21. "All those who come to see him will see the wealth inside the house." A gerund when repeated may show an habitual action. P: so tato nikkhamitvā araññe gāyitvā gāyitvā anto dārūni uddharantiyā itthiyā gītasaddam sutvā sare nimittam ganhi / Dh. Co. I. 12. “Coming out from there he took as the object of his thought the voice of a woman having heard her sound of the singing, the woman who was collecting wood in the forest by singing." JM : so tatha jimium jimium ahijjai / U. Tīkā. fol. 124. "He studies by taking his meal there."
In the Ardha-Māgadhī prose, however, we often find the number 2 used as a sign of repetition usually after the verbal forms and sometimes after other words. tae nam sā miyadevi miyāputtassa dāragassa anumaggajāyae catāri putte savvālankāravibhūsie karei 2 bhagavao goyamassa pāesu pādei 2 evam vayāsī / Vip. 14. “Then that queen Miyā adorned the four sons born after the child Miyāputta with all ornaments, placed them at the feet of the venerable Goyama and spoke thus." miyam devim āpucchai 2 miyāe devie gihão padinikkhamai 2 miyaggamam nayaram majjhammajjhenaṁ niggacchai 2 jeneva samane bhagavam mahāvīre teneva uvāgacchai 2 samaņam bhagavam mahāvīram tikkhutto āyāhinam payāhinam karei 2 vamdai namamsai 2 evam vayāsi / Vip. 19. “He takes leave of queen Miyā, comes out of her house, comes out of the town of Miyaggāma passing through the very centre, comes to the place where the venerable ascetic Mahāvīra was, salutes and bows him and says as follows." In the innumerable cases of this type the use of the figure 2 is not exactly the same as in the cases cited before. As shown by the meaning and the parallel passages the figure 2 does not here suggest the simple repetition of the verbal form after which it is placed but the gerund derived from the root after which it comes and points out the succession of acts thus enumerated. More accurately it should have been represented by the addition of the syllable ttä after the figure in the text. This itself is a peculiar Prākrit idiom which requires explanation.
More frequent is the repetition of the NOUNS in their inflected forms. Here also the locative forms are the most frequent. The use of the locative to express a particular point in place or in time is susceptible of repetition more than the meanings of other cases, and naturally repetition is favoured

Page #34

Repetition in Prākrit Syntax
in this case. Locative of time repeated : P: māse māse kusaggena bālo bhuñjetha bhojanam / Dh. 70. "The fool may take meals on the tip of a Kusa grass month after month.” māse mase sahassena yo yajetha satam sama / Dh. 106. "Who performs sacrifice for hundred years by spending a thousand month after month.” divase divase ti sattisatāni navanavā pateyyum kāyamhi / Th. 2. 473. “Even if three times seven hundred spears were to fall on the body anew on each day.” āhariṁsu dine dine / Mhvs. 5. 29. “They brought day by day.” adāpesi dine dine/ Mhvs. 5. 84. "He caused to be given day by day.” AMg. : māse mäse u jo bālo kusaggeņa u bhuñjae / U. 9. 44. “The ignorant person who eats with the blade of the grass month after month." JM : tinneva ya kodio addham ca dine dine ya rayaņānam, pādei dhanayajakkho / Pau. 21. 16. “The demi-god Dhanaya showered three and half crores of jewels every day.” The locative of place repeated :P : yojane yojane dentu mahādānam mahitale / Mhvs. 5. 179. “Let them give the great gift on the earth at every yojana.” kule kule appatibaddhacitto / Sn. 65. “With his mind unattached to all the families.” AMg. : miyaggāme nayare gehe gehe kālunavadiyae vittim kappemane viharai / Vip. 9. "In the town of Miyaggāma he lived maintaining himself by pitiously begging from house to house." caccare caccare khandapadahenaṁ ugghosijjamānam / Vip. 37. “At every square he was being proclaimed with the beating of a small drum." pacchā jāyā gamissāmo bhikkhamänä kule kule / U. 14. 26. "Later on, Osons, we will wander by begging from family to family.” pae pae visīyanto samkappassa vasaṁ gao / Das. II. 1. “Disheartened at every step and under the influence of his desires.” JM : thāne thāne jasaṁ lahai / Vaj. 6. 82. "He gets fame in many places.”
Other case forms of nouns are also found repeated but not to the same extent. AMg. appege paliyantesim coro coro tti suvvayam, bandhanti bhikkhuyam bālā / Sut. 1. 3. 1. 15. “Some ignorant persons bind a monk of good vows on the border land by calling him a thief.” chanaṁ chanam parinnāya logasannam ca savvaso / Ay. 1. 2. 6. 5. “At every moment knowing the view of the people from all sides.” Sometimes the two words are joined together in a compound so that the first word does not receive its proper grammatical inflection. P : anuvicarantānam maggāmagge kathā udapādi / DN. I. 235. “While wandering there arose the conversation on the way.” AMg.: urālāi māņussagāi bhogabhogāi bhuñjamāṇe vihara / Vip. 50. “He lives enjoying the great human pleasures of different types.” pakkamanti disodisim / U. 27. 14. "They go in different directions." JM : donni vi kheyālasangamangai/ Pau. 16. 80. “Both of them with all their limbs full of exhaustion."

Page #35

Amrita
suhamsuhenam kālam gamei / Vas. 75. 28. “He spends the time in happiness."
In one particular instance we find that the same noun is repeated but in two different grammatical forms both of which however have the same meaning. JM : nayaram ciya porānam rāyapuram nāma nāmenam / Pau. 2. 8. "The old town Rāyapura by name." saccamai nama nāmenam / Pau. 19. 32. "Saccamai by name." janayassa mahādevī asi videhi tti nama nāmenań / Pau. 26. 2. “The chief queen of king Janaya was Videhi by name." arihasano nama nāmenam / Pau. 31. 23. “ Arihasana by name." sunando nāma nāmao así / Vas. 74. 24. "Sunanda by name." mandaro nama nāmenam / Vas. 75. 22. "by name Mandara." This usage can also be met with in the Rāmāyana. śikharam nāma nāmataḥ / I. 27. 10; nandanam nāma nāmataḥ / I. 27. 13.; kesini nāma nāmatah / I. 38. 3. Metrical considerations and the desire to fill in the line may have been responsible for this repetition but its presence in Vasudevahindi would suggest that the close joining of the word nāma with the proper name as in jambūnāmo often occurring therein must have also helped to add the adverbial form namenań or nāmao in addition to the word nāma.
In this connection a peculiar idiom must be noted. Along with the usual phrase for expressing a continuous action from place to place, in which the first word is put in the ablative and the second in the accusative as the first is regarded as the starting point of the action while the second alone is thought to be the goal of the action, we often get phrases in which the word in the accusative is merely repeated. P. kālam kālam bhavā bhavam sakkāyasmiṁ purakkhatā / Th. 2. 199. "entangled in the view of believing in the permanent thing from time to time and from birth to birth." rattham rattham vicarissam sāvake vinayam puthu / Sn. 444. "I will wander from kingdom to kingdom teaching the followers." Such an idiom must be at the basis of the word gānaṁganie in U. 17. 17. where it refers to a pupil who is in the habit of wandering from one gana to another. Speaking grammatically in these cases both the words are in the place of the object and the expression should therefore mean going to one place and then to another. In the ultimate sense of both the idioms there is very little difference but while the first implies some kind of sequence the second is free from any such connotation. It should be further noted that the translation of a word when repeated as "every” is not very accurate. Often it leaves out the idea of succession implied in the original and adds the sense of totality which may not be always present in the original. Of the same import is the phrase in which the instrumental is used instead of the ablative of the first word which is followed by

Page #36

Repetition in Prakrit Syntax
27
the same word in the accusative : P: tena hi bho imam purisam dalhāya rajjuvā pacchābāham gālhabandhanaṁ bandhitvā khuramundam karitvā kharassarena panavena rathiyāya rathiyam singhātakena singhātakaṁ parinetvā..../ DN. 23. "then having bound strongly this person with strong ropes with his hands behind having shaved him, and carrying him from road to road and from square to square with the beating of a drum of a shrill sound....” JM : eyassa pavarakitti gehaṁ gehena bhamai jiyaloe / Pau. 15. 66. "His fame wanders from house to house in this living world.” A similar illustration can be found in the Rāmāyana in 1. 1. 30. te vanena vanaṁ gatvā nadīstīrtvā bahudakah. This idiom is simply a recasting of the first by changing the ablative case into the instrumental as in course of time the two cases were often confused and mixed together. Some change in the meaning of this idiom can be seen in such an illustration as anubandhi padāpadań / Sn. 446. "he followed him step by step” where the compound expression is clearly formed on the analogy of such expressions as : P: te mayam vicarissama gāmā gāmam nagā nagaṁ / Sn. 180. "We will wander from village to village and from mountain to mountain." vinīpātam samāpanno gabbhā gabbhaṁ tamā tamam / Sn. 278., "He fell down from one birth to another and from darkness to darkness."
A peculiar type of repetition of nouns is to be found in cases where we find the same word twice used but in two different syntactical relations, usually once as the subject and again as the predicate of the same sentence. This was already noted by the Sanskrit rhetoricians who gave it the name chekānuprāsa. As : gadhā să hoti piti piti dhammavijayasi / K. XIV. 13. “That love is the deep love which pertains to the conquest of religion.” JM : taha ri hu hamso haṁso kāo kāo cciya varão / Vaj. 358.“ Even then the swan remains a swan and the crow a crow." M : raikiranānuggahiyāi honti kamalai kamalāi / Vis. "The lotuses become lotuses when touched by the rays of the sun."
ADJECTIVES are often repeated to point out the intensity of the quality expressed by them. P : khippam giram eraya vagguvaggum / Th. 1. 1270. "Quickly put forth words which are very sweet.” paramam paramam ti yodha ñatvā akkhāti vibhajati idheva dhammam / Sn. 87. "Whosoever, having known it to be the highest religion preaches and explains it.” AMg. : mahayā mahayā saddenam ugghosemānā / Vip. 32. “Proclaiming with a very loud voice." mahayā 2 saddenaṁ... ārasie/vip.43."He cried with a very loud voice.” Ghodayapuccham va tassa maṁsüī kavilakavilāi / Upa. 94. "His moustaches were very tawny like the tail of a horse." In the frequent phrase

Page #37

28
Amrita
mahaimahāliyāe parisāe of the AMg. canon we find the adjective repeated with the first word retaining its locative form as mahai corresponding to Sanskrit mahati which however, loses its inflectional value as can be seen from its being used with a feminine noun.
Adjectives can also be repeated with a distributive sense. P: tena kho pana samayena sambahulā abhiññātā abhiñāttā brāhmaṇamahāsālā manasākate pativasanti / DN. I. 235. "At that time there lived in the village Manasākata many great Brhamins all of whom were well-versed.” paccati munino bhattam thokam thokam kule kule / Th. 1. 248. "The food for the monks is cooked in different families a little in each." AMg.: saehinto saehinto gihehinto padinikkhamanti / Vip. 23. "They come out of their respective houses."
Slightly different meanings like that of totality or diversity can also be i found expressed by the repetition of adjectives. AMg. : bhaddagam bhaddagań bhoccā vivannam virasam āhare / Das. V. 2.33. "Having eaten all that is good he may bring the food which is colourless and tasteless." P. evarūpaṁ vā evarūpas vā pāpam kammam akaramhā ti na jänātha / MN. I. 14. “You do not know that you have done such and such an evil deed.” Sometimes this repetition of the adjectives is concealed on account of the false etymology. AMg. : thaviyaṁ sankamatthāe tam ca hoi calācalaṁ/ Das. V. 2. 33. “It may be placed for crossing, and may be shaking” where the word calācala originally a repetition of the adjective cala was thought as a compound of cala and acala.
The repetition of the ADVERBS is fairly frequent. They also show a variety of meanings expressed by this device. As : esa cu kho mama anusathiya dhammāpekkha dhammakāmathā cā suve suve vadhitā / T. I. 6. "Now by my preaching of the law love for the law and the need of the law have increased in the respective fields.” P. yañce viññā pasaṁsanti anuvīcca suve suve Dh. 229. “That the wise praise him discriminately day by day.” vissavanto tato tato / Sn. 205. "Flowing from those different places.” iccetam attham bhagavā punappunaṁ akkhāsi / Sn. 251. “Thus the Lord preached this meaning again and again." tena kho pana samayena bhesiko nahāpito bhagavantam pitthito pitthito anubaddho hoti / DN. I. 226. “At that time the barber Bhesika was following the Lord from behind." aniccāni gahakāni tattha tattha punappunam / Th. 1. 17. "The houses are transitory and there again and again.” bhikkhave tadā mama puttena katakammaṁ pacchato pacchato anubandhi / Dh. Co. I. 17. “O Monks, at that time the acts done by my son followed him from behind." visum visum pure rajjam kamato anusāsisum /

Page #38

Repetition in Prākrit Syntax
29
Mhvs. 2.11. "They ruled in due order individually.” tahim tahiṁ cetiyāni akāresi / Mhys. 5. 175. "he erected the Cetiyas in different places.” AMg.: Mahabbalassa ranno abhikkhanam 2 kappāyaṁ genhai / Vip. 60. “He often took taxes from king Mahabbala.” payāhiņaṁ karento puna puno vandai sakko / U. 9. 59. "Sakka saluted him often and often while circumambulating him." neyāuyaṁ suyakkhāyam uvāyāya samīhie, bhujjo bhujjo duhāvāsas asuhattaṁ tahā tahā./ Sut. 1. 8. 11. "Following the right doctrines he exerts himself, as one becomes more and more the receptacle of misery so his bad thoughts increase.” pāvāiņo pudho pudho kittayantā sayam sayaṁ ditthi karenti pāu / Sut. 2. 6. 11. "The upholders of the wrong faith explaining their different views maintain their own opinions." anto anto puidehantarāni pāsāi / Ay. 1.2.5.5. “Inside and further inside he sees the rotten parts of the body." saniyam saniyam paccosakkai / Upa. 101. "He slowly recedes back.” JM : so pāvai abhiseyam uppajjai jattha jattha naro / Pau. 32. 78. “The man wherever he is born obtains coronation.” bhattumaranadukkhiyā mamam ca soyamāņi māyā me sukkakothararukkho iva vanadavena soyagginā anto anto dajjhai / Vas. 36.8. "My mother troubled by the death of her husband and lamenting for me was burning inside and inside with the fire of grief like a tree with a hollow by the fire of forest conflagration." ehi kunasu paikkhinam devaulassa mandam mandam parīti / Vas. 81.5. “Come and go round the temple slowly." suyanasamāgama vaggi niccam niccaṁ suhāvei / Vaj. 655. "Like the meeting with a good man fire gives pleasure constantly.” Ap. puņu puņu panavivi pañcaguru 7 P. ar-pr. 1.11.1. "Having saluted the five teachers again and again.”
The NUMERALS are repeated in a distributive sense. As. : pañcasu pañcasu vasesu anusayānam nikhamantu / D. III. 9. "In each five years they will go on a tour.” etāye ca athāye hakam dhaṁmate pañcasu pañcasu vasesu nikhamayisāmi / D. 21. I. “For this purpose every five years I will send out.” P: duve duve putte janayi kāle sā/ Mhvs. 6.37. “She gave birth to sons two at a time.” JM. : atthattha nādayāi dāre dāre ya naccanti / Pau. 2. 52. "At every door were dancing groups of eight shows.” causu vi disāsu majjhe havanti cattāri cattāri / Pau. 2.52. “In the middle of the four quarters there were four of them each." P: saggakāyam agamaṁ sakim sakiṁ / Th. 1. 259. “I went to the heavens once in each.”
The PRONOUNS are very often repeated to express the idea of totality. As. : taśi taśi pakalanaśi pujetaviya cu palapāśadā-tena tena akālana / K. XII. 32. "On occasions other religionists ought to be honoured in different ways.” athi ca hetha punaṁ puna lipite tasā tasā athasā madhuliyae yena jane tathā

Page #39

30
Amrita
patinajethe / X. XIY. 21. "Here it is written again and again on account of the sweetness of this and that subject so that men may follow them.” se tam apahatā tam tam dhammavadhi pāpovā / T. VI. 3. "Giving up this and that they should develop their religion." P: yam yam padesam bhajati tattha tattheva püjito / Dh. 303. “In whatever place he goes he is honoured there." Yena yena hi maññanti tato tam hoti aññathā / Sn. 588. “Whatever they think it becomes something otherwise.” te te āvikaromi / Sn. 84. "I will explain them all.” tam tam aham dhīra tatheva maññe / Sn. 349. “I also think the same O courageous man." vicariham tena tena lābhasakkāraussukā / Th. 2.92. "I wandered here and there desirous of profit and honour.” naggā pakinnakesi ham tena tena vicāriham / Th. 2. 133. “Naked and with the hair let loose I wandered here and there.” ubbāsiyati so so ca yam yam gamam upeti so / Mhvs. 6.22. “Every village in which he goes is deserted.” AMg. : jā jā vaccai, rayani na sā padiniyattai / U. 14.24. “Every night that passes does not return." jai tā kāhisi bhāvam jā jā dacchasi nārio / Das. 2. 9. “If you love every woman whom you chance to see.” tehi tehi uvāehim tam tam sampadivāyae / Das. IX. 2.20. “He should carry all that out with all means.” JM : te te thāvehi vase / Pau. 11. 102. “Put all of them under control." jā jā dālā lambai ........sā sā tadatti tuttai / Vaj. 124. “Every branch which he holds breaks suddenly." je je kulammi jāyā te te gaya- kumbhaniddalaņā / Vaj. 201. “All those who are born in the family are capable of breaking the temples of the elephants."
The second person pronoun is repeated. AMg. : tumam tumam ti amanunnam savvaso tam na vattae / Sut. 1.9.27. "To call thou is not elegant and so one should never speak it.” The reflexive pronoun is repeated with a distributive meaning. P: sakaṁ sakam ditthi paribbasānā viggahya, nānā kusalā vadanti / Sn. 878. “Referring to their own views the clever ones explain differently." sakam sakam ditthimakaṁsu saccam / Sn. 882. “They claim their own views to be true." AMg : evam annāniyā nāņam vayantā vi sayamsayam/ Sut. 1.1.2.16. “Thus the ignorant ones claiming that knowledge is their own." sayam sayam pasaṁsantā garahantā paraṁ vayam / Sut. 1.1.2.23. “Praising one's own views and blaming those of others.” sae sae uvatthāne siddhimeva na annaha / Sut. 1.1.3.14. “Liberation is possible in one's own view and not otherwise.” The interrogative pronoun repeated : JM : kim kim ti ullavantā uppaiyā nahayalam turiyā / Pau. 9. 73. “They quickly flew in the sky crying 'what is it, what is it?" ke ke ime pautthā mottūņa gharesu gharanio / Vaj. 650. "Who are those who have started leaving behind their wives in the houses ?” kim pi kim pi cintanto / Vaj. 23. “Thinking something." ..
PRONOMINAL ADJECTIVES repeated :P : ettakam vā dukkham

Page #40

Repetition in Prākrit Syntax
31
nijjhinnam ettakam vā dukkham nijjetabbam / MN. 1.14. “This much misery is conquered and this much is yet to be conquered." Paramam paramam ti yodha ñatvā / Sn. 87. “Who knows it to be the highest.” AMg. : jakkhā uttarauttară / U. 3.14. "The demi-gods, higher and higher.” samsāramāvanna param param te veyanti bandhanti ya dunniyāni / Sut. 1.7.4. "Coming to life they bind more and more and suffer miseries." samvaccharenāvi ya egamegam banena mareu mahāguyam tu / Sut. 2.6.52. "In a year having killed only one elephant by the arrow.” sāyāgāravie ege ege sucirakohane / U. 27.9. "Some hanker after pleasures and others entertain anger for a long time." JM : mā punaravi ahiyayaraṁ pāvihaha paramparam dukkham / Pau. 26. 74. "Do not suffer more and more misery again and again to a greater extent." patteyam patteyam vakkhāram kittaissāmi / Pau. 2. 54. "I will explain the regions individually."
PARTICLES are repeated : P: evam evam vicarantam pāpima upagacchasi / Th. 1. 1213. "O wicked one, you follow me while wandering in this manner." AMG : jāvam jāvam ca nam abhikkameitāvam tavam ca nam mahante udáe....seyamsi nisinne padhame purisajāe / Sut. 2.1.2. “While he proceeded further the first man sank deep in the great water and mud.” JM : eso suyanasahāvo namo namo tāna purisanam/ Vaj. 37. “This is the nature of good men a repeated salutation to these men.”
PARTS OF WORDS are repeated to give an intensive meaning to the expression. The first syllable of a verb is often repeated to give emphasis to the idea, and this must have been the origin of the reduplication. JM : e-ehi savadahutto majjha tumam mā cirāvehi / Pau. 8. 173. "Come in front of me, do not tarry." āyārai mattagayaṁ e-ehi mahaṁ savadahutto / Pau. 8. 218. "He calls the intoxicated elephant to come to face him." Ap : abbhatthio si de-dehi tema / Mahp. 1.6.14. “You are requested to give me.” Here also belong the innumerable cases of immitative words like M : jagajagentasche / Pau. 14.134. gumugumugumunta / Pay. 2.40. where the first syllable is repeated and the grammatical inflections are added only to the last.
Another type of repetition which survives in the modern languages is the one in which the PRESENT PARTICIPLE is repeated. Ap. rangantena ramantaramantē manthau dhariu bhamantu anante / Hp. 85. 6. 2. "Taking delight and playfully, Ananta took the churning rod." Bhūvaibhūmi kamantakamantaha / Hp. 87. 6. 9. "Going over the land of the king."
Finally we find WHOLE SENTENCES or long phrases are repeated to give emphasis. P : appeva nāma siyā bhesike, appeva nāma siya bhesike ti /

Page #41

32
Amrita
DN / 1. 226. “It may be so Bhesike it may be so." kaham ekaputtaka kaham ekaputtkā ti/ Dh. Co. I. 23. "How is it O my only son, how is it ?" In the following case the idea is expressed with greater force by repeating the sentence and changing the order of words there. JM : jam ajja duhiyāe giyam tam vāiyam padujainā jam vāiyaṁ padujaiņā taṁ gaiyam ajja duhiyāe / Vas 132. 8. “What was sung by the daughter to-day the same was sounded by the clever man, and what was sounded by him the same was sung by the girl."
ABBREVIATIONS
AMg. Ardha-Māgadhi Ap. Apabhramsa Aś. Aśoka's Inscriptions Ay. Ācārānga D. Dhauli Rock Edicts
Das. Daśavaikālika
Dh. Dhammapada Dh. Co. Dhammapada Commentary
MN. Majjhima Nikāya P. Pāli Pau. Paumacariya P-pr. Paramātmaprakāśa Sn. Suttanipāta Sut. Sūtrakrtānga T. Delhi-Topra Pillar edicts. Th. 1. Thera Gāthā Th. 2. Therī Gāthā U. Uttarādhyayana Upa. Upāsakadaśah Vaj. Vajjālagga Vas. Vasudevahindi Vip. Vipākaśruta
DN. Dĩgha Nikāya
Hp. Harivamsa of Puspadanta
JM. Jain Māhārāstrī
K. Kalsi Rock Edicts
M. Māhārāstrī Mahp. Mahāpurāņa of Puspadanta
Mhvs. Mahāvamsa
Vis. Visamabānalīlā.
OOO
Repetition in Prākrit Syntax
New Indian Antiquary 2. 1. 1939.
25

Page #42

Concord in Prākrit Syntax
Nearly all the inflectional languages give some amount of importance to the observance of concord, partly due to the law of euphony and partly due to the necessity of expressing some discursive relations which otherwise do not find adequate expression in the language. Even in Indo-European times there must have been a tendency to put words in formal agreement whenever these stood in some relation which had no special means of expression in the morphology of the language. This tendency must have given rise to the laws of congruence. Normally, it was found difficult, except by the use of concord, to express the relations that are found between a substantive and a predicate or a dependent adjective, or between a pronoun standing for a noun and its dependent adjective and sometimes even between two sentences dependent upon each other in some respects.
Concord relates mainly to such aspects as Gender, Number, Case and Person, regarding which the words must formally agree to form congruence between them. Usually the logical relations which subsist between things and their qualities, an action and its doer and other similar cases are clear enough to exclude any doubt about them. In such cases concord can be neglected or even dispensed with. But the necessity of congruence arose in the inflectional languages because they failed to give a crude and uninflected form of a word to be used and the speaker was forced to choose some form or other to express the idea. Thus if a predicate attribute was to be used it should be neutral according to the discursive relation but usually took the form of the subject because no such neutral form was available. Only the form of the Vocative of a noun can be said to approach the neutral form and even here the adjectives qualifying it assume a similar form.
As suggested by Brugmann, concord may have arisen out of the tendency to generalise such cases as words in apposition in the first instance where they were in the same case because both served the same function in

Page #43

34
Amrita
the sentence. Once this was felt as a formal agreement it may have been extended to other cases analogically. So it is natural that in a sentence like bhrātaram varunam avavítsva "Turn towards us Varuna, the brother” both the words 'brother and 'Varuna' should occupy the same position in the sentence as they are, in reality, a repeated reference to the same object from different aspects. They thus show a formal agreement. From such cases the agreeement was extended to cases of predicative adjective and thence to all adjectives. In this manner the aspects of Number and Gender, which are primarily applicable to substantives, were extended to cases of predicative adjective and thence to all adjectives which in their own nature cannot possess either of them. Other factors like the general tendency of assimilation must have contributed to the spread of congruence even though it is not always possible to point out its exact influence in individual cases.
The congruence with the verbal form has a different origin. According to Paul the agreement of the verb as regards Person and Number arose because of the fact that a pronoun corresponding to the subject was added to the verb to produce the inflected form which thus naturally agreed with the subject. The added pronoun sometimes served the function of the personal subject but often merely pointed out the relation with the subject through congruence.
But the ideal perfection of congruencial relation between the various parts of speech in any given language can scarcely be met with. Various disturbing factors help for its break-down in individual cases or even in whole group of cases. (1) One such factor is the impression produced by the actual fact as distinguished from its grammatical expression, when both of them do not fully agree. Thus the thing may be thought of collectively even when the grammatical expression retains the plural number, which thus breaks the formal congruence. So in AMg. abhū jiņā atthi jiņā aduvā vi bhavissai U. 2. 45. "There were Jinās, there are Jinās and there will be Jinās." Here the writer thought the subject collectively when he used the verbal forms. (2) Another such factor is the fixing up of a group of words which resist all further attempts of changing them. So, on account of the constant use of the form atthi at the beginning of a sentence we get expressions like. P. natthi khandādisă dukkha | Dh. 202. “There are no miseries like the Khandas". Here the form of the verb persists even when followed by a subject in the plural - number. (3) Sometimes an important word in the sentence disturbs the congruence. AMg. chajjīvaniyā nāmam ajjhayanam paveiya | Das.4: "The chapter called 'The Six Groups of Living Beings' is preached.” Here the

Page #44

Concord in Prākrit Syntax
predicate is attracted by the earlier word so as to make it agree with itself instead of the word ajjhayana with which it should formally agree. (4) Sometimes the speaker himself changes his mode of expression suddenly and the result is some kind of disagreement. AMg. siehi asie bhikkhū amokkhãe parivvaejjāsi | Sut. 1. 1. 4. 13. "A pious monk should live as a man free from fetters among those who are bound with fetters, till he reaches liberation." All such examples occur at the end of the various chapters and it appears that the speaker used the form of the second person when he came to the end of the lesson because he wanted to make the whole chapter an address to his audience. (5) Sometimes the actual situation is so difficult to deal with, that the speaker is at a loss to know what should be the method of agreement and naturally takes up a third course which he thinks best in the circumstances. Thus when the subject of a sentence is a group of two nouns of which one is Masculine and the other Feminine, the verb is often put in the Neuter as the speaker found it difficult to make it agree with any one word at the expense of the other. JM. tāhe rāyā sā ya jayahatthimmi ārūdhāi Erz. 34. 29. "Then the king and the queen mounted the victorious elephant." The more popular the language the more powerful are such agencies and the greater the effect on the concord of the language. A comparison between the rules of congruence in the Classical Sanskrit and the different Prākrits reveals the fact that these rules are often broken in the latter than in the former.
Tven in
Even in the Sanskrit language itself the Vedic speech on the one hand and the epic idiom on the other, are decidedly more flexible and admit of greater variations in the rules of concord than is the case with the Classical Sanskrit. In many such cases the Prākrits are found to agree with them more than the literary Sanskrit. Thus the Vedic language shows some cases where the Neuter plural subject is joined with a singular verb : dhrsnave dhiyate dhanā | RV. 1. 81. 3." To the bold man booties are offered.” This is a regular feature of Greek syntax. In Prākrits we find this usage extended to other genders as well. AMg. tattha je āriyā te evam vayāsi | Ay. 14. 2. "There those, who were the Aryans, spoke thus." soccā neyāuyam laddhum bahave paribhassai U. 3. 9. "Many, having heard the law and having practised it, again fall down from it." In all such cases the idea of collection is prominent and guides the usage.
I. THE AGREEMENT BETWEEN THE SUBJECT AND THE PREDICATE
The agreement found between the subject and the predicate of a given sentence pertains to the number and person when the predicate is in the form

Page #45

Amrita
of a finite verb and to number and gender when it is in the form of a verbal derivative. In the first case the verbal inflections are incapable of denoting gender while in the second the case terminations cannot express person.
The agreement in this part of concord is usually the same in Prākrits as in Sanskrit, the difference being mainly the greater freedom which Prākrits languages enjoy in not following the laws than is allowed in Sanskrit. It is, often, very difficult to assign precise reasons for this difference, but lack of greater refinement and proper care must have been one of them. But all cases cannot be explained on the mere supposition of mere lack of ability or carelessness on the part of Prākrit writers and at least in case of the vast number of illustrations in which the congruence of number is absent, we must suppose that the popular speech allowed greater freedom and latitude than was given to Sanskrit writers.
A. Agreement in Number. Normally the number of the subject and the predicate is the same. AMg. devā vi tam namamsanti | Das. 1. 1. “Even the gods adore him.” bhamaro āviyai rasam | Das. 1. 2. “The bee drinks the juice.” P. rāyadūtā vidhāvantu kahim bhikkhū gamissati | Sn. 411. “Let the royal servants hurry to know where the monk will go.” Because the dual forms are lacking in all the Prākrits, the verb is in plural with a dual subject. AMg. do vi gacchanti sôggaim Das. 5. 1. 100. “Both of them go to a good birth.” JM. varadhanunā bhaniyā bambhadatta-buddhila | Erz. 12. 15. "Varadhanu said to Bambhadatta and Buddhila." payattā kumāra-varadhanū| Erz. 12. 15. "The prince and Varadhanu started."
This agreement holds good with the grammatical gender and number even when the actual fact may not conform with it. When a person is spoken of in plural as a sign of respect, the predicate will be also in the plural. AMg. tao tubbhe (mama tão) eyamattham āikkhaha | Ny. 1. 14. “Then (O father) tell me this fact." JM. annesiyā savvao tubbhe na kahim ci ditthā | Erz. 15. 29. "You were searched everywhere but you were not found any where." Here tubbhe refers to one person only.
But a real difficulty arises when the subject is not a single word either in singular or in the plural but a group of words co-ordinate with each other. Here the predicate enjoys a good deal of freedom. Thus, if the words forming the subject are joined with the practical ca ‘and the predicate is naturally put in the plural. AMg. rāyāno rāyamccā ya....pucchanti | Das. 6. 2. "The kings and the ministers of the king ask.” koho ya mano ya aniggahiya | Das. 8. 40. "Anger and pride which are not controlled.” But sometimes the other

Page #46

Concord in Prakrit Syntax
alternative, namely that the predicate should agree with one of them, usually the nearest one, is also found. JM. pesio leho kosalliyāim ca | Erz. 63. 27. “The letter and greetings were sent." kulam silam jai jaso lajjā ya pariccattā | Erz. 24. 32. "Family, good conduct, birth, fame, and even shame was abandoned." kim vā karei mama virahe mahārāo amba ya Erz. 21. 10. "What is the king and my mother doing in my absence ?” In the first group of cases the speaker views the things at one stretch, while in the second the thought grows slowly, making additions to the original idea which is already expressed. The rule, that the predicate should agree with the nearest subject is however not strictly followed. JM. sampayam ca davvaharanam parimoso ya kemavi kayam Erz. 67. 19. "Just somebody took away the wealth and committed theft.”
If the nouns forming the subject are joined by the disjunctive particle vā ‘or', the predicate naturally remains in the singular. AMg. asanam pāņagam vāvi udagamsi hojja nikkhittam Das. 5. 1. 59. "Either food or drink may have been thrown in water." natthi punne va pāve vā| Sut. 1. 1. 1. 12. "There is neither merit nor demerit."
When the subject is made up of nouns of different numbers the verb agrees in number according to the followig possibilities. (1) When the subject consists of two singular nouns joined with ca ‘and' or any other particle like vi ‘also', the verb is found either in the singular or in plural with nearly equal frequency. (i) Cases with the verb in the singular : P. atha nam bhanati tāto ammā savvo ca me ñātiganavaggo | Th. 2. 424. “Then my father, mother and all my relatives say." abhijjhā ca vihesā ca cittam assūpahaññati | Th. 1. 795. “Greed and vexation trouble his mind” na me tappaccayā tattha pemaṁ doso ca hehiti | Th. 1. 719. “I will neither have affection nor hatred on account of it.” yathā jalo ca mugo attānam dassae tathā | Th. 1. 582. “Just as a lame and a blind man might show themselves. “yassa rāgo ca doso ca mano makkho ca pātito | Sn. 631. “Who has thrown off attachment, hatred, pride and rage." neva devo na gandhabbo na māro saha brahmunā jitaṁ apajitam kayiră | Dh. 115. “Neither a god nor a Gandharva nor Māra along with the Brahmā can make a conquered thing unconquered.” na taṁ mātā pitā kayirā| Dh. 44. "Neither the mother nor the father can do it." Ap. damsanu nānu anantasuhu samau na tuttai jāsu | P-pr. 2. 11. "Whose faith, knowledge and infinite bliss do not cease at once." atthị na punnu na pau jasu P-pr. 1. 21. "One who has neither merit nor demerit.” atthi na ubhau jaramaranu roya vi linga vi vanna | P-pr. 1. 69. “Who has neither birth nor old age nor death nor disease nor sex nor colour."

Page #47

Amrita
(ii) Cases with the verb in the plural : P. ubho pitä сa mātā ca buddhassa upanāmayum Th. 1. 244. “Both father and mother carried me to Buddhā." evam jarā ca maccă ca āyum pācenti pāņinam | Dh. 135." "Thus old age and death drive off the life of men." AMg. na tassa māyā ya piyā va bhāyā kālammi tamamsahārā bhavanti | U. 13. 22. "Neither his brother nor mother nor father take a share of him at the time of death.”
(2) When the subject consists of more than one word of different numbers the verb is usually in the plural and only occasionally in the singular. (i) Cases with the verb in the plural. As. hidā migaviyā annāni ca hedisāni abhilamāni husu | K. VIII. 22. "Here were to be found hunting and other similar pastimes." P. rājā ca aññe ca bahū manussā avītatanhã maranam upenti | Th. 1. 778. "The king and many other men die without being freed from desire.” bhātaro mahyaṁ mātā ca modanti kāmākāmino Th. 1. 242. "My brothers and mother take delight, being desirous of pleasures." tato ca devā pitaro indo asurarakkhasā adhammo iti pakkandum Sn. 310. Then the gods, men, Indra, the Asuras and demons cried 'o the irreligion.” AMg. sehanti ya nam mamāiņo māya piyā ya suyā ya bhāriyā | Sut. 1. 2. 1. 19. "His father and mother, children and wife, who claim him, admonish him." rāyā narasetthimāhanā ţhāņā te vi cayanti dukkhiya | Sut. 1. 2. 1. 5. "The king, men, merchants and Brahmāns, all of them, fall, from their places, becoming miserable." JM. so ya sattudamano rāyā kumārā ya appachandenam viņatthā | Vas. 64. 3. “King Sattudamana and the princes were destroyed by their own whim." AMg. pasubandhā savvaveyā ya jattham ca pāvakammunā na tam tāyanti dussilam | U. 25. 30. "The sacrifices of animals, all the Vedas and whatever is offered by the wicked, all these do not protect a mản of wicked conduct." palenti puttā ya pai ya majjham | U. 14. 36. "My husband and sons run off." bhajjā ya puttā vi ya nāyao ya dāyāram annam anusankamanti U. 13. 25. "Wife, sons and the relative approach another protector.” Ap. satthapurānai tavacaranu mukkhu vi karahi ki tāsu | P-pr. "The scriptures, legends, performance, of penance even liberation, what can they do to him?” hariharabambhu vi jinavara vi munivaravinda vi bhavva..mukkhu ji jhāyahi savva P-pr. 2. 11. "Hari, Hara Brahmā, the Jinās, the group of sages and good people, all meditate on liberation.”
(ii) Cases with the verb in the singular : P. na tassa puttā pasavö vā khettaṁ vatthus na vijjati | Sn. 858. "Neither sons nor cattle nor wealth nor field belong to him." AMg. jattha natthi jarā maccū vāhino veyanā tahā U. 23. 81. “Where there is neither old age nor death nor diseases nor pains."

Page #48

Concord in Prākrit Syntax
39
(3) When the subject consists of many words all of which are in the plural, the verb is naturally in the plural except in exceptional cases. (i) Cases with the verb in the plural : AMg. dārāni ya suyā ceva mittā ya taha bandhavā jīvantam anujīvanti | U. 18. 14. “Wives, sons, relatives and friends depend upon him as long as he lives.” (ii) Cases with a verb in the singular : AMg. daharā buddhā ya patthae | Sut. 1. 2. 1. 16. “Young and old may request."
Principal Exceptions to the Agreement in Number
The principal exceptions to the above rules can be grouped under various heads as follows. (1) Because the Prākrit grammar has suffered greatly in the richness of forms, many anomalies arise. Particularly in the past tense the distinction between the singular and plural is often neglected and the singular forms are used in place of the expected plural ones. AMg. parenam vinayenam te suminapādhage evam vayāsi | Ny. 1. 12. "With great modesty those astrologers said as follows.” But even when two distinct forms are preserved sometimes the plural form of the verb is used with a singular subject. AMg. pantam sejjam sevimsu | Ay. 1. 9. 3. 2. "He used a low bed." puttho vi nābhibhāşissu | Ay. 1. 9. 1. 7. "Even when asked, he did not answer." (2) The verb may remain in the singular even when the subject is in the plural when it is thought of collectively. P. puttā matthi dhanam matthi Dh. 62. “I have sons and wealth," te sabbe pi tuhyam abbhantare atthi | PR. 3. 25. "All of them are in your inside.” AMg. māhanā samanā ege savvam nānam sayam vae | Sut. 1. 1. 2. 14. “Some Brahmāns and monks may claim all knowledge as belonging to them." JM. jattha jara-marana-roga-dosā natthi | Vas. 4. 3. "Where there are not old age, death, diseases and faults." Aś, osadhāni atata nathi | D. 45. "Where there are no herbs.” asti pi tu ekacā samājā G. I. 2-3. "There are but a few assemblies. (3) Sometimes a singular subject of the first person is followed by a plural verb, which construction is frequent in some books. AMg. ahayam anucitthimo...näham cukkimo | Mah. Ni. "I will perform it, I will not commit mistakes," JM. taö ham tão lavāmo Vas. 290. 28. "Then I spoke to them." cintemi kammi paesāmmi va tāmahe Vas. 352. 22. I began to think, 'in what place do I find myself ?" n sobhai ihaṁ acchium avakkamāmo tti | Vas. 122. 5. “It is not proper for me to remain here, let me go." In some cases the speaker is making a statement for himself and for those who are accompanying him. Amg. avi yāvi cittā jānissāmo | Ry. 18. “Moreover, O, Citta, let us know.” (4) In some cases the subject which consists of two things which in Sanskrit would normally take the dual number, takes a singular verb and not the usual plural in Prākrit. AMg. do vi ee akvoiyā tivvam soyam nigacchai | Sut. 1. 1. 2. 18. “Both of

Page #49

40
Amrita
them, being ignorant, come to suffer excessive misery." do vi tattha nimantaye Das. 5. 1. 38. "Both of them may call him." (5) In few cases we must admit some amount of inadvertence on the part of the writers who were misled by some adjacent word in a different number. P. (setthi) te (putte) vayappatte ghara-bandhanena bandhimsu | Dh. Co. 1. 2. "The merchant bound those sons who had attained youth with the bond of family." JM. sa imie saha dhammatthaabhagga-pasaraṁ visayasuham anuhavissu | Sam. 407. "Along with her he enjoyed the pleasures of the senses without impeding the course of religion and wealth."
There is a group of cases which show a peculiar type of disagreement between the subject and the verb. The speaker changes his thought from one number to another while adding a second verb to the first which shows the normal agreement. The result is that the two verbs joined with the same subject show different numbers. AMg. suddham ravai parisāe aha rahassammi dukkadam karenti | Sut. 1. 4. 1. 18.” They speak well in the assembly but perform wicked acts in secret." pakkhande jaliyam joim necchanti vantayam bhottum Das. 2. 6. "They fall even in burning fire but do not wish to take back what is vomitted." tam appaņā na ginhanti no ya ginhāvae param | Das. 5. 1. 5. "They themselves do not take nor cause others to receive." aggi cītthai Goyama je dahanti sariratthe | U. 23. 50. "There is fire in the body, O Goyama, which burns." As. e cu hetā desam pi hāpayisauti se dukatam kachati | K. V. 14. “Those who make even a part of this to suffer do a bad
act."
B. Agreement in Person. The subject and the finite verb normally agree as regards the person. From earliest times this agreement was observed with care on account of the radical difference between the three persons.
If the subject consisted of more than one word of different persons a definite rule is observed in choosing the persons of the verb. The plural verb is in the first person rather than in the second or third and in the second person rather than in the third AMg. (aham ca tam ca) mā kule gandhană homo Das. 2. 8. "I and you, let us not become Gandhana serpents in our family." JM. aham tumam ca do vi jujjhāmo | Kum. 53. “Let I and you both fight."
This rule, though apparently arbitrary, in giving preference to the first two persons, can be easily understood by trying to ascertain the meaning of the plural forms of the first two persons. The plural form of the first person does not simply mean the plurality of the concept expressed by the singular,

Page #50

Concord in Prākrit Syntax
as the plural of Tis unthinkable, but means a combination of l' and 'You' - the first person singular and the second person singular or any other person
thought of together. Jespersen calls this a plural of approximation. Similarly the second person plural means the idea of the second person singular repeated with reference to many persons addressed or some other person to be expressed by the third person. Naturally the forms of the first and second person plural include the idea of a second and a third person in the first case and of the third person in the second. Thus it will be seen that the verbal forms in these cases exactly correspond with the combined nature of the words forming the subject and follow the rules of a very accurate concord.
Principal Exceptions to the rules of Agreement in Person
The illustrations showing the disagreement in person cannot but be mere lapses on the part of the authors or occasioned by some special conditions. In AMg. we often find such cases, as : bhikkhū amokkhãe parivvaejjāsi | Sut. 1. 1. 4. 13. "The monk should wander till he gets liberated." neva sayam dandam samārabhejjā nevannehim dandam samārambhāvejjā dandaṁ samārambhante vi aññe na samanujāņāmi jāvajjīvāe Das. 4. “He should not inflict punishment, nor cause others to inflict punishment nor consent others inflicting punishment.” In all such cases which usually occur at the end of a chapter we must suppose some kind of change of thought in the mind of the writers, who try to conclude the discourse on some moral point with a direct address to the hearers, forgetting for the time being that the whole of the preceding chapter was a general statement in the third person. Some other cases of the disagreement in person are : AMg. na tumam jāne anahassa attham pottham ca patthivă | U. 26. 16. “You do not know the meaning and import of the word 'helpless' O King," jahā vayam dhammam ajānamānā pāvam purā kammam akäsi mohā | U. 14. 20. "As we have done wrong acts in our former lives, not knowing the religion.” P. jarājajjaritā honti hatthapādā anassavā | yassa so vihatatthāmo katham dhammam carissasi | Dh. Co. 1. 5. "How can he follow the religion, when he has lost his vigour and when his hand and feet are shattered by old age and incapable of hearing ?"
C. Agreement in Gender. The agreement between the subject and the predicate as regards gender in not very rigidly fixed. Usually the predicate agrees with the gender of the subject whenever it is an adjective or a verbal derivative. This agreement pertains to the grammatical gender of the word only. JM. sayameva appă vivāhāvio | Erz. 54. 10. "I myself married my self.”

Page #51

42
Amrita
Here the word appā which is grammatically Masculine refers in reality to a woman. If however the subject is a pronoun of the first or second person where there is no distinction of gender, the predicate takes the natural gender of the subject. AMg. jaham tena pariccattā | U. 22. 29. “That I am abandoned by him.” Here aham refers to Rājīmatī. JM. ahań jakkho ettha-nivāsi | Erz. 21. 33. “I am a demi-god living here.” tao mi junnanteuraṁ gayā | Vas. 283. 16. "Then I went to the old harem." P. aham sāmi etassa kapirājassa hadayamamsam khăditukāmā | PR. 1. 7. “I am, O Lord, desirous of eating the flesh of the heart of this king of monkeys."
When the subject is made up of more than one noun of different genders and the predicate is a verbal derivative which must agree with the subject in gender, we find a peculiar development in Prākrits to satisfy the demands of congruence. Even in the earliest stage of Sanskrit, there is no distinction between the animate and inanimate gender, by which, in case of animate subjects of both the genders Masculine is used and in case of inanimate things the neuter is preferred. Sanskrit, however, prefers Masculine and Feminine and Neuter in all other cases. This position is reflected in the early stage of the various Prākrits. AMg. kumāra dovi purohito tassa jasa ya patti rāyattha devi kamalāvai ya kāmagune virattā | U. 14. 3-4. “Both the boys, the priest and his wife Jasā, the king and the queen Kamalāvai, all became disgusted towards pleasures.” rāyā saha devīe māhano ya purohio | mahani dāragā ceva savve to pariņivvudā | U. 14, 53. “The king along with the queen, the Brahmān priest, the Brahmān woman and the sons, all became liberated.” māyā ya loho ya pavaddhamāṇā | Das. 8. 40. “Deceit and greed growing." P. yesam rāgo ca doso ca avijjā ca viräjitā | Th. 1. 282.'"Those who have removed attachment, hatred and ignorance." cinnā angā ca magadhā vajji kasi ca kosalā | Th. 2. 117. "I have wandered over Anga, Magadha, Vajji, Kāsi and Kosala."
In the later stage of the Prākrits, however, it was thought better to put the predicate in the neuter gender whenever it referred to a subject consisting of both a Masculine and Feminine word. The neuter gender thus acquired the quality of a neutral word. P. puttādārāni posentā dhanam vindanti mānavā Th. 2. 177. "Men acquire wealth and support sons and wives.” Sometimes a compound of a masculine and a feminine noun becomes neuter. P. tasma sitä puttadāram dhanaṁ ca / Th. 1. 188. "Thus attached to wealth, wife and children."
In some cases, however, the predicate agrees with the nearest word. P.

Page #52

Concord in Prākrit Syntax
43
sīlāni brahmācariyam pavvajjā dukkarā puttāka Th. 2. 463. "O child, rules of good conduct, celibacy and asceticism are difficult.” All rules are however broken in such cases as : As. etāye athāye iyam likhite K. IV. “For this purpose was this edict written." JM. tato mayā se asilatthi dāvio | Vas. 41. 15. "Then I showed him the sword.”
Principal Exceptions to the Agreement in Gender
As seen above, if the subject is made up of two nouns one of which is Masculine and the other Feminine the usage of later Prākrit prefers the predicate in the Neutral plural and not in the Masculine as is the case in Sanskrit. This usage is continued in Modern Indian Languages like Marāthi and Gujarāti. It becomes very frequent from the stage of Jain Māhārāstri onwards. · suratūsave samatte donni vi kheyālasangamangāi I annonnabhuyālīñgaņasuhena niddam pavannāi Pau. 16. 18. “When the sport of love was over, both of them went to sleep with their limbs exhausted with exertion and with the happiness of embracing each other." tena amham pesiyam eu vairajangho sirimai ya tti | mhe viulena khandāvārena patthiyāni puttam nayare thaveūna Vas. 176. 15. “He sent us (a message) that Vairajangha and Sirimai should come. We then started with a big camp after placing the son in the town." tao amhe pattāṇi kamena saravane āvāsiyāni | Vas. 166. 21. "Then we came there in due course and lived in Saravana." pasuttaim pallavasattha e | Sm. VI. "We slept on the bed of leaves" thiyāim egadese | Sm. VI. "We remained in one place.” suyam ca amhānam āgamanań eehim | ao ahiyayaram lajjiyāņi | Sm. 145. “They heard of our arrival and became greatly ashamed." tāhe rāyā sa ya jayahatthimi ārūdhāim | Erz. 34. 29. "Then the king and the queen mounted the victorious elephant.” tena tāni ammāpiyaro pucchiyāni | Erz. 37. 29. "He asked them about his parents." (ammāpiyaro) virahamahāgahagahiyaim citthanti | Erz. 77. 31. "The parents remain being troubled by the great devil of separation." (rāyā devi ya) niyabhavane gayāim | Erz. 84. 6. “The king and the queen went to their own palace." (manippaho mayanarehã ya) uvavitthāim tayantie | Erz. 44. 14. "Manippahā and Mayanaraehā sat near him." Ap.devīnivãi romanciyai | Nk. 2.8 6. "The king and the queen were horripilated." māyāsuyäi amayai gehaho Nyk. 3. 9. 5. "The mother and the son came back to the house." According to this usage the neuter plural of a word denotes a pair of a male and a female, while in Sanskrit it was expressed by the dual of a Masculine noun. Ap. sari vi dovi cakkai San. 544. 3. "In the lake a pair of Cakravāka birds." A similar usage is pointed out by Paul to exist in the old Germanic dialects where a pronoun or an adjective referring to a maculine and a

Page #53

44
Amrita
feminine noun was put in the neuter gender. It apears therefore that the neuter gender is used where the difference of gender is not to be taken note of. This is also seen in some modern Indian languages where the neuter gender is used when the specific sex of a being is not known..
If the predicate is a noun it is not free to agree with the subject as regards gender and so retains its original gender. P. appamādo amatapadam pamado maccuno padam | Dh. 21. "Carefulness is the place of immortality and carelessness that of death." santutthi paramaṁ dhanam | Dh. 204. "Satisfaction is the best wealth.” saddhā bijaṁ tapo vutthi paññā me yuganangalam hiri isā mano yottam sati me phālapācanam | Sn. 77. "Faith is the seed, penance is the shower, intelligence is my plough, shame is the pole, mind the strap and mindfulness my ploughshare." AMg. tavo joi jivo joithānām jogā suyā sarīram kārisangam kammehā sañjamajogasanti homam hunami isinaṁ pasattham | U. 12. 44. "Penance is the fire, the soul the altar, the activities the laddle, the body the fuel, Karma is the wood, self-control is the control of activities, thus I perform the sacrifice praised by the sages."
D. Agreement in Voice. The general rule that in the active voice the verbal form agrees with the subject and in the passive voice it agrees with the object is on the whole closely observed in Sanskrit except in case of a few past pasive participles of some intransitive verbs. In the Prākrits, however we find the slow process by which the active construction has replaced the passive one though the grammatical forms remained passive. In this connection two important cases are to be considered. The one is that of the past passive participles which are used differently than in Sanskrit. Even when they are derived from transitive verbs, they are found used in an active construction. AMg. mukko mi visabhakkhanam | U. 23. 46. “I have abandoned the eating of poison." JM. tco ganaharo jambūnāmassa parisāe ya parikahio | Vas. 3. 15. “Then the Ganadhara preached to Jambū and the assembly." so ya kilesena ya tam mahilam moyavio Vas. 58. 4. "He freed the woman with great exertion.” dūram gantūna ya sarajālaṁ pavuttho rohinīsuyarahovari | Vas. 81. 26. "Going to a distance he started to shower arrows on the chariot of the son of Rohinī.” bhikkhā dinnā dinnā paaio | Vas. 95. 6. “He ate all the alms given.” kei nārayaṁ pasaṁsiyā kei pavvayam | Vas. 191. 18. “Some praised Närada and others Pavvaya.” dasaraho pavutto Pau. 27. 16. “Dasaraha spoke.” lankāhivo pavutto | Pau. 6. 111. “The lord of Lankā spoke." Ap. kai bhaniu | Nyk. 1. 3. 13. “The poet said." iya munivi to vi padilaviu bhāya | Kr. 3. 5. 2. "Knowing this he said to the brother.” Secondly many intransitive verbs use their past passive participles actively. AMg. tão so

Page #54

Concord in Prākrit Syntax
45
pahasio rāyā senio magahāhivo | U. 20. 10. “Then Seniya, the king of Magha laughed.” JM. tam ca datthūna parunno rāyā | Vas. 17. 21. “Seeing it the king lamented." koühallena me jão | Vas. 56. 20. "My curiosity arose."
There are cases as well where some kind of conclusion between the two voices is to be found. JM. paosena ya tena piyaram bandhiūna rajjam pasāsai | Vas. 119. 18. "With rage he bound his father and rules over the kingdom.” (It is just possible to regard tena as a demonstrative adjective of paoseņa.) Ap. to apparāiņa pabhanevi mahuravkharihi, asanam ti garuyanurāina, pucchai | Kum. J. 29. "Then the king Appa having uttered the word 'a seat' asked with great affection.”
E. Subject with More than One Verb. According to the Sprachgefuhl of the Indians the verb or the verbal form constitutes a sentence and so when a subject is joined with more than one verb, there result in reality more than one sentence. The verbs are usually joined with some particle either conjunctive or disjunctive. But in Ap. we very often find a subject joined with more than one verb without the use of any such particle like ca or api, and the relation between the subject and the verbs is made easy to understand by often putting the subject between the two verbs or verbal forms. In the early Prākrits two or more verbs were used with one subject without any conjunctive particle only when they expressed the same idea or were synonymous. AMg. bhāvao nam kevalanāni savve bhāve jānai pāsai Nandi. “As regards the modifications the omniscient teacher sees and knows all modifications.” JM. jo padhai sunai puriso so bohiphalaṁ samajjei | Pau. 20. 52. "The man who studies and hears gets the fruit of enlightenment. "The disjunctive particle vā is also absent in Apabhraṁsa. Ap. chijjau bhijjau jāu khau joiya ehu sarīru | P-pr. 1. 72. “O Yogin, let this body be cut or be broken or be destroyed.” vandau nindau padikamau bhāu asuddhau jāsu | Ppr. 2. 66. “Let him bow or censure or perform the Pratikramana, when his mind is impure.” In the following illustrations the subject is put between the two verbs. Ap. janasavaņasuhu janai | tā jinavaro bhanai | Hp. 82. 13. 1. "Then the Jina produced delight to the ears of men and spoke.” tā niyaņāņu payāsai, tāhā bhadārau bhāsai Hp. 82. 17. 4. “Then the great man showed his knowledge and spoke to them.” anukampai sarkhu sāhu cavai | Hp. 82. 17. 4. "The sage took pity and spoke.” A similar use is to be found with the participles of the past. osāriu nayaru ņīvāriu | Hp. 84. 1. 14. “He drove the people of the town away, and forbid them.” murāri nivāriu halinā mantimagge sañcāliu | Hp. 87. 6. 3. "He controlled Murāri and brought him back to the way of deliberation.” The contrary usage of joining one participle with two

Page #55

Amrita
subjects is found in : Ap. punu vi bhuyangaselapavisanu kiyaü tehi kiyayaninnäsanu Hp. 92. 10. 2. "Then they effected entrance into Bhuyangasela and the killing of Kiyaya.”
i nima sfound inp. na
pun
, withinyangos element restauranter
As pointed out by Alsdorf this curious construction in the Ap. language is due to the fact that the conjunctive particle ca is absent in this language. Other related words like vi, punu, conated something more than the mere meaning of ca and were not used to join two simple sentences. The natural result of this fact was the putting together of two verbs or two subjects without a connecting word and in course of time the stylistic peculiarity of putting the single word between the other two connected with it arose to give facility to the understanding which it admirably performed. II. AGREEMENT BETWEEN THE SUBSTANTIVE AND THE ADJECTIVE.
On the whole the adjective agrees with the substantive on which it depends in Gender, Number and Case. The exceptions to this rule are only occasional
When a substantive is pointed out by means of a superlative from a group of nouns usually in the Genitive and the two nouns differ in Gender, the superlative usually agrees with the noun in the nominative and not with the one in the Genitive. P. saccam have sādutaram rasanam Sn. 182. "Truth is the sweetest of all flavours." AMg. cando va tārāna mahānubhāve Sut. 1. 6. 19. "Just as the moon is the most lustrous among the stars." gandhesu vā candanam āhu settham Sut. 1. 6. 19. “As sandal is the best among the scents." tavesu vā uttamam bambhacariyam Sut. 1. 6. 23. "Celibacy is the best among the penances.” But the older method of agremeent is found in AMg? thaniyam ca saddana anuttare ū | Sut. 1. 6. 10. "The thunder is the best among the sounds." Thus it is clear that attraction is the chief cause of deciding the gender of the superlative form of the adjective, even though we expect, according to meaning, that it should agree with the genitive, as all such sentences mean that one of the many is the predicate which is put in the nominative. But as the word in the genitive was not repeated in the nominative form, as being not necessary for the expression of meaning, it was found impossible to make the adjective agree with a noun which was unexpressed. The other substantive also in the nominative present in the sentence as the subject naturally attracted the adjective and ultimately came to agree with it in gender. The contrary illustrations in Prākrit are merely occasional.
The agreement of an adjective with more than one substantive is on

Page #56

Concord in Präkrit Syntax
47
parallel- lines with the agreement of the nominal predicate with the subject. In case of adjectives, however the usual practice appears to be to make it agree with the nearest substantive. AMg. taham piyā surā sīhū merao ya mahūni ya U. 19. 70. “You like wine, drink, and other varieties of wine." pāio mi jalantio vasão ruhirāni ya U. 19. 70. “I was made to drink burning fat and blood." In case of predicative adjective we find the tendency to use only one form with reference to nouns of all genders. P. pavivekiyam duḥkhaṁ seyyo yaṁ ca kāmamayaṁ sukham Th. 1. 669. “Misery due to knowledge is better than happiness due to sensual pleasures.” maranam dhammikam seyyo yam ce jīve adhammikam | Th. 1. 670. “Death due to religion is better than life due to irreligion." sangāme me matam seyyo yam ca jīve parājito | Th. 1. 194. “It is better for me to die on the battle field than to live with defeat." garahā va seyyo viññühi yam ce bālapasaṁsanā | Th. 1. 668. “It is better to have censure than praise of a fool.” tasmă hi pañña va dhanena seyyo | Th. 1. 784. “Therefore knowledge is better than wealth.”
The cases of disagreement can be grouped under following heads. (i) Disagreement in Gender : AMg. teu cittamantam akkhāyā Das. 4. "It is preached that fire has life.” The adjective in this case is neuter even though the substantive teū is feminine, probably because of the gender of the word in Sanskrit. chajjīvaniyā nāmam ajjhayanam | Das. 4. "The chapter called Six Groups of Living Beings.” JM. ihaṁ puna samīve giriküdam nāma gāmo | Vas. 182. 6. "Here near about there is a village called Girikūda." kumbhakārāvekkho nāma pattanam tassa nāmena kayam | Erz. 34. 8. “A town was established after him with the name Kumbhakārāvekkho." P. attā have jitas seyo | Dh. 104. “The self is better when conquered.” Ap. nayanai honti juvāṇaha muddhao | tarunivayanadassanarasaluddhao | Bh. 48. 8. "the eyes of the youths become deluded, being attached to the flavour of looking at the faces of the maidens.” tinni mi täi asi guṇavantao | tinni mi jinavarasāsaņi bhattaoBh. 81. 5. "Those three were virtuous and also devotees of the preaching of the Jina." puri citthai padaliputtu namu | Kum. S. 1. “There is a town by name Padaliputtā.” As ese hi setthe kammaṁ am dhammānusāsanam | K. IV.” This is the best work namely to preach piety.” (ii) Disagreement in Number. Cases of this disagreement are extremely rare and doubtful. Ap. dīvāna pahānhi jambudīve | Kr. 1. 3. 1. "In Jambudvīpa the best of the continents." Even in this case it is better to admit some kind of misreading than to think of a conscious fact of disagreement. (iii) Disagreement in case. AMg. Sammaddamāni pānāñ bīyāni hariyani ya asañjamakarim naccā tārisiṁ parivajjae | Das. 5. 1. 29. “Knowing that a

Page #57

Amrita
woman who is trampling on living beings, seeds and green grass is devoid of self-control, one should avoid her (alms).” pāșai jattha kesīkumārasamane mahaimahāliyae mahaccaparisāe majjhagae mahayā saddenam dhammā maikkhamanam ry. 19. “Where he sees monk Kesi sitting in the middle of that great assembly and preaching religion with a loud voice." III. AGREEMENT OF NUMERALS
The agreement of numbers is not homogeneous. All the ordinals are regarded as regular adjectives and therefore agree with the noun on which they depend in all respects. But the cardinals are not of one and the same nature. Some of them i.e. the first 19 numbers are adjectives, while those above them are regarded as substantives. They are further used collectively or distributively, which affects their agreement as regards the aspect of number, with reference to the noun and also their inherent meaning. This was caused by the supposition that the application of the category of number to a numeral may involve some kind of repetition.
The cardinal number meaning one fully agrees with the noun as regards Gender, Number and Case thus behaving like a regular adjective. By its very nature it cannot have a plural form. But this numeral includes not only the idea of 'one' but also that of an indefinite article which is a purely adjectival sense. In this second capacity it has a plural form meaning 'some'. Probably this dual nature of the numeral is the cause of its fully adjectival concord. As. eke mrige | M. 1. 5. "one deer” eko mago | G. 1. 11. ekaṁ samvaccharam | S. 5. 11. "One year.” AMg. pasai tas maham egam paumavarapondariyam | Sut. 2. 1. 2. “He sees a big, beautiful, excellent lotus." māhanā samanā ege | Sut. 1. 1. 2. 14. “Some Brahmāns and monks." JM. egāe haṁsie gabhammi | Erx. 1.14. "In the womb of a goose.” vahena egena ceva sarena vinivāyiyā | Erz. 1. 13. "The hunter killed both of them with one arrow."
The cardinals from two to eighteen are regarded as adjectives. But by the nature of their meaning they are always used in the plural. Moreover they have only one set of forms for all the three genders. As. majura duvi | S. 1. 3. "Two peacocks." do morā| G. 1. 11. duvehi yeva ākālehi T. VII. "with two akālas.” etāni pi cu timni pānāni | J. 1. 4. “These three animals also.” catparo rājano G. XIII. 8. “Four kings." pañcasu pañcasu vasesu | K. III. 7. "In every five years.” sașu vi yojanasateșu | M. XIII. 9. “In the six hundred Yojanas." AMg. donni vi rāyāno | Nir. 26. “Both the kings." tassa bhajja duve asi U. 22. 2. "He had two wives.” jahā ya tinni vaniyā U. 7. 14. Just as

Page #58

Concord in Prākrit Syntax
49
three merchants." tassa nam ānandassa gāhāvaissa cattāri hirannakodio nihānapauttão | Upa. 4. "The householder Ananda has his four crores of gold coins deposited." eesim coddasanham mahāsuminānam annayare mahāsumine pāsittānam | Ny. 1. 12. “Having seen four great dreams out of these fourteen great dreams." atthārasahim samanasahassehim saddhis samparivude | Ny. 5. 53. “Surrounded by eighteen thousand monks." P. dve vă tinni vā rallindivāni | DN. 2. 327. "Two or three days and nights." callunnam māsānam accayena Sn. 99. "At the lapse of four months." solasannam puggalāham Mil. 310. "Of sixteen men.'
In case of ti 'three and catu 'four' Pāli preserves two sets of forms, one for the Masculine and Neuter and the other for Feminine. But even here confusion between the two sets of forms is met with, which thus paves the way for the later development. sakhiyo tīņi janiyo | Th. 2. 518. "we three friends."
A similar lax use of the numerals is to be found in the Purānas where the form trayah is used with words of all the three genders. varsāni bhavitā trayah | tasya putraḥ samāḥ trayaḥ | bhavisyati samāḥ trayaḥ. This similarity is taken by Pargiter as pointing out that the original language of these Purānic passages was Prākrit. It is, however, more natural to suppose that this syntactical similarity between the two is a result of the popular tendencies affecting both.
Numerals over twenty are all regarded as substantives and stand in apposition to the noun which they qualify. Their only agreement pertains to the case forms. AMg. samaņenam...nāyānam egūnavīsam ajjhayaņā pannatta | Ny. 1. 5. "The ascetic has preached 19 chapters of the Nāyas." evaṁ khalu amhaṁ sami suminasatthamsi bāyālīsam suminā tīsam mahāsuminā bāvattarim savvasumiņā ņidditthā | Ny 1. 12. "Thus indeed, O Lord, in our science of dreams there are 42 dreams and 30 great dreams and so in all 72 dreams preached.” bhogasamattham jāņittā battīsāe ibbhakulabāliy āņam egadivasenam pānim genhāventi Ny. 5. 53. “Knowing him capable of enjoyment they marry him with 32 daughters of merchants on one and the same day." eesim tīsae mahāsuminānam coddasa mahāsumine päsittānam | Ny. 1. 12. "Having seen fourteen great dreams out of these 30 great dreams." cattālisäe ajjiyāsahassehim saddhiṁ samparivude Ny. 5. 53. “Surrounded by 40 thousand nuns." ekkavīsāe titthayarehim ikkhāgakulasamuppanaheim | Kap. S. 33. "By the 21 Tirthakaras born in the family of the Ikkhagas." pañcahattarie vāsheim Nir. 24. "With 75 years." JM. satthim bhattāi anasanāe chetta | Erz.

Page #59

50
Amrita
34. 4. "Having cut short 60 meals with fasting." P. dasannaṁ pi dadāmi, vīsāya pi dadāmi, tiṁsāya pi dadāmi, cattārisāya pi dadāmi, paññāsāya pi dadāmi | Sn. 86. “I give to ten, twenty, thirty, forty or to fifty.” vīsatiyā yojanesu tiṁsāga yojanesu | M. 11. 162. "At the distance of 20 or 30 yojanas.” ekūnapaññāsāya kandehi | Jat. III. 220. “With 49 arrows." chattissatiyā sotehi | Dh. Co. IV. 48. “With 36 streams” atthacattārisam vassāni Sn. 289. "For 48 years." imesam tevīsatiyā buddhānam santike Dh. Co. I. 48. “Near these 23 Buddhās.”
Words expressing bigger numbers like hundred thousand, etc. are also regarded as substantives and so stand in apposition to the words which they qualify. But sometimes they may be also thought of as adjectives. As. yojanaśatesu | K. XIII. 6. "In hundred Yojans.” vassasatāni | K. IV. 9. "Hundred years." bahūni panasahasāni | K. I. 3. "Many thousand beings." P. gandhabbā cha sahassāni | Th. 164. “Sixty thousand Gandhabbas." bhikkhuniyo sahasam | Mhvs. 5. 187. "Thousand nuns.” pañcasatāni candālapurise | Mhvs. 10. 91. "Five hundred low-caste men.” gāthā satam | Dh. 102, "Hundred Gāthās.” pañcasate vānije | Jat. II. 123. “Five hundred merchants.” AMg. pañco sayā | Kap. 142. “Five hundred."
In many cases the numeral remains without any inflection. P. pannāsa yojanā sattahi yojana | Dh. Co. III. 217. “Fifty and sixty Yojans," pañcapaññsa vassāhi...pañcavīsati vassāni Th. 1. 904. “Fifty five years, twenty five years." ekūnapaññāsa jane | Jat. III. 220. “Forty-nine people.” AMg. solasa rogāyankā | Vip. 22. "Sixteen diseases and sicknesses.” sattari vāsāi | Kap. S. 70. “Seventy years." - Numerals over twenty being substantives, they can also govern the Genitive of the noun, P. parosahassam bhikkhūnam Th. 1. 1238. “More than thousand monks.” vihārānam pañcasalam| Mhvs. 12. 33. “Five hundred Vihāras." Satthi arahatam akā | 1. 14. "He made sixty Arhats.” AMg. atthasayam māhanadāraganam atthasayam khattiyadāragānam | Vip. . 107. “Eight hundred Brahmān and Ksatriya boys.” JM. pañcasayā suvannassa dāūna taṁ parenei | Erz. 39. 27. “Giving five hundred gold coins he marries her.” donni lakkhā mayagalānam donni sahassā rahanam pañca ajuyam hayānam satta kodio pāyāijanānam | Erz. 39. 3 1-32. "Two lakhs of elephants, two thousands of charitors, five Ayutas of horses and seven crores of foot soldiers.” puttanam satthi sahassa | Sag. “Sixty thousand sons."
Numerals also enter into compounds with the nouns. AMg. attthasahassalakkhanadharo | U. 22.5, “Possessed of eight thousand marks."

Page #60

Concord in Prākrit Syntax
duvālasangiņo cauddasapuvviņo samattaganipidakadhūragā | Kap. S. 2. “Knowing 12 Angas, 14 Pūrvas and the whole of the Ganipidaga." P. atthavassaṁ sattamāsas rājā rajjam akārayi | Mhvs. 35. 46. "The king ruled for 8 years and 7 months.” JM. sirāsayāi satteva nava nhācusayā bhave Erz. 4. 14. "There are 700 veins and 900 sinews." As. tedasavasābhisitena | D. V. 22. "Annointed king for thirteen years." dasavasābhisite | K. VIII. 22. "Annointed ten years.” Sampunnacatuvīsativaso OBJ. 1. 1. "Having finished 24 years. IV. AGREEMENT OF PRONOUNS
Usually the pronoun agrees with the noun in Number and Gender. But on account of the uncertain nature of the Gender of Prākrit nouns and the meagre remnants of the pronominal forms sometimes this agreement is lacking and the usage in Prākrit often disagrees with the one prevalent in Sanskrit.
A. Disagreement in Gender Many versions of the inscriptions of Asoka write, ayam dhrammadipi | S. I. "This religious edict." ayi dhrammadipiM . 1. I. Others write : iyam dhammalipi | K. I Cp. also yadā ayam dhammalipi likhiti G. I. "When this religious edict was written.” ese hi (sethe) kammam | K. IV. “This is the best duty." In AMg. we often find a neuter pronoun agreeing with a masculine noun. Jāvanti loe pānā | Das. VI 10. "All living beings in the world." jāvantavijjāpurisā savve te dukkhasambhavā U. 6. 1. "As many ignorant persons there are, are all born in misery." eyāvanti savvāvanti logamsi kammasamārambhā | Ay. 1. 1. 1. 5. 7. "All the activities in the whole world." āvanti keyāvanti logamsi samaņā ya māhaņā ya | Ay. 1. 4. 2. 3. "The monks and Brahmāns who are to be found in the whole world." yaim tumāim yāim te jānagā | Ay. 2. 4. 1. 8. "You and your knowers." je garahiya saniyānappaoga na tāni sevanti sudhiradhammā | Sut. 1. 13. 19. "Those blamable and sinful practices the pious monks do not observe." Other cases are only sporadic. Se dittham ca ne | Ay. 1. 4. 2. 3. "That we have seen.” Se gayaṁ eyam bhagavam | Bhag. 15.1. 11. "That is gone, O Lord.” annavamsi mahohamsi nāvā vipparidhāvai | jaṁsi goyamamārudho | U. 23. 70. "In the great ocean there is a boat running, on which you have boarded." jassi guhāe | Sut. 5. 1. 12. "In which cave."
If the pronoun is made to refer to more than one noun which differ in gender, it is placed in the Masculine when the words referred to are masculine and feminine. P. (sassu sasuro ca) te maṁ pitu gharaṁ pati nayiṁsu vimanā dukkhena Th. 2. 419. “They carried me to the house of my father

Page #61

52
Amrita
with a sorrowful mind." If the pronoun refers to words including a neuter noun, it is often put in the Neuter plural. AMg. cīrājinam nagininaṁ jadi sanghädimundanam eyāni vi na tāyanti | U. 5. 21. "A bark garment, nakedness, matted hair, upper garment and shaving, all these do not protect." P. uttitthapindo uccho ca psmsukūlam ca cīvaram etāni abhisambhonti Th. 2. 329. “Alms specially prepared (?) food collected, a garment thrown off and bark, attaining these.” Sometimes the form in the Masculine is used even in such cases. AMg. egappā ajie sattū kasāyā indiyāni ya te jinittu | U. 23. 38. "When the self is not conquered it is an enemy and so also the senses and passions. Having conquered them.”
B. Disagreement in Number P. tapena brahmācariyene samyamena damena ca etena barāhmano hoti | Th. 1. 631. “By penance, celibacy, selfcontrol and mortification, by these does one become a Brahmān.”.
C. Disagreement in Case P. kassa sāsanamāgamma labhate tam asokatā | Th. 1. “By following which law does one get the freedom from grief ?”
The relative pronoun agrees with its antecedent in number and gender, but its case is determined by the clause in which it.occurs. As. am kichi dakkhāmi hakam tam ichāmi | D. I. 2. "whatever I see I wish." ye ādikare kayanasa se dukaraṁ karoti | M. V. 19. "He who begins a good act does a difficult task.” yo tu etaṁ desaṁ pi hāpesati so dukataṁ käsati G. V. 3. “He who causes even a part of it to decrease does a bad act.” AMg. je ya buddha mahābhāgā...suddham tesiṁ parakkantam | Sut. 1. 8. 23. "Those who are enlightened and great...their activity is pure.”
Mūliyam te pavesanti mānusam jonim enti je | U. 7. 19. "Those who are born as men again are those who obtain their capital.” jesim tu vivulā sikkhā mūliyam te aicchhiyā | U. 7. 21. “Those who have much instructions have exceeded their capital." JM. jo eyam dandagam genhissai so rāyā hohii Erz. 37. 1. “He who will take this staff will become a king.” jaṁ jena kayam kammaṁ.tam tena veiyavvam | Erz. 42. 21. "Whosoever does an act will have to suffer for it.” jenäham hatthisambhamāo rakkhiyā tena saha pāniggahanam jai na hoi Erz. 17. 15. "If I cannot marry him who has protected me from the attack of the elephant." je sappurisā havanti te maranavasanam bahu mannanti Erz. 42. 15. "Those who are great consider the calamity of death highly." P. tam ca kammam katam sādhu yam katvā nānutappati Dh. 68. "That act is well done by doing which a man does not repent. yesim sannicayo natthi ye pariññātabhojanā...tesam gati durannayā | Dh. 92. "Those

Page #62

Concord in Prākrit Syntax
who have no store, and who know the measure of food, their track is difficult to find.” yassindriyāni samatham gatāni...devā pi tassa pihayanti | Dh. 94. "Gods honour him who controls his senses." abbahi vata me sallam yam asi hadayanissitam | Jat. IV. 62. "He removed the dart which was rankling in my heart.” yo na hanti na ghāteti...veraṁ tesa na kenaci | Jat. IV. 71. "One who does not kill nor cause to kill is not inimical with anybody.” evam kathenti ye saddahenti Th. 2. 522. “Thus tell those who believe."
There are, however, a few exceptions which go against this general rule. (i) Disagreement in Number. AMg. acchandā je na bhuñjanti na se cãi tti vuccai | Das. 2. 2. "Those who do not enjoy because unable to do so, are not called relinquishers." jai taṁ kāhisi bhāvam jā jā dacchasi nārio | U. 22. 44. "If you love all women you chance to see." ja ya saccā avattavvā saccamosā ya jā musā jā ya buddhehi nainnā na tam bhasejja pannavam | Das. 7. 2. "A wise man should not speak those words which are true but not to be spoken, those which are both true and false, those which are false and those not used by great men." ke ettha khattā uvajoiyā vā...khalejja jo nam | U. 1. 18. "Who are the door-keepers appointed here who will remove him?" Ap. appā laddhau jena paru te paramappa navevi | Jos. 1. "Having saluted those highest souls who have obtained the highest self.” (ii) Disagreement in Gender. AMg. jāe saddhāe nikkhanto tam eva anupaliyā | Das. Ay. 1. 1. 3. "Preserving the enthusiasm with which he has become a monk.” P. jānāmi attano satta jātiyo yassāyaṁ phalaṁ vipāko | Th. 2. 434. “I know seven of my lives whose bad effect and result is this."
The relative pronoun is neuter when it refers to a whole sentence, because the neuter gender is thought to approach the neutral form closely. P. dubbacanam kim sakka kalum yam mam videssate bhattā| Th. 2. 418. "Is it possible to use bad words because my husband hates me ?” tassetam kammaphalam yam mam apakaritūna gacchanti | Th. 2. 447. “This is its effect that they go abandoning me." kam te aparādhitam mayā yam mam ovariyana titthasi | Th. 2. 367. "What fault have I committed against you that you stand blocking my way ?” AMg. tassa me apadikkantassa imam eyārisam phalam jānamāṇo vi jam dhammam kāmabhogesu mucchio | U. 13. 19. "This is the effect of 'my deed for which I have not repented, that even knowing religion I am attached to plesures of senses.” JM. ajjautta sohanam kayam jam so durappā nihao | Erz. 8. 29. “O Lord, you did well in killing that wicked
man."
If the predicate of a relative pronoun is a noun having a different

Page #63

54
Amrita
gender from the gender of the antecedent noun, the relative pronoun agrees with the predicate rather than its antecedent. The demonstrative pronoun naturally agrees with the antecedent noun. As. ese tu parisrave yaṁ apuñam G. X. 3. “That peril is sin.” e hi etrake mangale samsayike tam S. IX. 6. "For the ceremonial of this world is of doutful efficacy.” ese hi srethe am dhrammanusāśana | M. IV. 17. “This is the best deed, the promulgation of religion.” P. appamatto ayam gandho yāyam tagaracanani | Dh. 56. “Very little is this scent, the scent of sandal and aloe.” yassa bhāriya...śīlavati patibbatā dāresu ve sotthānaṁ tadāhu Jat. IV. 76. “Whose wife is endowed with good conduct and devoted to her husband, she is called the best resort among wives." AMg. je māhanā jāivijjovaveyā tāim tu khettāi supesalāim | U. 12. 13. "Those Brahmāns who are endowed with birth and knowledge are the best fields."
If the sentence has more than one predicate the relative pronoun agrees with the one nearest to it. As. esa hi dhammāpadane dhammapatipati ca yā iyam d yā dāne S ce Socaye mad ve sādhave ca lokasa hevam vadhisati ti | T. VII. "The performance of law and the belief in law are but how kindness, charity, truthfulness, cleanliness, happiness and goodness will be increasing among the people." P. am mahārāja y ññevoñanām să eva paññā . Mil. 42. “Yes, o king, what is knowledge is intellect."
OOO
Concord in Prākrit Syntax
ABORI. XVI. 1. 2. 1939
.30

Page #64

Instrumental and Locative in Ardha-māgadhi
The syntax of the Prākrit languages is important in more than one way. It forms the connecting link between the earlier syntax of Vedic and Classical Sanskrit and the later syntax of Modern Indian languages. Further it shows the transitionary period in a development of a synthetic language into an analytical one and as such shares the peculiarities of both. But the earlier Prākrits like Ardha-Māgadhi and Pāli do not show many traces of an
ytical syntax, the method of formation current in a synthetic language being the usual one in them. This early syntax has passed through a stage of vicissitudes in which the distinctions between the different cases were obliterated and the seven cases were ultimately reduced to two : the Nominative and the oblique case in the Modern languages. One important instance of such a case-mixture and case-confusion is investigated here.
In the older books of the Ardha-Māgadhi canon and the early works in Jain Māhārāstrī we find very often a confusion between the uses of the Instrumental and Locative cases where each one takes the place of the other besides their correct uses which are also preserved side by side. I collect below such instances where some kind of confusion between these two is found :
1. In the early works of the Ardha-Māgadhī canon there are many instances in which Instrumental is used in place of the regular Locative. eehi muni sayanehim samane asi patelasą vase, Ay. 1.. 9. 2. 4. "the sage was a monk for thirteen years in those places"; (Cp. Sīlārka, etesu pūrvoktesu śayaneșu sa munīḥ p. 279) sayanehi tassuvasaggā bhīmā āsī anegarūvā ya, Ay. 1. 9. 2. 7. "in those places he encountered many fearful obstacles"; sayanehim tattha pucchińsu egacarā vi egayā rāo Ay. 1. 9. 2. 11. "in those places he was asked why he was wandering alone and alone in the night”; (Śīlānka reads sa janehim probably to evade the difficulty). lādhehim tassuvasaggābahave Ay. 1. 9. 3. 3. "in the country of Lādha there were many obstacles for him";

Page #65

56
Amrita
Cof the Locatiksolute constrinnaneh
(śīlānka ladha nāma janapadavisesaḥ tesu ca dvirūpeșu), duccaragāņi tattha lādhehim Ay. 1. 9. 3. 6. "in that country of Lādha the places were difficult to wander through” (Šīlāńka tatra teșu lādheșu p. 282); vihāya dandam pānehim Ay. 1. 9. 3. 7. "abstaining from punishment from living beings”; evam pi tattha lādhehim aladdhapuvvo vi egayā gāmo Ay. 1. 9. 3. 8. "thus in that country of Lādha he did not once get a village to live" (śīlānka tatra lādhesu p. 282).
2. A further development of the confusion between these two cases is the interesting fact that on the model of the Locative absolute construction the Ardha-Māgadhi canon shows and Instrumental absolute construction. tam jahā, soyaparinnāņehim parihāyamāņehim cakkhuparinnāņehim parihāyamānehim ....... abhikkantam ca khalu vayam pehāe Ay. 1. 2. 1. "for example while the cognizing power of the ears, eyes etc. is waning and knowing the period of life to be short” so also, sottaparinnāņehim aparihāyamānehim Ay. 1. 2. 1.; sayanehim vīimissehim itthio tattha se parinnāyā Ay. 1. 9. 1. 6. "in the places mixed up with householders he knew the real nature of women" (śīlāňka, sayanāni vasatayah tesu .....vyāmisritesu grhasthatirthikaih); In fact this construction has led śīlānka to write in Sanskrit taiḥ srotraparijñanājaih (? naih) jarāprabhāvāt parihiyamanaih sadbhih p. 15. an anomaly which cannot be otherwise explained. The supposition of such a construction will greatly facilitate the interpretation of passages like saehim pariyāehim logam būyā kade tti ya, Sut. 1. 1. 3. 9. "in their own views they speak the world to be created"; amohāhi padantihis gihamsi na raim lahe Utt. XIV. 21. "while the nights are flying uselessly I cannot get delight in the house"; chinnāhi sāhāim tameva khanum Utt. XIV. 29. "with the branches cut it is called a stump."
3. In a great number of cases we find the Instrumental used where with the rules of a refined grammar the Locative should be found. In fact in many other cases the Locative is used in the canon in similar instances along with this curious use of the Instrumental case. sattā kāmehi mānavā Sut. 1. 1. 1. 6. "men are attached to pleasures"; abhinūmakadehi mucchie tivvam se kammehi kiccai Sut. 1. 2. 1. 4; ajjhovavannā kāmehi mucchiyā Sut. 1. 2. 3. 4. "infatuated and attached to pleasures”; pāņehim sanjae Sut. 1. 2. 3. 12. "restrained towards the living beings”; jaipaham anuparivattamāne tasathāvarehim viņighāyamei Sut. 1. 7. 3. “wandering through the course of births he comes to destruction in the births of movable and immovable beings"; saddehi rūvehi asajjamānam savvehi kāmehi viniya gehim Sut. 1. 7. 27. "without sticking to all kinds of forms and removing attachment towards all

Page #66

Instrumental and Locative in Ardha-magadhi
pleasures'; virae gämadhammehim Sut. 1. 11. 33. "abstaining from the pleasures of the senses"; savvehi pānehi vihāya dandam Sut. 1. 13. 23. "abstaining from punishment from all living beings"; virae pāvakammehim Sut. 1. 16. 1. "abstaining from sinful deeds”; kannasokkhehi saddehis pemam nābhinivesae Das. VIII. 26. "one should not show attachment towards words delightful to the ears”; visannā pāvakammehim bālā pandiyamāṇino Utt. VI. 10. "ools thinking themselves wise are stuck to evil deeds”; appamatto pamattehim pindavāyam gavesae Utt. VI. 16 "himself careful he should search the alms among careless persons"; aha coddasahiỉ thāņehim vattamāne u samjae Utt. XI. 6 "one who behaves according to these fourteen points is a self-controlled one"; savvehi bhūehi dayānukampi Utt. XXI. 13. "full of compassion and pity towards all living beings.”
4. In all such cases the correct use of the Locative is also met with throughout the canon and their number far exceeds this use of the Instrumental in such a context. Annamannesu mucchiya Sut. 1. 3. 3. 9. "infatuated of each other"; agiddhe saddaphāsesu ārambhesu anissie Sut. 1. 9. 35. “unattached to words and objects of touch, not engaged in preliminary sin"; saddesu rūvesu. asajjamane Sut. 1. 12. 22. "not attaching oneself to words and forms'; kāmagunesu giddham Utt. XX. 39 "attached to tastes”; asamsattam gihatthesu Utt. XXV. 28 "not clinging to the householders"; tao se dandam samārabhai tasesu thāvaresu ya Utt. V. 8. "then he begins to inflict punishment on movable and immovable beings"; nārīsu novagijjhejjā itthi vippajahe anagare Utt. VIII. 19 "one should not be attached to women, a houseless monk should abandon women"; vihamgamă va pupphesu dānabhattesane raya Das. I. 3. "taking delight in begging food and gifts as the birds in the flowers"; gamanāgamane cevī bhattapāne ya samjae Das. V. I. 89 “selfcontrolled in going and coming as well as food and drink”; amucchio bhoyanammi Das. V. 2. 21. "unattached to food"; samjame ya tave rayan" Das. VI. 1. "taking delight in self-control and penance"; na ya rūvesu manam kare, Das. VIII. 19. "he should not put his mind on beautiful forms"; na ya bhoyanammi giddho Das. VIII. 23, "not infatuated with food"; gadhie mihokahāsu Ay. 1. 9. 1. "attached to mutual talk”.
5. Sometimes we find both the usages side by side and mixed up in a complex manner. Suyalābhe na majjejjā jaccă tavasi buddhie Das. VIII. 30 "he should not be proud of scriptural acquisition of birth, penance and intellect'; appegaiyā meham kumāram hatthehim samghattenti evam pāehim sise potte kāyamsi Nay. 1. 1. "some strike prince Meha on his hands, feet, head, belly and body."

Page #67

Amrita
:
6. There are further good many cases where Locative should be more appropriate but the use of Instrumental can be also imagined as being in agreement with the thought, the actual relation being two sided and viewed from two different aspects. As such the correct use in Prākrit for such cases must remain doubtful. Anne annehi mucchie Sut. 1. 1. 1. 4 "others attached to still others"; eehi tīti thānehim samjae sayayam muni Sut. 1. 1. 4. 12 "the monk is always self-controlled as regards these three points"; kivä jattha ya kissanti nāisamgehi mucchiyā Sut. 1. 3. 2. 12 “where the weak persons attached to the relatives are troubled"; ajjhovavannā kāmehiṁ noijjantā gayā gihaṁ Sut. 1. 3. 2. 22. "being infatuated with pleasures and being pressed went back to the household life”; ajjhovavannā kāmehim Sut. 1. 3. 3. 13; giddhā sattā kāmehim Sut. 1. 4. 1. 14 “men attached to pleasures"; kuvvanti samthavam tāhim Sut. 1. 4. 1. 16 "they make acquaintance with them"; tao verehi rajjai Sut. 1. 8. 7. "he takes delight in hatred"; aniyae ayam vāse nāyaehi suhihi ya Sut. 1. 8. 12 "uncertain is this residence among relatives and friends”; mettim bhūehim kappae Sut. 1. 15. 3 "he should have friendship with living beings"; bhūhi na virujjhejjā Sut. 1. 15. 4. "he should not be on inimical terms with living beings”; jo rāgadosehi samo sa pujjo Das. IX. 3. 11. "who thinks attachment and hatred as equal is worthy of honour"; iha kāmagunehi mucchiyā Utt. X. 20 "here attached to pleasures"; vādehim pamjarehim ca sanniruddhe sudukkhie Utt. XXII. 14 “much pained and restrained in enclosures and cages”; evam alittam kāmehim Utt. XXV. 27. "thus unsoiled with pleasures”. .
7. The contrary practice of using Locative for Instrumental is much less common in the canon but to be found greatly in the later work, the Paumacariya of Vimala. Usucoiyā sattisu hammamānā Sut. 1. 5. 1. 8 "urged on with arrows and beaten with spears”; savvesu kāmajāīsu pasamano na lippai tai Utt. VIII. 4 "the protector who is observant is not soiled with all the sensual pleasures”; a few doubtful cases are, pāosinānaisu natthi mokkho Sut. 1. 7. 13 “there is no liberation by early bathing etc.”; jo vā dae sattiagge pahāram Das IX. 1. "one who may strike with the point of a lance".
8. In the Paumacariya we find the locative regularly used for the Instrumental along with its use in other cases. uragamahā-phanimanīsu pajjāliam 1.6. "which was lighted up with the jewels on the big hoods of the serpents”; ujjānesu ya samantao rammo II. 7. "charming everywhere with gardens"; vappinasaesu aimanaharālovas II. 12 of a very beautiful appearance with hundreds of fields; damanesu tādanesu ya....dukkham anuhavantā II. 72 "suffering misery on account of training and beating"; kāūna udarabharanam

Page #68

Instrumental and Locative in Ardha-māgadhi
59
... kumjarāibahuesu II. 112 "filling his belly with many elephants etc."; gaesu angam pellijjai II. 109 "the body is crushed with elephants"; tīsu ya valayesu hoi parinaddho III. 20 "is surrounded with three layers"; Jiņaceiesu rammā III. 30 "charming with the temples of the Jinas"; nanesu tīsu sahio"; III. 69. “possessed of three kinds of knowledge"; Jinabhavanesu manaharam III. 155 "beautiful with the temples of the Jinas”; annapāņadāņāsanesu sampūiyāna IV. 77 "having honoured with food, drink, gift and seats”; pattharapaharesu hammantā IV. 83 "beaten with stones”; this list can be extended to a considerable extent, but the present illustrations are sufficient to give a fair idea of the author's usage of the Locative for the Instrumental.
Dr. Jacobi has suggested the reason for this striking fact while discussing the Apabhramsa features of the language of the Paumacariya (Bhavisattakaha, Introduction, p. 60). As the Paumacariya had illustrations of the interchange of these two cases in the plural only he points out that this must have been also the tendency of the popular speech of the time, a fact which is also found in the later literary Apabhramsa where the Locative and Instrumental plural forms are identical. The author of the Paumacariya has in this as in many other cases allowed the popular language current at his own time to influence his own style to a certain degree probably to make his work more popular. This very feature is further inherited by the literary Apabhramśa.
Now with the more copious materials collected above from the early books of the Ardha-Māgadhi canon we can see that this interchange of the two cases stands in a different light. The writer of the Paumacariya is merely following a long-standing usage only with a little marked aptitude for it. Further our collection of the illustrations do not confine themselves exclusively to the plural number even though the majority of them are of that type. And we know that the literary Apabhramsa has no similar forms for the singular number of these cases. Jacobi has rightly rejected the probable explanation of a metrical necessity on the ground that the metrical value of the two forms is the same, and further we now have cases in prose as well, which goes to confirm such a rejection.
That even in Sanskrit two cases of this type were mixed up in the mind of the author is proved by the fact that many prepositions govern more than one case. Further the breaking up of the old syntax helped such a confusion more and the tendency has grown up slowly in course of time to make the use of these two cases quite indiscriminate. It appears that the

Page #69

60
Amrita
Instrumental first got mixed up with the Ablative and later on with the Locative. With verbs meaning "to carry" both Instrumental and Locative are found used as in skandhe or skandhena dhr, as also with verbs to mean "conquer" as yuddhe or yuddhena; while on the other hand we find such expressions as samudra iva gambhirye and use of words like śāsane and mate where the sense is more in agreement with the use of Instrumental. Further cases for shadowing this confusion would be udnā na nāvam anayanta RV. 8 "as a ship by water”; eha yātam pathibhir devayānaih RV. "come here on the god-trodden paths”; jagmur vihāyasă Mabh. “they went through the sky”; an illustration in which the same relation has two aspects which can be expressed by Instrumental and Locative is made ahim indro jaghāna RV. "in intoxication Indra killed the dragon”. Na tvayātra mayāvasthitena kāpi cinta kāryā Pañc. shows that there may be here some kind of Instrumental absolute construction in germs.
From these sporadic cases in Sanskrit both Vedic and Classical of such tendencies it can be reasonably assumed that a further development of these very tendencies makes its appearance in Prākrits in a more marked way. While the Sanskrit tradition continued to be preserved with a more or less systematic and scholarly syntax the Prākrits brought into greater play the latent forces of development making for a simplification in this field of language as in others.
DOO
Instrumental and Locative in Ardha-māgadhi
IHQ. XIII. 1. 1937
31

Page #70

A Prākrit Tendency in Rgveda
A number of considerations give Rgveda a position of pre-eminence in the whole range of the Vedic literature. As regards the language we have not only the oldest phase of the Indo-Aryan preserved in it, but we also see there many dialectal features which bring it nearer the spoken language of the time than any other Vedic work. One result of the influence of the spoken language on the literary idiom of the Rgveda is a number of Prākritism which it reveals here and there! Scholars are now beginning to see that not only the phonology of the Rgvedic dialect shows traces of Prākrit influence in a number of words, but also the so-called Prākritic tendencies of the later days find some trace in the Sandhi-rules and the morphology of its language? The general implication of these facts is clear. In spite of the attempts of the later redacters to put the Rgveda Samhitā in a form consistent with a later linguistic tradition and theory, the original features of its language have survived in places, particularly when the normalising process would go against the metre or when some plausible explanation could be imagined which can bring them in agreement with later usage. We hope to find such a Prākritic tendency in the present Samhitā of Rgveda incidentally preserved though in a limited number of cases.
The Rgveda shows, as is well known, a number of cases of double Sandhi, a feature also observed in the epic literature. Among these scholars have included a few cases where we find a Sandhi of final -am with a following vowel giving rise to a long vowel by combination. Thus Whitney? explains a case like sasvānanobhayankāram (RV. VIII. 1. 2.) as due to the loss of the final -m and the combination of samvānana with the following ubhayankaram. Support is sought in the reading of Sāmaveda which gives samvānanam in the corresponding place, although this interpretation contradicts the reading of the Padapātha samvänanā. Wackernagel notes the explanation given by Kuhno to the effect that the final -m becomes a mere

Page #71

Amrita
anusvāra as seen in the Prākrit tendency of changing all final -m's into -m. Being thus weakened the final nasal resonance disappears. Parallels to such a change can be found in abundance in the Prākrits : kaham, kaha; evvam, evva etc. As pointed out by Benfeys, Latin supplies a good parallel to this phenomenon. Quintilian has already noted that final -m in expressions like multum ille, quantum erat was slightly pronounced, though it did not absolutely vanish. Latin inscriptions in the earliest times show that final m was frequently omitted also in writing. The later development of the Vulgar Latin makes final m disappear completely?. Obviously Wackernagels does not approve of this suggestion. He himself remarks that the poets of the preclassical period allowed the final m to drop, being misled by analogical cases like tubhya : tubhyam (cp. Avesta : maibyā, taibyā etc.) and then combine the final vowel with the following one. He illustrates it with AV. rastréhá for rāstram iha (so written). Macdonell' has mechanically followed Wackernagel and has also reproduced his illustrations.
The real explanation of these cases appears to be somewhat different. It is only recently that we are understanding the importance and wide scope of a purely phonetic alternance between a long vowel and a short vowel followed by an anusvāra which is used without any morphological significance. Thus the Prākrits often show alternative forms like atthā, attham, hetthā, hettham; sammā, sammam and so on. This Prākrit alternance has, however, a far wider scope. It is found in the body of the word (stha, simha, sanamtana, sanātana; vangīsa, vāgīša) or may give rise to doublets like pabhii, pabhiim, iha, iham, vīsā, vīsam. The full implication of this 'alternance is not grasped by the earlier Prākrit grammarians like Pischel and Geiger, though they have noted a number of cases of this alternance under some empirical laws or the so-called Dreimorengesetz". With the collection of more extensive material, we now see that this alternance between a long vowel like -ā, i, or -ū and the corresponding short vowel followed by an anusvāra is purely phonetic and is not confined either to the final of a word or dependent upon the presence of some kind of a conjunct requiring simplification 2.
In the light of this well established Prākrit tendency, it is necessary to see whether a few cases in the Vedic literature and particularly in Rgveda admit of a similar explanation. Cases of adverbs of identical nature have been already noted 13 from Rgveda : itthā, ittham; kathā, katham. But no explanation has been suggested for them. If this tendency is visible in the adverbs in Rgveda it is quite legitimate to find it in other similar cases also. Thus the case of Sandhi referred to above may be brought under this

Page #72

A Prākrit Tendency in Rgveda
63
tendency with great probability. RV. VIII. 1. 2. reads : (indram)....vidvesanam samvananobhayankaram mamhistham ubhayāvinam / "to Indra, who hates (the enemies), wins over (his friends), does both, the greatest and possessing both 4. Sāyana explains saṁvanana as samvananam sammyak sambhajanīym stotrbhih. There can be no doubt, whatever his semantic explanation, that he agrees with the Padapātha in reading samvananā and taking it as an equivalent of saṁvananam, a clear case of this alternance under discussion. With this admission of the word there is no difficulty of the Sandhi. The peculiar rule of the loss of the final m has no scope here and we have the perfectly normal case of -ā and u- combining into -0- here. Following Wackernagel and Macdonell we need not put any blame on the author of the Padapatha, who has obviously admitted a word like samvānanā in this case. Such doublets appear to be the peculiarity of the popular or spoken speech. A continuation of this very fact on a greater scale is found in the Prākrit stage, though unfortunately the Prākrit scholars have not taken any note of it so long. We have a large number of cases like moyanā (mocanam) Ācārānga 1; pūyanā" (pūjanam) Ācārānga 1; māraņā (māranam), ghāyanā (ghātanam), vahanā (vahanam), lovanā (lopanam) Praśnavyākarana I. 1; anuvāsanā (anuvāsanam), tappaņā (tarpaņam) Vipākaśruta 23; gabbhsādaņā (garbhaśātanam), pädaņā (pātanam), gälaņā (galanam) Vipākaśruta 20; camara (câmaram) Kalpasūtra 61; aosaņā (ākrośanam) Nirayāvali 32; etc. The Prākrits have been extended the tendency to masculine nouns also : takkā (tarkah) Ācārānga 1; viyakkā (vitarkah) Sūtrakrtānga 11.2.21. bhuyā (bhujah, L. Sk. bhuja) Kumārapālapatibodha p. 273; ghāyā (ghātaḥ) Paumacariya 33.109. But there is no doubt that it was originally confined to neuter nouns only and was extended later on to the masculine nouns by analogy or attraction. All these cases make it highly probable that the spoken language at the earliest times had a doublet in -a for -am of a neuter word and a remnant of it may be seen in the RV. word samvanánā.
Two more cases from Rgveda may now be considered in this context. In RV. 18.2 we get : nāhamato nirayā durgahaitat tirascatā pārsvān nirgamāni / "I will not come out from this place. This is a bad passage. I will come out from the side obliquely." The Padapātha reads : durgahā etat; and Sāyaṇa remarks: durgahā durgaham. The word is derived from dur + gāhconnected with gādha 'deep'. It may therefore mean 'a low passage' (Geldner) or 'difficult to enter' (Velankar)". Wackernagel has explained the Sandhi as coming from durgaham + etat. Macdonell follows him. Prof. Velankar accepts the Padapātha reading durgahā and takes it as neuter plural

Page #73

Amrita
in spite of the singular pronoun etat and a singular meaning from the context. It is true that neuter plurals are often thought of as collective in meaning, and, as in Greek, RV. also shows a singular verb with a subject in neuter plural 16. But here we have no trace of a collective sense and we have no single case of a plural noun with a single demonstrative. Here there is no necessity of either accepting an irregular Sandhi or a breach of the usual congruence. In agreement with the Padapātha and Sāyana, the most natural course is to accept durgahā as an alternative form of durgaham. The same is the case with ĶV. J. 162. 10, Sukrtā tac chamitāraḥ krnvantu / "Let the samit priests make it well done." Here also the Padapātha has Sukrta and Sāyana equates it with sukstam. Obviously the explanation of a peculiar Sandhi is not possible here and no incongruence between tat and suksta need be accepted. We can take the word as an alternative form of sukrtam and not as neuter plural. All the three cases fully justify the Padapātha and the explanation of Sāyana to the extent that they are neuter singular forms.
Two more cases in later Vedic literature now get explained. In Taittiriya Samhitā 1. 4. 44. 2. we have the form savanedam, explained so far as savanam idam against the Padapātha savanā idam; and in K. S. 1. 4. 12. sadanedam astu. The different variants noted on the second passage!? are best explained on the supposition that sadana was either taken as Nom. Plu. Neu. (sadanāni santu) or as Acc. plu. Neu. (sadanā akarma) or even changing sadanā straightway into sadanam.
We are thus led to the conclusion that in all these cases, RV. and occasionally some later Vedic works know of an alternative form in -ī to the Neuter sing. form in -am, in conformity with a well known Prākrit tendency, probably a result of the spoken dialect influencing the Vedic language.
Annotations :
1. Wackernagel, Altindische Grammatik. I, XVII-XXI and the sections referred to
there. 2. Bloomfield. Edgerton, Emeneau, Vedic Variants, I-III. Introductory Summaries. B.
Ghosh. Indian Linguistics. VII, 54-57; IX. 30-37. S. M. Katre. Präkrit Languages,
pp. 65-73 mainly based on Wackernagel and Vedic Variants. 3. Sanskrit Grammar, 213, h. 4. K. Beitr. 4, 208. 5. Goett. Abhand. 19. 159.

Page #74

A Prākrit Tendency in Rgveda
6. IX. 4, 40 7. Wartburg : Evolution et structure de la langue francaise, pp. 40-41. 8. Altindische Grammatik, I, 283. 9. Vedic Grammar, p. 65. A Vedic Grammar for Students, p. 33. 10. Ghatage, Introduction to Ardha-Māgadhi, 127; 11. Grammatik der Prākrit-Sprachen, 75, 76. Pāli Literatur und Sprache, 6. 12. Jacobi : Bhavisattakaha, 60; Alsdorf; BSOS, VIII, 329 foll. Ghatage : NIA., II. 50
foll. Upadhye, Dhūrtākhyāna. Intro. p. 51. 13. Bloch : L'Indo-Aryen, p. 91. 14. For the sense cf. Velankar : JBU., XIV., p. 8. 15. JBU. VI. p. 38. . 16. Delbruck, Altindische Syntax, 53. 17. Vedic · Variants, III, 164.
A Prākrit Tendency in Rgveda
IHQ. XXI. 1945
33

Page #75

An Unassimilated Group in Apabhramsa
In the course of the last few years the study of the Apabhramsa language has made rapid progress. A few years back, our knowledge of this important stage of the Indo-Aryan languages was confined to the rules of the Prākrit grammarians and the quotations found in Hemacandra and in the works of Alankāraśāstra. But, of late years a voluminous and ever increasing literature in Apabhramba is being brought to light and since the publication of Dr. Jacobi's edition of the Bhavisattakaha, a steady stream of Apabhramsa works is being given out to the public by various scholars. This increase of knowledge about the language and the fulness of material at our disposal have also brought in its wake a number of problems, which demand a close study for their solution.
One such problem is found in the phonology of Apabhramba. The Middle Indo-Aryan stage of linguistic growth is characteristically marked by the extensive working of the process of assimilation of groups of consonants, from the Old Indo-Aryan stage, where various groups of different consonants are preserved. From the earliest records of the Middle Indo-Aryan in the form of Asokan inscriptions to its very late phase as represented by the Apabhramsa literature, we find this law working with great consistency and thus it imparts to this stage a kind of uniformity of appearance. But there are a few exceptional cases, one of which is preserved in Apabhramsa, although earlier stages of the Middle Indo-Aryan do not show traces of it. This is all the more remarkable as the natural trend of development would make us expect it the least in such a late phase of the Middle Indo-Aryan.
Groups of consonants with r as one member and the vowel r have both suffered the normal changes in the Middle Indo-Aryan languages. Thus groups of the type consonant + r and r + consonant, are normally assimilated and cases of such assimilations can be found in the inscriptions of Asoka, the Pāli canon of the Buddhists, the Ardha-Māgadhi language of the

Page #76

An Unassimilated Group in Apabhramsa
67
Jairs and all the dramatic Prākrits. As regards the vowel r there is hardly any exception in the whole range of the Prākrit literature. But as regards the assimilation of groups of ļ, a few exceptions are met with in the early stages, particularly in the Aśokan dialect of the North-West and in a number of words in Pāli. But the Prākrit languages and the Prākrit grammarians always insist upon the assimilation of these groups, especially of r + consonant. The position of the Prākrit literature in this regard is quite unambiguous. The literary works always show assimilation and that also in a consistent manner. The only case without assimilation noted in this connection is that of the solitary word vandra (vrnda) which occurs in Jain Māhārastri a couple of times. Otherwise the rule of assimilation is observed with the greatest rigour.
The Prākrit grammarians, however, show that they know a few words at least where such groups could be tolerated. All of them agree in forming a special rule for the group dr, which may be optionally preserved in Prākrit. Thus Vararuci (III. 4) allows optional preservation of this group and his commentator Bhāmaha illustrates it with words like droho (drohah; doho), candro (candraḥ; cando), rudro (rudraḥ; ruddo). According to Hemacandra (VIII. 2. 79) vandra is always preserved without assimilation and he also allows optional retention of dr (VIII. 2. 80). His illustrative words further include bhadro (bhadrah; bhaddo), samudro (samudrah; samuddo) and draho (hradah; daho) which last also implies metathesis. He further remarks that the words like vodraha meaning 'a young man' are always without assimilation and are regarded as Desya-words. Mārkandeya has a similar rule for dr (III. 4). Moreover, according to him, the word vindrāvana keeps dr because the word is so enjoined and as such requires no operation of the general rule. All this would suggest that the grammatical tradition about the Prākrits preserved the memory of dr as an unassimilated conjunct to be kept optionally. The vast amount of the Prākrit literature, on the other hand, shows no trace of such forms.
For a long time it was thought that the position of Apabhramśa was nearly identical with that of the Prākrits, with this difference that a slightly greater emphasis was placed on the side of preservation of groups of consonant r. Here the grammarians taught that the Apabhramśa language has a general tendency of keeping groups with r and retaining the vowel s. But the literary works refused to substantiate it by their normal practice. Naturally, in the absence of confirmation from the literary form of the language the statements of the Prākrit grammarians were either looked upon with suspicion or otherwise explained in different ways.

Page #77

68
Amrita
Among the Prākrit grammarians, Vararuci shows no knowledge of Apabhramśa and the case of Canda remains doubtful. Among others we find a difference of treatment which is usually explained by grouping them into the eastern and the western schools. Whatever be the merits of such a division into schools, the important fact which immediately concerns us is that while Hemacandra and others like Trivikrama who closely follow him, treat of Apabhraíśa as one uniform language, the so-called eastern school of Purusottama, Kramadīśvara, Mārkandeya and Rāmatarkavāgīša knows different dialects of Apabhramsa and they treat of them separately.
Views of Hemacandra as regards the present problem can be seen from his three sūtras VIII. 4. 398, 399 and VIII. 4. 329. His sūtra 'svarānām svarāh prāyopabhramće' implies the preservation of the vowel ? as can be seen from his illustrations trnu (trna) sukrdu (sukrta). His other two rules lay down that r as the second member of a group may be optionally elided and that in some conjuncts an inorganic r is inserted without being found in the corresponding Sanskrit word. Thus priya may remain as priya and vyāsa may be met with in Apabhramsa in the form vrāsu. That these are not isolated cases to be noted by the grammarians as exceptional and out of the way forms going against the general trend of the Apabhraṁsa phonology and that these rules have a different import than similar ones in other Prakrits, becomes evident from other considerations. Besides these specific rules, Hemacandra gives other words and phrases which are peculiar to Apabhramsa and which evidently go to prove that the preservation of groups like consonant + r and the Sanskrit vowel ? was a feature of this language. Hemacandra tells us that in Apabhraíba yad becomes dhru and tad becomes trum (VIII. 4. 360); the genitive singular of yusmad is tudhra (VIII. 4. 372); the root brū- becomes bruva- (VIII. 4. 391) drs- becomes prassa- (VIII. 4. 393); the word prāyah is represented by prāu, prāiva and prāimva (VIII. 4. 414); dhruvam becomes dhruvu (VIII. 4. 419). All these rules presuppose that the Apabhramśa phonology can keep groups of the type consonant +r. For the retention of the vowel r we have his rule that the root grah- becomes grnha- (VIII. 4. 394). Other grammarians like Trivikrama add nothing new in their treatment (Cf. Trivikrama III. 3. 5, 6).
The eastern school of the Prākrit grammarians makes a distinction of dialects in their treatment of Apabhramśa and the present case has a vital bearing on this problem. Among writers who tell us of the Apabhramba dialects, we have the statement of Rudrata who, in his Kavyalankāra (2. 12) says that it has many varieties according to the country in which it is current.

Page #78

An Unassimilated Group in Apabhramsa
69
His commentator Namisādhu notes that there was another view current according to which Apabhramśa had three dialects called Upanāgara, Ābhīra and Grāmya. It should be noted that Namisādhu is more akin to Hemacandra in his grammatical treatment of the Prākrits, as far as we can judge him from the few remarks in his commentary on Rudraţa II. 12. Though Hemacandra makes no mention of Apabhramśa dialects in his Prākrit grammar, he himself names another variety called grāmyāpabhramsa in his Kavyānuśāsana, in which were composed kāvyas like Bhimakāvya in avaskandhaka metre, while works like abhimathana were written in the ordinary language. We do not meet this grāmya variety of Apabhrassa anywhere else than in Namisādhu and Hemacandra and both fail to give us any idea about it. Jacobi* has suggested that this grāmya variety of these two writers may be the same as the upanagara of the other grammarians.
Among these others Purusottama knows a number of Apabhramsa dialects (XVIII. 16-23), but his treatment pertains to the usual three viz. nāgara in detail and vrācada and upanāgara as far as they differ from it. In nagara, according to him, ļ and r after consonants are kept (XVII. 15). In vyāsa and other words we haver after the consonant of the first syllable (XVII. 25); the accusative of yad and tad is also jram and trus (XVII. 55); the genitive and locative is jatru, tatru (XVII. 56); drs- becomes prassa- (XVII. 79); kim is represented by kimpradi, kimpradu and kimpru (XVII. 25), all of which show that, as in case of Hemacandra, group of consonant + r was allowed in Apabhramsa, The retention of r is enjoined in the use of grnhafor grah-(XVII. 86). In most of these cases there is an obvious agreement with the rules of Hemacandra. Next, Purusottama adds rules for vrācada wherein he again states that (XVIII. 3) ? and groups of r are kept except in words like bhrtya etc. In place of vrs- comes varha- (XVIII. 12). Kramadīśvara also knows of this threefold division and his treatment agrees with that of Purusottama. In his case the most important point to be noted is that he takes as the characteristic feature of the vrācada dialect the preservation of conjunct of the type of r+consonant (IV. 67). He illustrates it with the word sarpi (sarpih). The comment further adds that in this dialect are used jrum, trum, which the author has also allowed for the nāgara dialect (V. 49). Jacobis had first misunderstood him in regarding that groups like consonant +r were characteristic of vrācada. Later on he corrected himself in making it consist in groups like r+consonant', but he has not changed all the conclusions that follow from his first position and in this manner, he is responsible in starting a fictitious distinction.

Page #79

Amrita
Mārkandeya knows a list of 27 dialects of Apabhramśa which he quotes at the beginning of his grammar (on 1. 4.), but he himself treats of the usual three. For nāgara, which he bases on both Māhārāstri and Saurasenī (XVII. 1), he tells us that y becomes r in words like vyāsa (XVII. 3). His list includes vrāsu (vyāsah), vrākrosu (vyākrośah), vrādi (vyādi) and (vrāgaranu (vyākaranam). it keeps ? and r in groups : indru (indrah), candru (candrah), mrgajudhu (mrgayūthah). But otherwise his treatment shows no such groups. It is all the more striking to note that for jram, tram, of others, he has jattim, tattim; for jatru, tatru he has jattha, tattha, for prassa- pumma (which according to Nitti-Dolci-? may be a mistaken reading) and for grnha- he has gunha., all of which are attempts at getting rid of r and r-groups. For vracada also he lays down that ļ and ļ as the second member of the conjunct may be kept except in case of words like bhrtya (XVIII. 4). He gives cases like prānaharu (prānaharah), krvānu (krpanah). For Purusottama's varha- he has simply vaha- (XVIII. 9), while he allows brodi like him (Purusottama XVII. 34 : broppi, broppinu). Thus there is nothing in his treatment to show, which was the distinguishing feature of vrācada from among groups like consonant +r like and r+consonant. There is a natural suspicion that the editor of Mārkandeya's grammar has made the changes which have removed all traces of groups of consonant+r, where the views of other grammarians may lead us to suppose their presence. Grierson says that in a number of cases his Mss. do not agree with the readings of the edition.
The rules of Rāmatarkavāgisa agree remarkably with those of Mārkandeya. He also knows a large number of Apabhramsa dialects but treats of the nāgara and vrācada and makes only a remark or two about the others. According to his treatment, nāgara uses vrāsu and vrādi for vyāsa and vyādi, it preserves words like priya and mrga (II. 4); the nominative and accusative singular of yad and tad is jadrum, tadrum, while the locative and genitive singular is jadru, tadru (II. 19, 20); drs- gives prassa- (to be read for pummaand probably in Mārkandeya as well) and grah- becomes grnha- (to be so corrected for gunha- and probably in Mārkandeya also) (II. 29, 30). For vrācada, he says, ļ and consonant+r are kept except in words like bhrtya (III. 2), brū- becomes bro- and vrs- becomes varha- (III. 4).
The general agreement among the grammarians of the eastern school is evident enough. We find them maintaining that the nāgara or the normal type of Apabhramsa allows (i) the vowel r, (ii) keeps groups of consonant+r and (iii) adds an inorganic r in a few words like vyāsa etc. This tendency is

Page #80

An Unassimilated Group in Apabhramsa
71
confirmed by the use of words containing r-groups or the vowel r as (iv) jraṁ, truṁ (or other variants all containing this group) (v) prassa- 'to see (with the doubtful exception of Mk.) (v) jatru, tatru (Mk. doubtful) (vi) grnha- 'to take'. To these Purusottama alone adds (vii) forms like kimpradi etc. The vracada dialect is distinguished from this normal type of Apabhramsa by the additional rules like (i) the preservation of r and groups of consonant+r except in words like bhrtya, (ii) vrs- becoming varha- (Mk. doubtful) and brū- becoming broppi etc. (Pu. doubtful). Kramadīśvara has the additional rule that this dialect keeps groups of r+consonant. For the present we are not concerned with other features which are said to distinguish these two dialects.
The position of the Prākrit grammarians may be summarised as follows. Hemacandra and others do not know any dialectal distinctions in Apabhraṁśa and allow for this language the preservation of ļ and groups like consonant+r, and the presence of an inorganic r in a few words. The grammarians of the eastern school, Purusottama Markandeya and Rāmatarkavägisa allow the first rule for both nāgara and vrācada, while the peculiar feature of vrācada alone is varha- coming in place of vrh-, because it is not possible to ascertain what is peculiar to vrācada in the forms broppi and broppinu, whether the base or the termination. For Kramadīśvara, the distinguishing feature of vrācada is the retention of r+consonant, of which varha- may be regarded as a specific illustration. Now in this fact Kramadīśvara is alone and the base varha- is too isolated to form clear distinction between the two dialects. Thus there is no authority for regarding groups of consonant+r as a feature which, should separate vrācada from nāgara and the group r+consonant has the authority of Kramadīśvara only with the solitary of sarpi.
In the light of these facts it will be now necessary to examine the Apabhramsa stanzas quoted by the early rhetoricians and particularly by Hemacandra in his grammar, in order to see how far they keep traces of conjuncts with r. The two stanzas of Rudrata illustrating bhāsāślesa of Sanskrit and Apabhramśa (IV. 15, 21) show forms like abhramada, prasara, krīdanti, prasaranti, pranaya, bhramarā, mitra and suvibrama all of which keep groups of plosive+r. The word durdhara, which occurs in the first verse, however, has suffered assimilation in becoming duddhara. In the verses quoted by Hemacandra, we find traces of r and groups of r in the following words : grnhai (336). prangani, dhrum, tram, bhrantri (360), krdantaho (370), tudhra

Page #81

72
Amrita
(372), broppinu, broppi (391) bruvaha (391) grnheppiņu vratu (394), priya (401), prayāvadī (404), prāu, prāiva, prāimva (414) dhruvu (418) drammu, dravakkau, drehi (422) trnai (422) pemmadrahi (424) grņheppiņu, dhrum, priu (438) antradi (445). In these very stanzas we find, side by side, forms like mahaddumu (336) mianku (401) bhaṁtadi (414) and pemma (preman) (424) which show that the normal assimilation and the vocalisation of r were current in the same dialect. Another fact which emerges by the scanning of these stanzas is that, as far as the metre allows us to determine, nowhere do these cases of groups make the preceding syllable long by position. The two apparent exceptions, antradi scanned as -U- and tudhra with the metrical scheme u-u are due to the anusvāra in the first and the double plosive in the second, which should be better written as tuddhra.
In his introduction to the Bhavisattakaha, Jacobi was misled by the supposition that groups of consonant+r were characteristic of the vrācada Apabhramśa and all these verses should be, therefore, regarded as being written in this dialect. This led him to enquire whether the language dealt with by Hemacandra is hemogeneous or contains traces of different dialects and he came to the conclusion that besides this vrācada, there are traces of two more dialects, one which may be called the Sauraseni Apabhramsa and the other showing softening of all intervocalic stops'. Later on in his introduction to the Sanaṁkumāracariu, he had to give up the fact, attributed to Kramadīśvara, that groups of consonant+r are characteristic of vrācada, but groups of r+consonant were meant by him as such. Naturally Jacobi 10 now considered that these verses which contain groups of consonant+r and the stanzas of Rudrata mark an older stage of Apabhramsa, which thus differs from the normal Apabhramśa of the other verses of 'Hemacandra's grammar. Dr. Upadhyell also thinks that features of the so-called Sauraseni basis and the retention of groups of r suggest dialectal differences in the Apabhramsa of Hemacandra. He differs from Jacobi in holding that the relation between the dialect showing such unassimilated groups and the normal one which assimilates them, is not chronological, the one being older than the other, but dialectal, the two coming from two different regions.
The idea of either a chronological or a dialectal difference based upon the treatment of groups of r was supported by most of the Apabhramsa literary works so far published. All of them uniformly followed the rule of assimilation and thus agreed with the practice of the majority of the verses quoted by Hemacandra, which left the few stanzas with the unassimilated

Page #82

An Unassimilated Group in Apabhramsa
73
groups either as a relic of an archaic stage of growth or remnants of a different dialect. But when Dr. Alsdorf published the Harivamsapurāna", a part of the Mahapurāna of Puspadanta, a writer of the 10th century and therefore decidedly older than Hemacandra, the problem assumed a new appearance. He pointed out that two of the three Mss. of that work do show the retention of r and groups of the type consonant+r, and he was able to collect some 11 words which have conjuncts of this nature and some 7 words showing the vowel r. Thereby he has naturally to reject the view of Jacobi that the retention of such conjunct can suggest an older stage of the language and his own conclusion amounts to the fact that Apabhramśa has kept an older phase of phonetic development, older than that of the Prākrits with respect to such groups.
With the publication of the whole of the Mahāpurana' in three volumes by Dr. Vaidya, it becomes again necessary to examine the problem anew in the light of all the material that can be collected from this work, which is the only major one which has kept traces of these unassimilated groups and the vowel ?. This will suppliment the collection made by Dr. Alsdorf and in part modify him, as the portion from which he has put together his 70 and odd cases, has been reedited with ampler material and which necessitates a revision of his cases, though to a slight extent.
As can be seen from the introductions of Dr. Vaidya to the different volumes of the work, the Ms. material for the whole work is not uniform. Because the problem of these sound changes depends greatly upon the way in which we may choose to handle the readings supplied by them, it will be necessary to note all the variants of the different Mss. as regards the preservation or removal of such sounds. I have been able to collect some 450 and more cases from the whole work, which show either the vowel r or r in conjuncts. I have also noted all the variants shown by the Mss. which either replace them with other vowels or change the groups by the process of assimilation and cases where the editor did not feel himself justified in adopting these sounds in the body of the text as the Mss. evidence was slender, being confined to one or two of inferior value. It is just possible that a few cases may have escaped my notice, but the collection is ample enough to evaluate it for linguistic purpose and a few more cases, if detected, are not likely to change materially our general conclusions based on the available material. It will not be out of place to point out that the editorial work has been so thorough and well done that the reader is never at a loss to make

Page #83

74
Amrita
out the exact significance of the readings recorded and except for a case or two where by oversight the names of the Mss. are missing in the critical apparatus, it is always possible to state the reading of a given Ms. as regards these changes.
To understand the implications of the readings of the different Mss. noted here by the indications used by the editor with minor changes, it will be necessary to summarise the important facts recorded by him, as far as they pertain to their nature and classification. Further details may be looked into in the introductions themselves. The first 37 sandhis of the Mahāpurāna are based upon five Mss. designated by the editor with the letters G. K. M. B. P., P., however, breaking off in the middle of sandhi 28. These are classified into two groups, G. K. giving the older version and M. B. P. a slightly later version. shown by the presence of a large number of introductory stanzas, various readings and a few additional passages. This division is valid for the present problem as well. The group G. K. uniformly presents ļ and r-groups, while M. B. P. change them to i and assimilated stops. In one case (9. 22. 9) the original reading of K. trya is modified into tiya by a later hand, which thus comes nearer to tiya of P. and tiha of M. B. Out of 157 cases of the first volume, there are only four cases where G. and K. differ among themselves : 22. 1. 13. G. vrata K. vraya; 34. 8. 6. G. prānapriu K. prānipriu; 36. 2. 6. G. priyahu K. pryahu; 10. 4. 11. G. dhruvu K. dhuvu, where only the last case shows a real difference as regards the retention of r-groups. Equally close is the agreement of the remaining three Mss. M. B. P. and there is only one
se 23. 11. 11. where P. agrees with G. K. as against M. B. P. prāņiprāņa M. pānipānu B. pānipāna. Otherwise they uniformly agree in getting rid of these sounds.
For sandhis 38-80, Dr. Vaidya has used three Mss. called K. A. P. which can be grouped into K. which keeps these sounds and A. P. which remove them. Out of 189 cases in all, only in 13 cases this grouping is disturbed : 54. 4. 7. is doubtful as no readings are recorded (uddiyaprānai) and this will be the only case where P. may be said to preserve the r-group which, however, is naturally suspect. In 60. 16. 4. A. K. nrvahu P. nivahu; 60. 20. 8. K. A. mrga P. miga; 61. 19. 8. K. A. mrgaloyaņāi P. miga; 61. 19. 14. K. A. nrvai P. ņivai; 61. 19. 15. K. A. mrganayaņa P. miga; 61. 20. 12. K. nrvu, A. nru P. niu; 62. 4. 14. K. A. mrganettahi P. miga; 74. 10. 6. K. A. nrvabalai P. niva; 74. 12. 12. K. A. mrgena P. migena; 80. 5. 11. K. A. nrvai P. nivai, A. agrees with K. In 71. 7. 2. K. P. migi A. mrgi; 73. 18. 15. K. P. migamuddai A. mrga, A. is alone in keeping r.

Page #84

An Unassimilated Group in Apabhramsa
The sandhis 81-92 are based on the Mss. K. A. P. (=Alsdorf B.) B. (=Alsdorf A) and S. (=Alsdorf C). Here also the relation between the first three K. A. P. remains the same. Out of 72 cases K. keeps these sounds for 56 times, P. in none and A. agrees with K. in 10 cases in having them : K. A. nrva P. niva 81. 12. 1. 84. 2. 3.; K. A. prangani P. pangane 90. 4. 13; 92. 9. 8. K. priu A. prya P. piya 81. 12. 2. 82. 1. 11; 91. 13. 11; K. A. mrgāyanahu P. migāyaṇaho 82. 8. 10; K. A. vrndu P. vindu 81. 18. 3. K. A. samprāiu P. sampaiu 86. 1. 25. In one case 81. 1. 11. K. niva A. nrva, it has r against K. In 88. 24. 13. the reading of P. remains doubtful, and with the express statement of Alsdorf we may regard it as not having the sound r in mrga. B. has 24 cases of retaining these sounds and S. has 72 cases of ļ and r-groups, thus 16 times more than K. : S. priya K. tiya 89. 13. 5. S. trya K. tiya 91. 6. 6. S. nrva K. niva 81. 1. 11; 88. 2. 15; 88. 9. 12; 89. 18. 8; 90. 2. 6; 90. 12. 5; 91. 9. 10; 92. 3. 3; 90. 6. 15; S. priya K. piya; S. mrga K. miga 88. 23. 15; 88. 23. 17; 91. 2. 4; 92. 1. 16; S. vraya K. vaya 92. 14. 14. S. srya K. siyā 90. 3. 7. and one case in which K. has the sound as against S. Thus S. agrees almost always with K. in having them and shows them in addition in many cases. B.'s agreement with any other Ms. is not obvious.
For the last portion of sandhis 93-102 the Mss. are again K. A. P. as in volume two and their interrelation remains the same. K. has 17 cases and A. has 15 cases and agrees with K. in 101. 3. 18. K. A. P. prāna which is doubtful. Otherwise P. has no case of such sounds. This is all the more remarkable when we find K. and A. agreeing in volume two in 10 cases and in Harivaņģa in 10 more cases.
Our results about the Mss. may be thus summarised :
K. no date, available for the whole text, has 414 cases with and 34 cases without these sounds.
G. date 1518 A.D. for ādipurāņa has 155 cases with and 2 without the sounds.
M. date 1826 A.D. for adipurāna, no case with the sounds. Ba. date 1602 A.D. for ādi. no case. Pa. no date, for 28 sandhis, no case.
A. date 1558 A.D. for Uttarapurāna, cases 37 with and 254 without the sounds.

Page #85

76
Amrita
Pu. date 1573 A.D. for Uttara cases 3(?) with and 288 without the sounds.
Bu. date 1584 A.D. for Harivamsa, cases 28 with and 46 without the sounds.
S. younger than Bu. for Harivamsa, cases 71 with and 1 without the sounds.
From this list the interrelation of the dates of the Mss. and their tendency to keep the sounds of and r-groups becomes evident. The case of S. must be judged by the fact that it is itself a recent copy of an older Ms. of the Senagana. Moreover no very strict correspondence can be expected in such matters. But the conclusion should become apparent that these forms go back to the oldest period of the next tradition and cannot be attributed to later importation by the copyists. That the original text must have contained more forms than what we now possess is suggested by the fact that even the best and oldest of our Mss. do not give them in cases where others have preserved them. A direct proof of the tendency to remove such forms from the text is supplied by K. in which we find how a few forms of these sounds are changed into those without them.
More important for deciding the phonological nature of Apabhramsa is the fact that in spite of so large a number of such cases the range of the words in which they occur is much limited. The following are the words with the number of cases occurring in Mahāpurāņa distributed according to the Mss. where they are found. M. Ba. Pa. and Pu. may be left out of consideration, as they show either no cases or one or two of doubtful value. (1) Words with r-groups :
G. A. bu. S. anghriņi 1 ānāprāna 1 indriya
1 kriya 1
1 : dratti 1 dravakkiyau 1 draha 4 druma 1
K.

Page #86

An Unassimilated Group in Apabhramsa
16 6
i 2
2 4
6
24
1
4
5
- dhruvu 25 prangana 26 pravari 1 prāe 1 prāņa 49 prānaya 7 prāņi 10 prāva 2 prāsiya 2 priu 15 priyadatta 1
6
4
3
4
vraniu 1
2 1
8 3
3
um
1
A.
Bu.
S.
vrana 1 vraya 38 · 11 samprāiya 25 14 samprāviya 9 5 samprāsanu 1 (2) Words with r : K. , G. krya 1 trya 196 6 nr 7
2 nrva 53 5 nrvai 5 parivrddhi 1
10 pryadattā 5 mrga 57
25 mrganāhi 1
1 1 10
3 3 3
21
3
pru 14
1
6
3
9

Page #87

Amrita
mrgavai 1 mrgāyanu 2 1 1 mrgāvai 5 mrgi 1 vrnda 2
1 1 1 srya 12 - 3 1 3 (3) Words with inorganic r :
G. A. Bu. nipriha 1 prāě(?) 1 prāyanti 1 bruha 1
S.
vrahiu
1
vrāsu 1
Traces of such forms are rare in other Apabhramsa works. But it will be incorrect to say that they are confined to the Mahāpurāņa only. In the Kathākośa of Śrīcandra a few forms with r (particularly nrva) are noted. This author is placed in the 11th century. The Apabhramśa verses in Hemacandra's Chandonuśāsana also supply us with a few words of this type : krva (6. 116), cakru (7. 3) trutti (6. 32), drahi (6. 60), druo (7. 37) dhruvu (6. 121; 6. 129; 7. 57) priya (6. 18; 6. 25; 6. 38; 6. 46; 6. 82; 6. 127) prānahara (6. 118) vandri (6. 34). Jacobi has already pointed out that these Apabhraṁsa verses are the compositions of Hemacandra himself and not quotations from earlier works as is the case with his grammar. This is not wholly true. A few of them are quotations and recur in his grammar, as the verse cüdullao cunni hoisai (Ch. 6. 119 = Gr. 4. 395). Moreover, to ascertain the range of such words they are of importance, as it is scarcely possible that Hemacandra would have kept them even in his own compositions if he had not found them in current literature. In this sense they differ from his stanzas in the Kumārapālacarita where they are expressly written for illustrating the rules of his grammar and as such may distort the facts of the natural language by overemphasis. In this metrical treatise he has no such need to change the phonology of the language, particularly when these groups are metrically harmless, i.e. make no difference by their position.

Page #88

An Unassimilated Group in Apabhramsa
" As said above, though the verses of the Kumārapālacarita may not be taken as good evidence for linguistic purpose, it is nevertheless important to see how far they may help us in indicating the scope of Hemacandra's rule vādho ro luk, for even a poet writing ad hoc for the purpose of illustrating his grammar cannot write a non-existent form, especially in words not meant for illustration and when there is no metrical or other necessity for their choice. Thus besides the words expressedly taught in his sūtras, we have, in addition, krva (8. 66) krvā (8. 82) krvālū (8. 82) drahi (8. 68) grnhia (8. 45) nrva (8. 82; 8. 83) prāvai (8. 58) prāmvei (8. 69). That these stanzas do show the same phonological features as the quotations in his grammar can be seen from an interesting coincidence, which cannot be reasonably said to be intentional. Thus all the above words do not make long the preceding syllable in spite of the conjunct of r. But we find the word tudhra used with the metrical scheme-u which is the same as found in the line tudhra anuttara khanti of his grammar (4. 372).
These facts about the phonology of Apabhramśa have raised a number of problems, many of which Dr. Alsdorf has dealt with, with reference to the cases found in the Harivamsa and has come to some tentative conclusions. With this fuller material, they can be now in part confirmed and in modified. Of his forms one pittrya must be rejected as being not substantiated by the Mss. evidence and his reading of C (=S of Dr. Vaidya) in 84. 10. 11 vrahena (vadhena) is also doubtful as it is not recorded by the later editor.
As regards the question whether all such forms can be the result of Sanskrit influence on the scribes, Dr. Alsdorf has rightly pointed out that the number of such forms, which has now become quite considerable, the agreement between the illustrative words of the grammarians and the words found in the Mahāpurāna, the agreement of the different unrelated Mss. and the absence of any other traces of Sanskritisms on the part of the scribes, all go to show that they are not due to Sanskrit influence. But he has added two more considerations of a different nature to prove the same fact. The writing of srya for śrī and trya for stri cannot be due to the influence of Sanskrit, in which case we should expect writings like sriya and triya. Secondly there is uniformity in the writing of nrva (with dental.n) on the one hand and niva (with cerebral n) on the other. In case of a simple Sanskritism we should have expected a p as well in this word. But the choice of n with the vowel ļ may be due to the form current in Sanskrit.
If we accept the fact that the written r in these words in Apabhramsa

Page #89

80
Amrita
really stands for the sound ri, which Alsdorf himself suggests, it is but natural to suppose that the choice of the form with either ? or conjunct with r was mostly decided by the nature of the Sanskrit word, as, on the whole, the distinction between the two modes of writing agrees with Sanskrit orthography. This lack of agreement in case trya and soya may be attributed to the obscurity of their relation with the corresponding Sanskrit words, which was not quite apparent. Moreover we have the express statement of the grammarians that both r and groups with r remain as in Sanskrit (prakrtyā). The second point, probably, is to be explained differently. The contrast between nrva and niva cannot be pressed too far, as even Mss. and books which prefer to write niva continue to write nrva. That the n of nova was caused by the r sound is more probable, as the two cerebral sounds cannot be easily pronounced together and we find a similar contrast as early as in Asokan inscriptions. There we see the eastern dialect with cerebralisation and assimilation of r-groups contrasted with the western dialect with r-groups and absence of cerebralisation.
In spite of statemant of the grammarians and the nearly concordant usage of the Mss. Alsdorf is right in suggesting that all these words with r and r-conjuncts represent the same real pronunciation., His argument to prove this, however, is not quite intelligible. The writing of srya and trya to represent śrī and strī can only prove that their relation with the Sanskrit words was not evident for the writer, for otherwise, with the same sound and the Sanskrit mode of writing to guide him as in all other cases, we should really expect triya and sriya. That priya is actually written prya both as adjective ('dear') and noun (=priyā 'wife') and in proper name pryadattā is clear from the above illustrations. These cases, however, were lacking in his portion of the text. Fortunately we have some more evidence to show that both the modes of writings represented the same sound and that it was the sound ri. In the first place, we have now the same word, now written with r now with ri. Thus in 69. 15. 7 there occurs the word paramakrya (paramakriyā), while in 34. 1. 6. we find sanāhakriya (sanāthakriyā). We have both priu, priyadattā and pru and pryadattā in the various readings. In 36. 2. 5. while G. writes priyahu K. prefers pryahu. All the Mss. which assimilate the r-groups always write i in case of words with r which means that they had before they forms with ri in groups. But probably the best evidence for this sound of r is to be found in the rhymes of Apabhraíśa poetry. It is well known that final rhymes in Apabhramśa occupy a prominent place and the whole of the Mahāpurāna shows them. The one peculiarity of

Page #90

An Unassimilated Group in Apabhramśa
these rhymes in Apabhramśa is that they pertain not only to the final syllable, which must be identical, but also to the vowel preceding it, to make it a good rhyme. Thus if a word of two syllables occurs at the end of a line so as to rhyme with another we can naturally suppose that their vowel elements must be identical. In Mahāpurāna we find the following cases of end-rhymes :
*69. 15. 7.
paramakrya
paccakkhasrya
34. 1. 6.
dehi sṛya
31. 8. 8.
prāṇapru
*45. 12. 4.
tāsu prya
*45. 3. 8.
gampi thiu
82. 1. 11.
pañcamahu piya
Three of these cases marked may not prove anything. But the other cases clearly indicate that both the modes of writing r and ri represented the same sound, which made good rhymes with the first syllable of piu (= pitā) and piya (= priya), and this sound can only be ri in these circumstances.
saṇāha kriya
kuberapiu nai srya
prāṇapriu
paccakkha srya
81
As regards the forms of inorganic r as taught by Hemacandra VIII 4. 399. Alsdorf was not able to get any in his part of the text. Now we have a few cases in the following places: 49. 4. 2. nipriha (T. explains niḥspṛha), 42. 3. 5. prayanti (pādānte), 16. 11. 7. bruhasamgamu (budhasangama), 99. 3. 5. vrahiu (vadhita), 98. 8. 6. vrāsu (vyāsa) of which last is the illustration given by Hemacandra and Märkandeya. A doubtful case is 25. 5. 7. prae samprāviu prāṇanihanu vani vāhe viddhau harinamihuņu which can give a better meaning by taking prae as an adjective of vähe in the sense of 'sinful' as the reading of M. B. P. pave suggests, though the explanation of T. prāyaḥ is just possible.
These cases amply justify the rule of the grammarians which is cautiously worded and would inspire confidence in their statements, when we see them confirmed even in obscure cases like the present. The real explanation of such forms is, however, difficult to find. Following Märkandeya (XVII. 3. vyāsāder yasya rephaḥ syat) Alsdorf thinks that this r may have developed out of y. Tram may be from tyad and not from tad; bhantri may presuppose bhrantyi. This explanation cannot hold good in our cases. If vrahiu is the result of contamination between vadhitaḥ and vranitah (which actually occurs in 99. 3. 5) others remain without explanation. Here, I think, Hemacandra is right in calling the r extraneous (abhūta non-existent) than

Page #91

82
Amrita
Mārkandeya who takes it as developing out of y in a conjunct. It is quite probable that in such stray cases Apabhramśa has kept a genuine popular element, the mode of pronouncing some words with an intruding r current among the illiterate people. In Marāthi we actually find village folk pronouncing the word dusta as drusta and a similar habit may be at the basis of Apabhramśa.
Finally we come to the most important problem about these forms, their relation to other forms of these words which show either the full vocalisation of r or the normal assimilation of r-groups. These latter predominate even in Mahāpurāna and are exclusively found in the majority of Apabhramśa works. They are normally so represented in the Prākrits as well. Two explanations are possible and both are put forth by scholars. The forms. with r-groups may be regarded as indicating a dialectal and regional feature as is done by Dr. Upadhye or they may be thought to represent an older phase of growth, thus suggesting the earlier and archaic form of the language as compared to the other forms current in the younger form of the same language. This is the view of Dr. Jacobi. Dr. Alsdorf has further suggested that these forms with r and r-groups also indicate a phase of growth which is even earlier than the normal forms of Prākrits and in this respect at least, the latest stage of MIA. is more archaic than the earlier dialects.
With the above evidence before us the following facts become clear and help us in deciding which of these suggestions is most acceptable. Firstly we have now forms which preserve these sounds side by side with others which assimilate the groups in the same literary work. This will certainly go to exclude the possibility of regarding them as dialectal and regional in origin. It is possible to imagine that a grammarian like Hemacandra may have formulated his role of the language by taking into consideration all available literature, in which finer dialectal differences were neglected, and thus laying down rules applicable to different dialects side by side. But it is quite inconceivable that a poet writing a literary composition remaining in one place, can make use of forms from two different dialects, without apparent reason. That the same dialect can preserve both types of forms side by side should become evident from the Girnar version of Asokan inscriptions and the literary Pāli.
The second fact about these forms to be noted is that in nearly all cases the conjuncts are of a different nature than those of Sanskrit and even of Middle Indo-Aryan, where they are preserved. This difference becomes

Page #92

An Unassimilated Group in Apabhramsa
plain when we consider the metrical value of these words. In the majority of cases, all these groups in Apabhramśa do not make the preceding syllable long by position as do Sanskrit and Prākrit conjuncts. The only exceptions to this rule are, in Mahāpurāna : 20. 23. 7. G. K. rattadrahi M. B. P. rattaddahi; 16. 21. 2. G. K. niprānau M. B. P. nippānau; 48. 19. 4. K. nipraniya A. P. nippāniya. In the last two cases the prefix nis is responsible for the length of the first syllable. This metrical nature of the conjuncts with r is also apparent from the fact that in all cases one or the other Ms. does not show it and yet their readings are metrically faultless. As against the tendency of the Prākrits, they can remain even after long vowels without affecting them : 10. 12. 7. ānāprana, 35. 11. 20. mahadrahi 42. 2. 7. prangane prängane 47. 8. 15. cuo pränaindo and so on. This rule is applicable to the forms found in Hemacandra's grammar and Chandonusāsana as well. But the two stanzas written by Rudrata do not conform to it. In fact, the conjuncts there are of a different type as far as their metrical value is concerned.
Thirdly we find that all these Apabhramsa words preserve the group only initially. This is, in fact, a result of the preceding peculiarity. Here also Rudrata's stanzas differ in allowing words like mitra, vibhrama etc. where the conjunct occurs medially. These two facts would naturally lead one to enquire whether in all such cases we are dealing with real conjuncts or a mere graphical habit of writing them while the actual pronunciation had no groups. Such a supposition can explain their metrical value and their presence initially where alone can such conjuncts be written without violating the metre. But here, I think, there is much more than a difference of orthography. In fact, the Old Indo-Aryan and the Middle Indo-Aryan conjuncts written with r after consonants represent a sound in which the plosive is long or double while the conjuncts in Apabhramśa are pronounced with a single plosive as the first member, which satisfactorily explains the metrical value of the preceding syllable in the two cases. Thus mitra really represents mit/tra while the Apabhramsa form like anaprāna is anā/prāna.
With these facts ascertained, it is obvious that we cannot put side by side the Apabhramsa-Sanskrit stanzas of Rudrata and the genuine conjuncts of Apabhramsa works and put them together as marking an earlier stage of growth as Jacobi does. In fact, the Apabhramba verses of Rudrata are in a way artificial as they combine the Sanskrit and Apabhramśa conjuncts in spite of their difference in sound, with the help of a defective orthography.
The other argument of Dr. Jacobi is that Haribhadra in his

Page #93

84
Amrita
Nemināhacariu writes an Apabhraíśa which uniformly assimilates thergroups, while his contemporary Hemacandra describes in his grammar a language which preserves them and it is natural to suppose that the grammarian is describing a slightly older phase of the language which has become classical by his time, probably following an older grammatical tradition. As rightly pointed out by Alsdorf this cannot mean that works which do not show such groups are younger than others having them, for the text tradition may have removed such group which was easy in view of the fact that such a change could not disturb the metre. Moreover, we now find a poet using side by side forms with and without these groups. The only reasonable conclusion to which we can arrive is to regard them as simply alternative forms current along with the others or at most, as. archaisms continued in the literary language far beyond their natural date. . :)
If this is what can be ascertained about the conjuncts of the normal type of Apabhramsa, it will be possible to reconsider the question about the dialectal difference between the nāgara and vrācada dialects. Jacobi rejected the presence or absence of groups with consonants +r as marking the difference between the two in favour of regarding vrācada as keeping r+consonants, while nāgara as having consonants tor, on the authority of Kramadīśvara. But expect for the illustration sarpi given by his commentator and two forms in Hemacandra's Kumārapālacarita, narmada and sarmada (8. 80) we have no trace of such forms and thus we are unable to ascertain what value should be attached to Kramadīśvara's statement. On the contrary, all the other grammarians of the eastern school agree in attributing the preservation of the same type of conjuncts, consonant+r to both nāgara and vrăcada, and the mode of their writing implies a kind of distinction between the two. It may be that they implied the rule as having a limited application in nāgara while it was operative everywhere in vrācada. The exceptions noted by Mārkandeya to his rule in vrācada with the gana bhrtyādi only pertain to the vowel and not to r-groups, as can be seen from his list, bhrtya-bhica, nrtya nicca, krtya = kicca and krtyā = kiccā. If this is possible, it follows that vracada, as distinguished from nāgara, must keep these r- conjuncts both initially and in the middle as well and in that case, these conjuncts can only be of the type in which the first consonant is geminated. Then the verses of Rudrata can continue to represent the vracada dialect and we may suggest that the two distinct pronunciations of the r-groups distinguished vrācada from nāgara with the further consequence that while in the medial position they can continue in the first, only the initial groups can be preserved in the

Page #94

An Unassimilated Group in Apabhramsa
85
nāgara.
In the light of these distinctions it is not possible to accept without the reservation the conclusion of Alsdorf that the Apabhraíśa forms represent an earlier stage of growth than the Prākrits. In some cases at least it remains doubtful whether the Prākrit forms went through the stage represented by Apabhramśa, especially in words like nrva, nrvai, etc. In others, the Prākrit and Apabhramśa forms represent divergent lines of growth. Thus the Apabhramsa form mahādraha cannot form an earlier stage of Pkt. mahaddaha, obviously because the geminated consonant in Prākrit cannot be a further development of the Apabhraíśa conjunct dr and the same holds good in case of Ap. āņāprāna Pkt. anappāņa. Because the majority of the Apabhraíśa words keep these groups initially and in such cases the Prākrit represent only a single consonant at the beginning of a word, the relation of the two is not amenable to exact proof. So the conclusion of our enquiry can only be negative. The light conjuncts of the normal Apabhramśa cannot be taken as a dialectal feature and may or may not represent an earlier stage of growth with reference to the assimilated forms of Apabhramśa itself and the Prā If their relation to the other Apabhramsa forms may be that of archaic survival among the current usage and in this limited sense they may be earlier, their priority to the Prākrit forms remains doubtful. It is however possible that r-groups of the heavy type were a dialectal feature of vrācada Apabhramśa.
Annotations : 1. Cf. Pischel : Grammatic der Präkrit-Sprachen, p. 186. 2. Cf. Grierson : The Eastern School of Prakrit Grammarians (Mookerjee Vol. III, 2.
p. 119 foll).. 3. Nitti-Dolci : Les Grammairiens Prakrits, pp. 165 foll. 4. Introduction to Bhavisattakaha, München. 1918, p. *76. 5. Ibid. p. *71.
6. Jacobi; Introduction to Sanamkumāracariu, München, 1921, p. XX. 7. Le Prākstānnuśāsana de Purusottama, Paris, 1938, p. 109.
8. The Apabhraṁsa Stabakas of Rāmaśarman IA. (reprint). 9. p. *72. 10. p. XXI.

Page #95

86
Amrita
11. Introduction to Paramātmaprakāśa, Bombay, 1937, p. 47.
12. Harivaṁsapurāna : Alsdorf, Hamburg. 1936, pp. 137-140.
13. Mahāpuranam : Dr. Vaidya, Vol. I, Bombay, 1937; Vol. II, 1940; Vol. III, 1941,
MJGM.
An Unassimilated Group in Apabhramśa,
12th AIOC. 1944
38

Page #96

Group 2 : Prākrit Literature

Page #97


Page #98

The Title Daśavaikālika Sūtra
Even from the earliest times, it appears, there was no agreement among the traditional writers about the form and the interpretation of the name of the work usually known as the Daśavaikālika Sūtra. Like many other works of the Ardha-Māgadhi canon there is no occasion to give the title name either in the introductory or concluding portions of the text. References in other works and the comments upon it are also not unanimous.
In the Nandi Sūtra', where a list of the works forming the canon is given, we find the name of the work in the form Dasaveyāliya, and stands at the beginning of the Ukkāliya section of the Suyanāna. Bhadrabāhu, the oldest commentator of this text, used in his Nijjutti? the form Dasakāliya six times (w. I, 7; twice, 12, 14, 25) and the form Dasaveyāliya twice (v. 6, 397). Of these two forms of the title, he decidedly favours the first as the name of the work and he used the second only incidentally. This will be clear from the fact that in all the three places (w. 7, 12, 15) where an attempt is made to explain the name the form is invariably Dasakāliya and not Dasaveyāliya. Jinadāsamahattara in his Cūrņi? on the text, however, and following him Haribhadra in his Sanskrit Tikā, usually use the form Dasaveyāliya, even though the other form is found in their works incidentally (Cūrni, p. 4, Hari. p. 1). They have not seen any discrepancy between the two names and explain the title always in the form of Dasaveyāliya.
No material help can be derived from the names of the other books of the canon. Even though the first word of the name, Dasa occurs in many names of the canonical works, and according to W. Schubring“ all these works form a group by themselves, the meaning of the word is clear and it always refers to the number of the chapters found in the particular book. It is only in case of the Vanhidasāo that we find a disagreement between the title and the number of the chapters. In the present case also, there is no objection to take the word to mean ten chapters, because the additional two sections are

Page #99

Amrita
expressly called Cūlikās and are clearly intended to be later additions. The second element of the name Veyāliya occurs in only one other text of the canon, viz., the Tandulaveyaliya a book included among the Painnas but also in the Ukkāliya section along with the Dasaveyaliya. But there the name means a calculation (veyāliya=vicāra) of the number of rice grains, and cannot have anything to do with the second element of the name of Dasaveyāliya.
A correct interpretation of the name is equally hard to find out. In this connection it is difficult to decide what meaning of the title was intended by the author of the Nijjutti, in spite of the three different attempts made by him to explain the name. In the introductory portion of the Nijjutti we find the analysis of the title as dasa and kāla both the words receiving further elucidation. To explain the import of the word Dasa the author was led to explain the word ekka. On this verse the Cūrņi has preserved an interesting passage which runs :
एत्थ कतरेण इक्कगेण अहिगारो । भद्दियायरिआवदेसेण जम्हा दस एए पज्जाय-अज्झयणा संगहेक्कएण सङ्गहिया तम्हा सङ्गहेक्कएण एत्थ अहिगारो । दत्तिलायरिओवएसेणम् जम्हा सुयनाणं खओवसमिए भावे वहइ (? वट्टइ) तम्हा भावेक्कएणम् । दोनि वि एए आदेसा अविरुद्धा । भावेक्क एणं 3 II (p. 4).
In spite of the assurance of the author and his own inclination towards the second view, it is undoubtedly the first which is historically correct and offers one more proof to say that the present work is a collection. After the explanation of the number one' the Nijjutti goes to explain the number 'ten', and after that it states :
दव्वे अद्ध अहाउय उवक्कमे देस-काल-काले य ।
तह य पमाणे वगण्णे भावे पगयं तु भावेणम् ॥ Here it clearly states nine different senses of the word kāla and points out that in the present context the bhāva sense is applicable. What is meant by the bhāvakāla we are left to guess. The explanation of Haribhadra that it refers to ksāyika and other bhāvas of the soul is of no great use and like the above one of the Cūrni on ekka is a convenient way for the commentators to pass over the difficulty. In fact Haribhadra has noticed the discrepancy between the words of the Nijjutti and his own explanation, and so he remarks
यदुक्तं । पगयं तु भावेणं ति । तत् कथं न विरुद्ध्यते इति । उच्यते, क्षायोपशमिकभावकाले शय्यम्भवेन नियूदं प्रमाणकाले च उक्तलक्षणे इति अविरोधः । अथवा प्रमाणकालोऽपि भावकाल एव ।
The remarks are sufficient to point out his inability to explain the

Page #100

The Title Daśavaikālika Sutra
91
intention of the author. The Cūrni offers no more light on the point.
That Bhadrabāhu really meant to explain by the present remark the title of the work and was not merely speculating about the bhāva meaning of the word can be proved from many other places in the Nijjutti. While
lining the word mahugara he remarks; ihayam puna ahigaro vihāyagamanehi bhamarehim/117. While explaining the title of the third chapter he says paikhuddaena pagayam/185. So also we find him remarking: ettham puna ahigāro nikāyakāena hoi suttammi/289 ettham davvesaņāe ahigāro/304 niddesapasamisāe ahigaro ettha ajjhayane/316 and in all these cases he is perfectly right.
There are two more verses in the Nijjutti which offer another interpretation of the name. Verse 12 runs :
mifc437494397 quoti, fampreferito 31 निज्जूदं किर सेज्जम्भवेण दशकालियं तेण ॥
This suggests that the work was called Dasakāliya because it was composed or culled out by Sejjambhava when the period of time called Pauruşī was over. Verse 15 of the same text runs :
मणगं पडुच्च सेज्जम्भवेण निज्जूहिया दसज्झयणा । वेयालियाए ठविया तम्हा दसकालियं नाम ||
Here also a nearly identical explanation is offered. Because the ten chapters which he culled out were placed at the time of Vikāla the work was called Dasakāliya. Besides the apparent disagreement between the two words of explanation vigayaporisi and veyāliyā with the title dasakāliya, there are two interpretations of the word vikāla possible. It may mean the time of the evening, as the commentators take it or it may mean an improper time, as is suggested by some modern scholars. The choice between the two for Bhadrabāhu's own interpretation cannot be decided on the mere authority of the commentators.
The Curni throws very little additional light on the question of interpreting the title. It remarks :
विगतः कालो विकालः । अथवा विकालः कालः असकलः खण्डश्चेत्यनर्थान्तरम् विकालवेलायां परिसमाप्तं वैकालिकम् । अथवा विकाले पठ्यत इति वैकालिकम् । अथवा दशैतानि अध्ययनानि व्यवगते faa rifa ya hifciant (p. 5)
Here he accepts the usual interpretation but makes a new suggestion in the form that because it is studied at an improper (or evening time) it is

Page #101

92
Amrita
called Vaikālika. Haribhadra only accepts the usually accepted explanation that it was written in the evening and tries to explain the text of the Nijjutti accordingly. After him both the forms of the title and its interpretation were settled once for all and all later writers follow him closely?.
Now all these explanations except the first obscure one of the Nijjutti are based upon the traditional story about the composition of the work. But even taking the story as it is, it is difficult to see how such a small detail, that the work was written in the evening, should give the title to it. This is much more striking in face of the fact that there was nothing abnormal about the time itself. It is true that the Uttarāddhyayana prescribes the first and the last watch of the day and night for study (XXVI. 12, 18) and the Cūrņi remarks that the work was composed in the third Porisi, a little earlier. But Sejjambhava could have well waited a little more, and it would have made little difference, as Managa was to live six months more. The suggestion of the Cūrni is more to the point. The story tells us that Managa was to live only six months and it was not possible for him to complete the study of the scriptures in the usual method which extended over a very long period. We know that the Pūrvas can be studied by a monk in the 19th year of his Paryāya and it was impossible for Managa to study them. This naturally led Sejjambhava to have selections from these works for the benefit of his shortlived son and he taught them to him irrespective of the time which are prescribed for the study of these works. As such the work would well receive the name Vaikālika. In this very sense we can understand the words of the Nijjutti veyāliyāe thaviya' (15). In fact, all these extractions from the Pūrva books were intended to bring the important contents of the works within the province of study of monks who cannot wait for the regular period of time prescribed for their study. In this connection one remark in the Prakrit story as preserved in the Cūrņi is instructive. For pointing out the motive of Sejjambhava to cull out these texts from the Pūrvas he remarks:
तं चोद्दसपुव्वी कहिं पि कारणे समुप्पन्ने निज्जूहइ । दसपुव्वी पुण अपच्छिमो अवस्समेव fooles 11
So, this rule has same value when we consider that the earlier monks were not allowed to violate the rules of study unless some specific cause was available. But when the knowledge of the Pūrvas began to grow scarce it was allowed for the few who knew them to make extractions from them with the intention of preserving whatever little they can. It can be easily seen that these statements confirm the view of Charpentiero that the Pūrvas were lost

Page #102

The Title Daśavaikälika Sutra
on account of their study being placed late in the regular plan of mastering the scriptures. I cannot see any strong reason to suppose that the work got the name Vaikālika because it was culled out against the rules of doing so.
All these explanations, however, accept the name to be Daśavaikālika and see no contradiction between it and the other form Daśakālika. But seen above, the older name appears to be Daśakālika and not Daśavaikälika. The story itself, probably gave greater currency to the second form of the title.
To explain the title Daśakāliya we must try to know the meaning of the word kāliya. Two meanings of this word are of importance to us. There is a method of dividing the canon into four Anuyogas and it is common to both the sects of the Jaina community and as such must be very old. The very first of these Anuyogas is called the caranakaraṇānuyoga and the Daśavaikālika Cūrņi remarks : tattha caranakaranāņāogo ņāma kāliyasuyam p. 2. From this it appears that the canonical works dealing with carana or rules of good conduct and karana or rules of begging food were called by the name Kālika Śruta. This description passes very well with the contents of the Daśavaikālika. . We have further the authority of the Nijjutti to group the present work in this Anuyoga, because it remarks :
अपुहत्तपुहत्ताइं निद्दिसिउं, एत्थ होइ अहिगारो ।
चरणकरणाणुयोगेण तस्स दारा इमे होन्ति ॥ There is, however, another meaning of the word kālika in connection with the texts of the canon. In the Nandi we get the older classification of the canon into those Angas and Angabāhiras, the second of which is divided into Avassaya and Āvassayavairitta. The last is divided into Kāliya and Ukkāliya. The explanations of the two terms is given by Malayagiri', which runs :
तत्र यद्दिवसनिशाप्रथमपश्चिमपौरुषीद्वय एव पठ्यते तत्कालिकम् ।
यत् पुनः कालवेलावर्जम् पठ्यते तत् उत्कालिकम् ॥ and quotes a passage from the Cūrni to the same effect. This second meaning also harmonises with the one suggested above. But this meaning of the word kālika cannot be seen in the title because the text is included in the Utkālika section and stands first in that list and not in the Kālika one, which we should naturally expect if the word has this meaning in the title.
From the facts stated above we can conclude something about the real state of facts at different times in the history of the text, even though it must

Page #103

94
Amrita
be admitted that it is something of a convenient supposition to explain the conflicting facts. To my mind, originally the work was called Daśakālika and not Daśavaikālika, as is amply proved from the words of the Nijjutti. Thus it really meant 'ten chapters dealing with the rules of conduct and of begging food', the word Kālika being used in the sense of a part of the canon called caranakaranānuyoga or kālikaśrūta. Later on when the book was canonised, as is suggested by the story, it was included in the Utkālika group of the texts because it could be studied at any time of the day, though taken from the Pūrvas and at any year of the monk's paryāya. Here, there must have arisen some confusion about the name Kālika which, in close connection with the name of the group Utkālika, was taken in the sense of a book to be studied at the prescribed time, as there was the other group of texts in contrast with it, even though the word was used in the title in quite a different sense, namely to mean a kind of part of the canon dealing with rules of conduct. Naturally to overcome the supposed discrepancy the title was changed to Daśavaikālika, a term identical in meaning with Utkālika and a trace of which meaning is preserved in the remark of the Cūrņi. Later on the title was explained in the light of the story to mean the book composed at the time of the evening, another meaning of the term vikāla. This was tried to be supported by the facts of: the story as best as they could, and thus both the name and its interpretation were settled in a form quite different from their original nature.
Annotations :
1. Ed. Āgamodaya Samiti p. 201 b. 2. Ed. Prof. Abhyankara at the end of his edition of the text. The numbers of the
Nijjutti gāthās refer to his edition. 3. Ed. of Jāmnagar 1933. 4. Die Lehre der Jainas, p. 58. 5. Ibid., p. 75. 6. Cp. for a discussion of the title M. V. Patwardhan The Daśavaikālikā : A Study,
pp. 9-10. He himself accepts the traditional explanation. 7. Cp. Hema. Pari. V. 86. Samayasundara, p.1 8. Ind. Stud., XVI. 223. 9. Uttaraddhyayana, Intro. pp. 23. ff. 10. p. 203b.
The Title Dasavaikālika Sūtra
Indian Historical Quarterly Vol. XIV. 1938

Page #104

Māhārāștrī Language and Literature
The Māhārāștrī Prākrit
Māhārāstrī is one of the important Prākrits receiving a full treatment at the hands of both the Prākrit grammarians and Sanskrit rhetoricians. It is moreover the only language of the Dramatic group in which extensive literature is to be found. Dandin?, an author of the seventh century, says that Māhārāstri Prākrit is regarded as the language par excellence and Setubandha and other great works are written in that Prākrit. All the Prākrit grammarians from Vararuci onwards treat in their works Māhārāstrī as the chief Prākrit language and in case of other Prākrits give rules about those features only in which they differ markedly from Māhārāstrī. The value and importance attached to this Prākrit can be clearly seen from the fact that from comparatively early times the name Prākrit usually meant this language out of the whole group of Prākrits, a usage to which nearly all the writers subscribe? In the usual six-fold division of languages to be used in literature Prākrit occurs immediately after Sanskrit and stands there for Māhārāstri which is made evident by the specific mention of the names of other Prākrits that follow in the list. Both Vararuci and Hemacandra when they want to refer to Māhārāstrī write sesam māhārāstrīvat and seșam prākstavat", expressions which mean that all other features which are not dealt with there are to be the same as those in Māhārāstrī.
Reasons for this prominence are not far to seek. The grammarians of the Prākrit languages were in need of some system to be followed which would save them a good deal of their labour. They found in Māhārāstrī the only language in which nearly all the forms were to be met with and thus afforded them a very convenient starting point for the study of the whole of the Prākrit group. It supplied a very good standard of comparison for othe languages. Hemacandra finds it convenient even in case of such a language as Ardha-Māgadhi which is, to all intents, quite different from Māhārāstrī?.

Page #105

96
Amrita
The Sanskrit rhetoricians, on the other hand, were in need of illustrations to their rules of poetics and found in this language an extensive literature to hand, both epic and lyrical in character which would serve their purpose of illustrating the minute rules of poetic composition to their satisfaction. The major portion of the examples in these works is taken over from the literature in this language which shows that it was rich enough to serve their purpose well. The existence of anthologies of stray verses which would require no context for their understanding also helped them in citing these verses, a fact of no small importance to them. MĀHĀRĀȘTRI AND ŚAURASENÍ
This view of regarding Māhārāstrī to be the normal Prākrit is recently challenged by Mr. Ghose who has tried to prove that this suppositi based upon some kind of misconception. All along it was supposed by most of the scholars that the specific use of the word Präkrit as distinguished from its generic use to denote the whole group of the middle Indian languages, always meant Māhārāstrī and was taken to be true on the authority of the grammarians like Vararuci and the statement of Dandin in his Kāvyādarśa.
Mr. Ghosh has put these facts to a searching criticism and has come to think that all the evidence on which this view is based proves to be unreliable and in fact is a mistake on the part of later writers, who failed to understand the stand-point of the earlier ones. He comes to the conclusion that Sauraseni was really the Prākrit par excellence and what is known as Māhārāstrī is nothing but a later phase of it, developing under the influence of the Prākrit grammarians, if at all any reason is to be attributed for this change.
This view requires a careful consideration as it aims at the revaluation of the values of the various Prākrits and if true would lead us to a different solution of the whole of the Prākrit problem.
Mr. Ghosh relies upon a number of arguments to prove his position. Of them many may be admitted as facts proved, but it remains to be seen to what conclusion they should reasonably lead us. Hemacandra and many others do not mention the name Māhārāstrī and we can also add Vararuci to them, for his 12th chapter in which the word occurs only at the end, may be regarded as spurious and a later interpolation for which Mr. Ghosh has made out a strong case. Whether Vararuci wrote any 12th chapter which is lost and the deficiency was made up by some later writer producing the present inconsistency of arrangement or Vararuci could never have written

Page #106

Māhārāștri Language and Literature
97
any thing more than the first eleven chapters as he knew no Sauraseni other than what he treated in the first nine chapters as thought by Mr. Ghosh, remains doubtful. If accepted as true, Mr. Ghosh's position leads to the conclusion that the first nine chapters of Vararuci treat of Saurasenī. Now, many of the rules in them are taken over and worked out by Hemacandra and others to compose their Prākrit grammars and are made to serve the purpose of Māhārāstrī, a fact which cannot be explained unless one supposes that Hemacandra regarded the Prākrit of Vararuci's grammar to be Māhārāstrī. And this tradition can be pushed back to the times of Rudrața. Mr. Ghosh”s argument that the Saurasenī rules are not addding anything new to the first nine chapters of Vararuci's grammar is also due to his neglecting many differences of a minute nature. The difference admitted by Hemacandra is also small and mostly of a restrictive nature. In fact Vararuci has more rules and shows many more differences than given by Hemacandra. Vararuci says that in vyāprta t becomes d while Hemacandra includes this change in the features of Māhārāstrīl. Vararuci notes that in words like grdhra r is changed to i which is put by Hemacandra in his treatment of Māhārāstrīl2. The Sauraseni peculiarity of the change of jñ to nn in sarvajña and ingitajña is a normal feature of Māhārāstrī according to Hemacandra3. The form accharia is also allowed in Māhārāstrī by him and he makes no provision for the peculiar verbal bases, kar sumara, pekkha and pulla in Saurasenī4. All this is sufficient to show that many of the peculiar forms of Saurasenī were included under Māhārāstrī by the time of Hemcandra. So one fails to see, if Hemacandra's rules are sufficient to make a difference between Māhārāstri and Saurasenī how Vararuci's many more are insufficient to do the same. By a careful comparison of Vararuci's Sauraseni rules and their inclusion under Māhārāstrī by Hemacandra it appears that the natural development in Prākrits appears to be a fusion of the dialectical differences due to their becoming more and more literary languages and not a greater differentiation to give rise to new languages as supposed by Mr. Ghosh.
Mr. Ghosh has pointed out that rule 2, 7 of Vararuci is a great obstacle in accepting his Saurasenī as different from Māhārāstrī. But as proved elsewhere this fact should lead us to suppose that the peculiar change of t to d is not a peculiarity of Saurasenī, but it is not legitimate to conclude that there is no difference between these two languages on account of that single fact. The explicit statement of Dandin cannot be explained away on any preconceived theory.
Mr. Ghosh has also tried to support his position from
the theory of

Page #107

98
Amrita
Prākrit' which he finds in the Nātyaśāstrals. He argues that the Prākrit that is scantily dealt with in that book can be no other than Saurasenī, for all others named by Bharata can have no claims to the eminent position given to it there. Māhārāstrī the only, other Prākrit which can claim that position is not mentioned by Bharata in the list of his Bhāṣās and Vibhāṣās and was unknown to him. Sauraseni is the only important Prākrit in the dramas and therefore Bharata must have dealt with Saurasenī in his theory of Prākrit. This however remains at best doubtful and Bharata's treatment, coming as it does as a digression, gives us no clear hint that he is dealing with Sauraseni. If Bharata names sauraseni and prescribes its use in the dramas to a considerable extent, one fails to see why he should give his treatment of Saurasenī under the general name Prākrit, while it is equally probable that by Prākrit he means its generic connotation. This is made almost certain when.. we find him contrasting the Sanskrit and Prākrit Pathya with which his treatment of Prākrit begins 16
That the word Prākrit denoted either the whole of the Prākrit group of languages or else in its specific sense, the Māhārāstrī language can be proved from the consideration of a number of facts about its usage in Indian literature. Rudrata's enumeration of six languages in which Sauraseni and Prākrit occur side by side leaves no doubt of their separate nature. His illustrations!? also show the same thing and it is certain that Rudrata clearly regarded them as distinct from each other and used Prākrit to mean Māhārāstrī. The evidence of Dasarūpal and Kavyamīmāṁsā only shows that there the word Prākrit is used in its generic sense and that Māhārāstrī played no important part in the dramas.
Rājasekhara's views on Präkrit' are of some interest and importance. Nowhere does he mention the languages like Māhārāstrī, Sauraseni and Māgadhi with their specific names. In his Kavyamīmāmsā we find many references to Prākrit languages but in all the places he appears to give a fourfold division of languages into Sanskrit, Prākrit, Paišācī and Apabhramsa. Thus, while describing the Kāvyapurusa 20 he says that his face is Sanskrit, his arms Prākrit, his hips are Apabhrama and his feet are Paisāci. According to his conception of a Kavirāja2 he should be able to compose in all languages and for him there is no restriction of language, all being of equal value to him. He further notes that a particular language is favoured by the people of a particular country. The Gaudas favour Sanskrit, the Lātas Prākrit, those living in Maru, Takka and Bhādāna Apabhramśa, the people of Avanti; Pāriyātra and Daśapura favour Paiśācī while the poet living in the

Page #108

Māhārāștri Language and Literature
99
Madhyadesa favours al122. He says that one should recite Sanskrit and - Apabhramsa with grace and Prākrit and Bhūtabhāsā with dignity. People living
in countries like Magadha, east of Benares, speak Sanskrit well but are not able to recite Prākrit to prove which Rājasekhara quotes a famous but ironical verse. The Lātas speak Prākrit charmingly but hate Sanskrit23. In the hall of the king Sanskrit poets should sit to the north, Prākrit poets to the east, Apabhramśa to the west and Paiśācī to the south, a distribution based upon the places where these languages were current24.
In his Bälarāmāyana25 Rājasekhera calls himself well-versed in all the languages and points out that one and the same thought becomes different if written in Sanskrit, Prākrit, Apabhramba and Bhūtabhāsā or even by their combinations. He describes Sanskrit as beautiful to hear, Prākrit as elegant and possessing natural sweetness, Apabhramsa as very smooth and the Bhūtavacana as well-formed. In the Karpūranañjarī26 he contrasts the tenderness of Prākrit with the harshness of Sanskrit.
rit.
From the information supplied by Rājasekhara it is difficult to conclude that he knew of Prākrit to mean Sauraseni only. He nowhere mentions Saurasenī, nor Māgadhī nor Māhārāstrī and the natural conclusion appears to be that under the generic name Prākrit he includes all these three languages. His four-fold classification of the language is an extension of the one of Dandin and Bhāmaha into the three-fold one with the addition of Paiśācī, which by this time became independent of Prākrit, for what reasons it is difficult to say. That Paiśācī was regarded as different from Prākrit and Apabhramsa is also known from the tradition of the four-fold division of Prākrits27 as preserved by the eastern school of the Prākrit grammarians in later days. Dandin knew of Paiśācī and Māhārāstrī28 and yet classified literature into Sanskrit, Prākrit and Apabhramśaa. So also Rājasekhara divided it into four and thus his use of Prākrit is not in a specific sense. His use of Sauraseni in the drama, Karpūramañjarī proves that he knew of it and did not identify it with Prākrit. His one quotation from the Gathāsaptaśati30 also shows that he knew of Māhārāstrī and included it under the general term Prākrit.
Even though Hemacandra does not mention Māhārāstrī by name, that he makes a distinction between Prākrit and Sauraseni follows from a number of considerations. His two Sutras sesam prākrtavat 31 and sesam saurasenivat32 make a difference between the two. His Kumārapālacarita makes it still more evident, while his Kāvyānuśāsana makes a distinction between the two varieties of the Bhāsāślesas Sanskrit-Prākrit and Sanskrit-Saurasena33. He

Page #109

100
Amrita
further states that works like Setubandha are written in Prākrit and are divided into Aśvāsakas34. That Hemacandra deals with Māhārāstrī in VIII, 1-3 and also regards that language to be the same as the one used in works like Gäthäsaptaśati, Gaudavadha. Setubandha and Visamabānalilā follows from his questions from these works35.
Bhoja appears to draw for his remarks on Prākrit 36 both on Rudrata and Rājasekhara which renders his own views obscure. He regards the verses in the dramas to be written in Saurasenī7 and yet his remarks leave no doubt that he also regards Māhārāstrī and Sauraseni as distinct. With these facts before one it is evident that Dr. Chakravarti misses the real meaning of the word Prākrit when he calls the six-fold division of Rudrața to be logically defective 37
LINGUISTIC NATURE OF MĀHĀRĀȘTRĪ
The nature of this Prākritis fairly illustrative of nearly all the tendencies working in the whole of the Prākrit field. It is, as all others are, based upon the vocabulary of Sanskrit in the main and shows only a few words which cannot be traced to a Sanskritic origin for which reason they are designated as Deśī. The number of such words is however very smalll and it: is also possible that many of them which are found in the usually accepted list of these words are nothing but obscure Sanskrit words or words used in a figurative sense38 or lastly words greatly transformed so as to become different beyond ready recognition. Therefore except for a small number of these words the whole of the vocabulary of this Prākrit can be traced to a Sanskritic origin.
In the early works of this language the vocabulary is marked by the introduction of many Deśī words. In the Sattasai we find them in abundance which is not the case with later works like Setubandha and Gaudavadha. The cause for this difference appears to be that the former work is much more popular in character and is therefore found to preserve the language in a form nearer to the vernacular of the people than the language of the later works of the well known Sanskrit scholars. This change is further due to the growing influence of Sanskrit which was working over the Prākrit literature throughout its development, on account of its greater vogue. This fact invariably led the Prākrit writers to give preference to a Sanskritic vocabulary over genuine Prākrit one as being intelligible to a wider circle of readers and which they supposed would give their works a greater durability and which proved true to a considerable extent. Another characteristic of the vocabulary

Page #110

Māhärāștri Language and Literature
101
which is on increase in later-day works is the choice and favour shown to Sanskrit words which undergo no phonetic change or minimum change while coming over to the Prākrit language and which are technically called Tatsamas. This was again due to the desire of the writer to bring up his work as near Sanskrit as possible. On the whole this tendency of writing Prākrit in a purely Sanskritic manner must have been the chief cause of removing these popular languages far away from the spoken dialects of the people their original source and transforming them into literary languages.
The changes that have affected the Sanskrit vocabulary while coming over to this Prākrit are of some interest. There is, first of all changes in the genders) of words which differ in few cases between Sanskrit and Prākrit. Words like sarad, prāvrt, and tarani which are feminine in Sanskrit are regarded in Prākrit as masculine. Similarly words which end in s or n and which are neuter in Sanskrit usually become masculine in Prākrit. A few purely neuter words like aksi and vacana now slowly take on the role of masculine words while a few originally masculine words like guna become neuter. Words ending in įmā and the word añjali now become feminine while bāhu when in the form of bāhā is regarded as feminine. If we look a little scrutinisingly at these changes we find that the rules which are operating here are the natural effects of the tendency of regarding the grammatical gender to be determined by the ending vowel of the word and not on the intrinsic nature of the object denoted by the word. Therefore if a word is found to end in à or i it is regarded as feminine in gender for the majority of words of these endings in Sanskrit are feminine. This is well exemplified by the case of anjali and bahu when it becomes bāhā, the first of them lengthens its ending vowel in Prākrit while the second now ends in ā. The three words dāman, śiras and nabhas, because they become dāmam, siram and nahaṁ in Prākrit remain neuter while others which become jaso tamo regularly become masculine. There is further some confusion between masculine and neuter nouns which can ultimately be traced to the similarity of forms of these two classes. The evidence of Prākrit, it will be seen, is in full agreement with Brugmann's theory of the origin of the grammatical gender40.
In the field of phonology of this Prākrit we find nearly all the laws of phonetic change working in more or less general manner. Changes in the quantity of vowels due to the loss or shift of the accent are illustrated by the shortening of the vowels in words 41 like pānīya, alika, karısa, śirisa and lengthening in prakata, pravacana, and others. These changes in the place of the accent itself are due to the fact that the accent is chiefly a stress accent

Page #111

102
Amrita
in these languages while it was a musical one in the ancient Sanskrit. Syncope is found in words like kumbhakāra, bhājana and others where it is attributed to the influence of accent. It is possibly the reason why in a few words a consonant is doubled 42 as in tella, nedda etc. Māhārāstrī is very rigorous in the application of rules of dropping the intervocalic consonants. Compared to other Prākrits it has a very airy appearance and affords greatest difficulties in identifying its words with their exact Sanskrit counterpart. A probable reason for this fact is suggested by its usage in songs and lyrics, a fact evinced by its exclusive use in verses in the dramas. As a result of this particular use the language lost all its consonantal structure and its hardness, leaving behind only a string of vowels and a few consonants to hold them together. It is however doubtful whether it is the result of this specific use of this Prākrit in the songs or that it was chosen for that purpose because of its loose and melodious character. Unless we get materials to determine the nature of this Prākrit before its introduction in the dramas we are not in a position to determine the question one way or the other. A more plausible explanation of this peculiarity is, however, to be found in the tendency of the grammarians to formulate general rules from scanty material. It was further accentuated by the close following of the rules by the later writers in Prākrit.
The distinctive features of this Prākrit are the wholesale dropping of the intervocalic stops43, the reduction of all the aspirates to h44, the change of n to n at the beginning and middle of a word45, where however the change at the beginning is against the testimony of the modern vernaculars, the use of single sibilant s and the changes of the unvoiced cerebral stops to voiced ones. Of the vowels, r is changed preferably to a wherein it agrees with Pāli 46, ai and au to aï and aii while Hemacandra says47 that ai and au are also allowed by some grammarians, a fact which may be true in case of Māhārāstrī, as it is found in its later stage of the formation of Marathi. The conjunct ry becomes ij and ks changes more to cch than to kkh as in other Prākrits. Jñ and ny become nn and a few other tendencies which also appear in Marāthī.
Unlike other Prākrits Māhārāstrī shows traces of phonetic changes which are otherwise regarded as the distinctive features of other languages like Sauraseni and Māgadhī, and these are found in a few words. Thus the Sauraseni peculiarity of softening 48 the hard consonants like t into d, k to ğ, and th to dh is found in words like dā, dāva, gendua, pidham, or pudham, savvao or savvado. Vararuci states that in words like rtu and others t is changed into d, kirāta becomes cilādo, the ablative singular terminations are

Page #112

Māhārāștri Language and Literature
103
do and du while it is also found in prepositions like kado, jado, tado. The Māgadhi change of r to l is met with in words like calana, hallidā, juhitthila and sukumāla49. But the most important instance of this tendency of mixing characters of other languages is the features of Apabhramba that we meet with in Māhārāstrī. There we have the change of m to my or vṁ in ahivannu, navai rovais" and the opposite change of v tom as in gamesai, kamandha, nima and amelast. The existence of forms like jaunā, sumino52 the presence of conjuncts with r in vundra and bodraha53 and the change of s into h in daha, coddaha, pāhāna, divaha54 and others all betray the Apabhraṁsa influence over this Prākrits
Of these, the change of t to d is not a borrowing from the features of Saurasenī but a natural stage in the development of this language itself, mostly a survival of its traits in its early stage56. The presence of other features it appears, is due to the extensive use of this language in nearly all the provinces outside its original home which in its turn appears to be the result of its being taken to be the only language suitable for writing literary works of some pretension. Naturally writers from outside its original country, while writing in it, introduced consciously or unconsciously forms from their own provincial dialects which were very familiar to them. Later on the grammarians, drawing their material from the literary usage, appear to have included them in their treatment of this Prākrit as being sanctioned by good usage and so correct in it. In the present state of our information it is not possible to separate all such non-Māhārāstri element from its original features. It is equally possible to regard that this foreign element in this language is due to some kind of confusion on the part of the Prākrit grammarians and lack of finer distinctions between the different dialects which they never attempted. The Apabhramsa element, however, was introduced at a later stage and went on increasing as Apabhramśa spread widely.
The morphology of this language shows similar simplification and decay. There is no distinction to be observed between nouns ending in vowels and consonants. All words which originally ended in consonants are made vowel-ending either by dropping the final consonant or adding a vowel at the end, a tendency found in the classical Sanskrit itself. Further the whole of the declension was assimilated to the standard of the a-forms and words ending in other vowels also were given forms from the analogy of a-forms. Again no distinction was observed between long and short vowels which greatly simplified the intricacies of the Sanskrit scheme of declension. But along with this simplification there goes on the process of invention and a vast amount

Page #113

104
Amrita
of new forms are due to the use of alternative bases for the declension of nouns and pronouns. Of the cases, the dative is lost and genitive takes its place. Here also, first the plural form was lost and the singular followed. The whole of the dual number was suppressed and its function was discharged by the plural, a natural course of development to be observed in many IndoEuropean languages57. It was mainly due to the absence of any practical necessity of making a distinction between the things in pair which were usually denoted by the dual in Sanskrit and other ancient languages and things in group of more than two. Other forms of nouns show the effect of the working of the phonetic laws over the finished forms of the Sanskrit language while a few are to be explained by the mixture of the pronominal and nominal declensions and the effects of the law of analogy.
Hemacandra's treatment of the pronominal forms leaves no doubt that this vast array of forms must have been the result of some confusion of dialectical differences which are neglected by the grammarian and the indiscriminate use of different bases due to a optional working of a phonetic law or even some scribal error. The ablative appears to be particularly rich both in case of nouns and pronouns. The most distinctive features of Māhārāstrī in declension are the locative singular in mimi, ablative in a and nominative in o.
The conjugation of the verbs also has a similar line of development. The ten tenses and moods of the classical Sanskrit and the Vedic dialect are reduced to two tenses and two moods. The present and future are retained and the imperative and the potential moods preserved their full ,conjugations. The past tense was represented by one form derived from the aorist and served the purpose of all the persons and numbers. But the function of the past tense was taken over by the past passive participle which was regularly corrupted from the Sanskrit forms. In Sanskrit itself this method of using nominal forms for the finite verbs was on increase and gave rise to what is called the nominal style of writing59. The distinction between the two Padas which was loosened in the Epic Sanskrit now completely vanished and terminations of both were indiscriminately used to form forms. The Parasmaipada was favoured and soon drove the other out of the field.
In the formation of the verbal derivatives this Prākrit, like many. others, appears to draw more on the Vedic language than the Classical Sanskrit. This forms one of the links which connect the Prākrit language with the Vedic dialect without the intervention of the Classical Sanskrit60. The

Page #114

Māhārāștri Language and Literature
105
richness of the forms of the gerund betrays their dialectical origins. The peculiar features of Māhārāştri appear to be the passive formed by the addition of ijja, future with the help of hã and hi, the gerund in üna and a few verbal substitutes which also find their way in the modern vernacular, Marāthī.
THE HOME OF MĀHĀRĀSTRĪ
The name of this Prākrit would indicate that the home of this language was the country of Māhārāstra at least at the time of getting that nomenclature, in its early stage. It is often contended that various Prākrits that are preserved to us in literature are not the real vernaculars of the people of various localities but are to be considered as literary languages and therefore have no local basis to be called their home. Such a view is based, however, on the mistaken belief about the nature of these Prākrits and the nature of a spoken language. The Prākrits are no doubt literary languages and are essentially known to us in a form preserved by scholars in their literary productions of more or less scholarly nature. In this sense they are emphatically literary languages. But it is equally true that they are in no way to be regarded as artificial languages created by some particular social group without any spoken language at their basis. Originally Prākrits were the spoken languages of the people and their true vernaculars. In course of time they were refined and polished greatly with the help of the grammarians and were made suitable for literary expressions. Such a process naturally involved the removal of irregularities, the use of a few favoured forms and a more strict and accurate use of words and forms which come up to the standard of the grammarians. It is true that in this process of refinement and elevation some rules were carried to a greater extent outside their proper sphere than was warranted by the facts of the case, in the zeal to form comprehensive rules of grammar. All the same, these Prākrits are based on definite dialects which were the means of communication of the masses and so their vernaculars. The home of a Prākrit naturally means the locality where the dialect at its basis was current and spoken by the people.
In determining the home of Māhārāstri it is but natural that the country of Māhārāstra should be regarded as the place where this language arose. But such a suggestion is often repudiated and even attempted to be disproved. It is pointed out that Māhārāstrī was used throughout India and as the language of the Prākrit epic and lyric was current all over the country. As early as sixth century A. D. the Setubandha was written in that language

Page #115

106
Amrita
is Kaśmīr at the other end of the country. Others try to prove that Māhārāstrī has nothing to do with this country as its present vernacular Marathi is not closely related to this Prākrit. But viewed impartially one has to admit that the real home of this Prākrit is no other than the country of Māhārāstra. Besides the name which itself is an indication of this fact, we have clear traces of the Māhārāstrī language to be met with in Marāthī62, which shows in many ways a further development of this Prākrit. This is shown by the vast amount of Prakrit words which are peculiar to Māhārāstri and to be found in Marāthi, while other Prākrits differ from both of them. A comparison of words like maharastra, marahatta, marhatha; sadrśa, sarisa, sārakhā; pakva, pikka, pikane; angāra, ingāla, ingala; lalāta, nidāla, nidhala; khudia, khudane; saymā, sejjā, seja; valli, velli, veli; ārdra, olla, ole; nirjhara, ojjhara, ojharaane; and a host of others would convince us about the close connection between these two languages63 Add to this, the formation of the gerund in ūna and a few verbal bases and the Deśī words common to both and we have to admit that they are the languages related as mother and daughter towards each other. As regards its use throughout India it is found only when the language had attained prominence. But even then we have evidence to think that in its early stage it was used in the country of Māhārāstra and later spread all over India for the writing of Prākrit epics on account of its great popularity. This is sufficiently proved by the fact that the oldest work in this language the Sattasai of Hāla is a purely Māhārāstrian production which appears from its constant references to mount Sahya and the river Godāvarī. THE NAMES OF THE COUNTRY AND THE LANGUAGE
To understand the origin and development of this Prākrit we must briefly review the early history of this country to see the conditions in which this language must have originated and developed. It is not possible to determine with precision the date of the colonisation of the Deccan by the Aryans of the North and particularly the aryanisation of Māhārāstra. There is no mention of the country south of Vindhyas in the early Vedic literature64. Pāņini refers to Kaccha65, Avanti, Kośala67, Kalinga6 all of which lie to the North of the river Narmadā. Only Aśmaka69 that is mentioned by him, lies to the south of that river. So he appears to be ignorant of the major portion of the South India. The Aitareya Brāhmana' calls Bhīma a prince of Vidarbha and the story of Sunaśśepa? mentions by name such tribes as Āndhras, Pundras and Mūtibas which are probably South Indian tribes. Therefore, most of the scholars agree in regarding that the colonisation of South india must be dated later than the time of the great grammarian Panini who is variously

Page #116

Mähārāştri Language and Literature
107
assigned from 800-400 B. C. In later literature we find clearer traces of the knowledge of this country as in the Suttanipāta72 one of the ancient works of the Pāli canon which gives us the story of Bāvari who leaves off the country of Kośala and comes to Asmaka in South India. Kātyāyana73 refers to Pāndya and Cola two famous countries in the extreme south. There is a reference to Tāmbraprani in the inscription of king Aśoka74 while Kautilya refers to a king Bhoja Dāndakya who lived long before him.
According to a tradition preserved in the Rāmāyana we find that there were two different settlements in South India of the colonising Aryans. The first of them was led by Bhrgu and his descendants who entered that part of Konkana on the shore of the western ocean and which was known by its other name Aparānta. The eastern part of the country called Vidarbha was early colonised by the sage Agastya and his followers who later on occupied the middle country called the Dandakāranya a name explained by the tradition with the story of a king Dandaka who insulted the sage Cyavana and was cursed by him. Even though the traditional names are susceptible to objections this appears to be the only way of overcoming the natural obstacle of the mountain Vindhya and must have been the most easy route for the incoming Aryans. Dr. Bhandarkar75 has suggested the precise route to be from Māhismatī through Vidarbha to Pratisthāna mostly relying on the correctness of the account of the route of the pupils of Bāvari as given in the Suttanipāta. From the name Pāndya and the name of the southern town Madurā he comes to the conclusion that the people who colonised this part of the country came from Mathurā and belonged to the tribes to which the Pandavas belonged. S. K. Ayangarah puts the date of this southern migration on either side of 750 B. C.
From the nature of the Prākrit it appears probable that immediately after the Aryan migration in the country this Prākrit language was formed as a result of the intermixture of the Aryans and the original inhabitants of the country. It is however impossible to decide whether this Prākrit was a result of the change that affected the ethnological structure of the incoming Aryans or due to the transference of the language from the Aryan population to the aboriginal people who destorted it and turned it into the present Prākrit form.
The explanation of the name of this language offers unusual difficulties. If we are to follow the analogy of the names of other Prākrits we will have to say that like Sauraseni and Māgadhī, Māhārāstrī also derives its name from the country from which it took its origin. But there are some

Page #117

108
Amrita
difficulties in accepting this natural explanation which cannot be ignored. Unlike the other two names of Magadha and Saurasena, Māhārāstra is not the oldest name of the country and is said to be very late in its origin, according to the majority of scholars. It is sufficiently proved that the original name of this part of the country was Daksināpatha or Mahākāntāra or even Mahātavi. Vidarbha alone is known from very early times. The country got the specific name Māhārāstra very late. Asoka's inscriptions refer to Rāstrīkas which must be located in South India and the inscription of Samudragupta mentions one Devarāstra which is usually taken to be identical with the present Mahārāstra but which Allan? tries to identify with the Vizagapattam district of the Madras province. The oldest reference to this country is to be found in the Brhatsamhitā78 of Varāhamihira who flourished about the fifth century A. D. In A. D. 611, it is mentioned in the inscription of Satyāśraya Pulakesin. In his Kāmasūtra" Vāstāyana mentions both the name of the country and its inhabitants Mahārāstrīkas and his work is approximately dated 400 A. D.
This late origin of the name has led scholars to speculate about its probable significance. The most complicated and unsound theory put forth is that of Rajawade 80 who says that the Mahārāstrīkas were the followers of the king of Magadha and because of the spread of the Buddhist religion in that country, they migrated to the south and in course of time absorbed the earlier tribes of Rāstrīkas and Vairāstrīkas which gave rise to the Trimahārāstrīkas which is found mentioned in some inscriptions. They further mixed themselves with the Nāga people about 400 A. D. Vaidya has suggested that the name Mahārāstra is due to the mixing up of a number of small kingdoms like Vidarbha, Asmaka, Pandurāstra, Goparāstra and others which took place about the beginning of the Christian era. Kane8 also agrees in taking the word to mean the great kingdom. Dr. Ketkar82 however insists in taking the word to mean a country inhabited by the tribes of Mahārs and Rāstrīkas and thinks that the country got the name at about the seventh century A. D.83
We have, however, ample evidence to show that this Prākrit was called Māhārāstrī at a time much earlier than the date to which the name of the country is assigned at the earliest. Vararuci mentions it and deals with it at great length and his date is fairly early. The Sattasai can be reasonably assigned to the beginning of the Christian era and what is still important is that it presupposes a vast literature earlier to it. It is equally improbable to suggest that the country got its name from the language. The grammatical form goes against it. The only alternative left open is to regard that both the

Page #118

110
Amrita
to be the date of its origin. On the contrary it suggests that it existed long before it and was taken over in the dramas because of its popularity in literature.
It is not easy to determine the date of the beginning of this language. Jacobi has suggested that the third century is the probable time for the beginning of its use in writing lyrics and songs and it was later adopted in the dramas at the time of Kālidāsa. He points out that Bharata nowhere, mentions it in his Nātyaśāstra and the dramas of Aśvaghosa and Bhāsa show no trace of this language. This evidence would lead us to date this language after the third century when Bhāsa and Bharata flourished. This position is further strengthened by his theory that even in case of epics and stray verses Māhārāstrī was not used at the time of Nātyaśāstra but its place was taken by some other Prākrit dialect. Keith86 also emphasises the same facts and puts the language after the fourth century adding that the dropping of the intervocalic consonants was a very late tendency and is absent in ancient Prākrits.
The evidence adduced above leads to no definite conclusion. The fact that the older dramatists have no verses in Māhārāstri and that Bharata does not mention it is no doubt to be explained on the supposition that from early times this language played no important part in the Sanskrit dramas. Therefore neither the dramatists nor the theorist found it necessary to use or mention it in their works. But its use in other literary fields should not on that account be dated so late. The Pre-classical Prākrit or the older Sauraseni had a restricted use for the Dhruvās in the dramas. But for the writing of the epics and lyrics there is no evidence to show that it was used instead of Māhārāstrī and the later usage makes it extremely probable that from the very beginning Māhārāştri was used in them. The oldest work, the Sattasai confirms the same and when we find that it presupposes a vast earlier literature from which it was selected, it is but natural to date the origin of this language a few centuries earlier than the date assigned to it by Jacobi and Keith. DIALECTS AND USE IN LITERATURE
Prakrit grammarians do not give sub-dialects of this language. There are many dialects found in case of other Prākrits like Saurasenī and Māgadhi and it appears strange that Māhārāstrī should have no dialects particularly when it extended over a very wide area, a fact which should contribute to the emergence of many dialects. This fact may be due to its early

Page #119

110
Amrita
to be the date of its origin. On the contrary it suggests that it existed long before it and was taken over in the dramas because of its popularity in literature.
It is not easy to determine the date of the beginning of this language. Jacobi has suggested that the third century is the probable time for the beginning of its use in writing lyrics and songs and it was later adopted in the dramas at the time of Kālidāsa. He points out that Bharata nowhere mentions it in his Nātyaśāstra and the dramas of Aśvaghosa and Bhāsa show no trace of this language. This evidence would lead us to date this language after the third century when Bhāsa and Bharata flourished. This position is further strengthened by his theory that even in case of epics and stray verses Māhārāstrī was not used at the time of Nātyaśāstra but its place was taken by some other Prākrit dialect. Keith86 also emphasises the same facts and puts the language after the fourth century adding that the dropping of the intervocalic consonants was a very late tendency and is absent in ancient Prākrits.
The evidence adduced above leads to no definite conclusion. The fact that the older dramatists have no verses in Māhārāstrī and that Bharata does not mention it is no doubt to be explained on the supposition that from early times this language played no important part in the Sanskrit dramas. Therefore neither the dramatists nor the theorist found it necessary to use or mention it in their works. But its use in other literary fields should not on that account be dated so late. The Pre-classical Prākrit or the older Sauraseni had a restricted use for the Dhruvās in the dramas. But for the writing of the epics and lyrics there is no evidence to show that it was used instead of Māhārāstrī and the later usage makes it extremely probable that from the very beginning Māhārāstrī was used in them. The oldest work, the Sattasai confirms the same and when we find that it presupposes a vast earlier literature from which it was selected, it is but natural to date the origin of this language a few centuries earlier than the date assigned to it by Jacobi and Keith DIALECTS AND USE IN LITERATURE
Prākrit grammarians do not give sub-dialects of this language. There are many dialects found in case of other Prākrits like Sauraseni and Māgadhi. and it appears strange that Māhārāstrī should have no dialects particularly when it extended over a very wide area, a fact which should contribute to the emergence of many dialects. This fact may be due to its early

Page #120

Māhārāștri Language and Literature
111
standardisation and the activity of the grammarians must have been wide and valued enough to fix its form early. This left no scope for later writers to deviate from the norm laid down by the grammarians and must have been the chief agency to make the language uniform. But there are traces of the existence of some dialects in the present language itself like the use of so many forms for a particular formation of grammar. It is best explained by supposing that the grammarians overlooked the minute and fine distinctions of dialects and put all the forms together in writing their grammars. The presence of a few phonetic tendencies which are really foreign to the spirit of this language goes a long way to prove that it had formerly dialects which were soon forgotten and their peculiarities incorporated in the standard language. Its wide use in writing the epics and its early stereotyping in the form of a líterary language are responsible for this to a certain extent.
The practice of the dramatists also gives us no additional information about the dialects of this Prākrit. Prthvīdhara87 says that both the characters Candanaka and Viraka in Mrcchakatikā speak a dialect called Āvantikā and gives its peculiarity to be the presence of the sibilants and richness in proverbial sayings. According to Mārkandeya this language is a mixture of both Māhārāstri and Saurasenī which agrees with the few passages found in the drama. From these facts it is difficult to classify this dialect as either that of Māhārāstrī and Saurasenī. From a remark of Candanaka that the southerns speak indistinctly Pischel suggests that Candanaka is not speaking Avantikā but another dialect called Dāksinātyā which is mentioned by Bharata and the Sāhityadarpana. If this is true it differs very little from Māhārāstrī or even Avantikā. Woolner would take both these dialects as nearly related to Sauraseni because the passages in the Mss. show many Saurasenī features88.
Both Mārkandeya and Rāma Tarkavāgīsa mention the dialect Dāksināty, and Viśvanātha89 tries to make it identical with the dialect Vaidarbhikā. Lassen long ago classified it along with Māgadhi and ArdhaMāgadhī because he found them mentioned together but for no other logical reason. The names of both the dialects would suggest that they are the dialects of Māhārāstrī Prākrit and current in South India and the country to the north of it. Grierson has further suggested that these late references to the dialects may be to the dialects of the present day Marāthī which is certainly old enough to receive mention from these writers.
All the writers on poetics are unanimous about the role this Prākrit plays in the Sanskrit drama. The song of the ladies who usually speak

Page #121

112
Amrita
Sauraseni are to be written in Māhārāstrī. But Bhoja alone calls a song from the sakuntalā to be written in Saurasenī. His readings also preserve the change of t into d in that gātha. But it is difficult to believe in the statement of Bhoja in this respect. The softening of t is not foreign to Māhārāstrī and this must have misled him to call the verse to be written in Sauraseni as he came at a time when this change was taken to be the distinctive feature of the Saurasenī Prākrit.
The epics were no doubt written in this Prākrit alone. As remarked by Dandin” Kathās were also composed in it but we are not sure whether they were in prose or verse. Taking into consideration the fact that there is no prose in existence in Māhārāstrī it appears more probable to regard them as written in verse. This is confirmed by Namisādhuo who says that they are to be written in other Prākrits but not in prose. Bhojao says that it is used by characters of middle status. MĀHĀRĀȘTRĪ AND MARĀȚHI
Another complicated problem about this Prākrit is to trace its legal descendant among the modern vernaculars of India. It is no doubt Marathi which can lay best claims for it and it is the present day vernacular of the country of Mahārāstra where Māhārāstrī also took its rise. The older view that Marāthi is to be derived from Pāli is clearly wrong. Garenzo first tried to prove the close relation between Māhārāstrī and Marāthī. He pointd out that the Marāthi gerund in ūna is from the Māhārāstrī one in ūna, both have a feminine form of the demonstrative and relative pronouns, the emphatic particle cia in Māhārāstrī is found in Marathi in the form of eca and lastly both have a good many peculiar words. Pischel has accepted this view, which was first called in question by Grierson". Later on Grierson also admitted it as true and adduces a few more points of similarity. Marathi shows its near relation to Māhārāstrī in features like the formation of gerund, the change of ai and au into ai and aü which later revert to their original forms, the favour shown to the rule of simplifying the conjunct by lengthening the preceding vowel, the change of ks to cch, the peculiar basis of the personal pronouns of the first and the second persons, and a host of words showing abnormal changes common to both of them. Many Deśī words in Māhārāstri are found in Marāthi and similar is the case with the Dhātvādeśas.
Even when we accept that Marathi is a descendent of Māhārāstri it remains yet to trace the exact line of development by which it originated. Grierson following his general scheme of the development of all the modern

Page #122

Māhārāştri Language and Literature
113
vernaculars postulates the existence of an intermediate stage between the two languages which he prefers to call by the name Māhārāstra Apabhramśa%. In two cases our present day Marathi appears to draw from the Apabhramśa of the grammarians, the use of the interogative particle kona from kavana and the formation of the present participle like karata jata. To this may be added the nominative singular ending u if we take the older stage of Marāthī into consideration. Beyond this there is no evidence to show that Marāthi has much to do with Apabhramsa as in the case of Gujarātī. The case of nominative singular can be explained as the weakening of the Māhārāstrī termination. The peculiar. Apabhramśa features like the formation of gerunds, the use of ha terminations in declension and conjugation, the retention of r in conjuncts are not found in Marathi. On the contrary in all these cases Marathi follows closely the Māhārāstri Prākrit. Further as the old Marāthi words in y, V, and h do not show the nominative ending u, it is more probable to regard it as a case of weakening. Another suggestion is made to the effect that there was only one Apabhramśa from which all the vernaculars are derived. On the whole the available evidence is favourable to think that Marāthī is directly descended from Māhārāstrī without the intervention of the Apabhramsa stage, 'unless we like to call the works in old Marāthī to be written in a Mahārāstra Apabhramba. MĀHĀRĀȘTRĪ LITERATURE
By a comparison of Māhārāstrī with other dramatic Prākrits it becomes apparent that it stands on a different basis than the other Prākrits. It is a language primarily used to write works of the nature of artificial epics while its role in the drama is, after all, secondary. In fact speaking strictly this language is not a dramatic Prākrit at all. Like others it was not found in the Sanskrit dramas from the very beginning but was introduced later probably at the time of Kālidāsa. The reason appears to be the prominence in which it was held in other departments of literature. Another distinction between the dramatic Prākrits and Māhārāstrī appears to be while other languages are restricted to the use in the dramas Māhārāstri is found in other than dramatic literature in its real importance. This literature is vast and of great value to merit detailed consideration. THE SATTASAI OF HĀLA
The Sattasaī or Gāhāsattasai" which is its fuller name, is an old anthology of Māhārāstrī verses in the gāthā metre and amounting to some seven hundred pieces. It is traditionally attributed to a king Hāla or

Page #123

114
Amrita
Śātavāhana. It is the most famous and best known of the Māhārāstrī works. Its value as an anthology is fairly high and it also affords good historical evidence to show that Māhārāștrī literature was once very extensive and widespread. Its popularity can be easily seen from the good number of commentators and the vast amount of quotations from it in later works on poetics, as well as the use made of it by the Prākrit grammarians. Bāna98 bestows on it high praise in the introductory verses of his Harsacarita and Govardhana thought it suitable to compose a Sanskrit Aryāsaptašati on its model. Besides it gives us useful information of the rural and rustic life of the Deccan and reveals in a charming manner their proverbs and sayings current among them.
The work is divided in seven Satakas or collections of hundred verses each, which however differ very much in various Mss. preserved to us. This Prākrit anthology is mostly of erotic contents, love in its various phases and stages being the dominating sentiment either of the writers themselves or of the taste of the collector. The Indians have notably developed greatly in the composition of gnomic poetry with a peculiar force of its own. These ancient Indian poets were masters of this art which produces pictures in a few broad and bold strokes giving us a small incident or a miniature picture complete in itself. The present collection contains many such master-pieces which depict village life. The chief feature of the present work is that it mainly occupies with the painting of the village life and the peasantry more than the highly polished and sophisticated living of the courts and palaces. Naturally it is rich in pictures of nature cottages, country-side gardens, fields, thickets forests, rivers, streams and mountains, animal and plant worlds, and other things essentially rural. The family life of the lower society is given from all possible points of view but the main stress lies on the erotic side. The commentators, however, coming as they do in an age fully dominated with the irresistible force of theoretic considerations, try to take almost all the verses as erotic in sense and this they are able to do with that much-valued suggested sense. In the majority of cases their interpretation is justified, but in this absurdly consistent procedure of theirs they appear to lose sight of the real meaning of the poet when they come to see in verses dealing with the other sides of village life, studies of nature pictures and simple animal behaviour the same suggestion of illicit love when no erotic contents can be reasonably seen in them. All this should not lead us to think that the present work is wholly devoted to the treatment of village life and nothing else. There are verses about other topics as well and in the traditional list of poets we find a few

Page #124

Māhārāştri Language and Literature
115
names of kings and famous court poets.
The scope of the anthology is extensive and deals with varied subjects drawn from different walks of life. Winternitzo would not like to call them specimens of ancient songs of the Indians dealing with the joys and miseries of their lives, but only artistic poems closely modelled on them, because he suggests that Prākrit was not the language of the people but a refined and literary vehicle of expression based on such a dialect. Most of the verses are detached and self-sufficient but in few cases two or three verses relate to one topic.
The majority of the verses deal with the love of the village maidens and youths, shepherds and hunters, and even common labourers. In few words and in an admirable way a tune is struck, a pain is sounded or a joy is pictured. Very often a brief picture of a labourer is given in a single verse. We often hear the lamentation of a love-lorn lady, her longings, her plaintive
on her situation, her misery and pain of separation to which she sometimes gives vent. The usual conventions of Sanskrit poetry are also followed. As usual the winter season brings the lovers together while lightning and storm make the beloved to embrace him. The wishes of a lady in love are well expressed. She desires the moon to touch her with the same rays with which he has touched her lover, and wishes that night should last perpetually because her husband is going on a journey on the next day. A traveller prays that rain and wind should strike him as hard as they can but spare his beloved at home. The onset of the rainy season enkindles the flame of love in the hearts of the lovers. The poets also describe in minute details the varied effects of love when he points out that one looks deeply into the eyes of the other and drinks with them. The beauty of the ladies is also described in a microscopic manner, in which every one of their limbs gets its due share. Many times the poet gives us situation not without a comic effect as when the boy climbs the back of his father when he falls at the feet of his wife 100
Next to these come the pictures of nature. In four short lines the poet can effectively conjure before the eyes of the readers various scenes of the seasons, summer, autumn, winter and rainy-season, by pointing out their striking features. Winter lays bare the fields of corn while the summer burns them down. Equally fine are the pictures of animals, bees hovering round flowers, the peacock and the crow enjoying the rains, the antilope and his mate loving each other and rubbing with their horns, the female crane sitting

Page #125

116
Amrita
silently on a lotus leaf and the many pranks of apes. Besides these the plants get their share and the mountain Sahya with the river Godā figures prominently. There are a good many maxims and popular sayings of a gnomic nature while incidently many gods, situations from the epics and folk-lore are also mentioned.
The work also gives us a few references of general interest. Of kings we meet the two famous names of Vikramāditya and Sālāhana, beside the name of Hāla the compilor of the anthology at the beginning and the end of the śatakas. The geographical atmosphere of the work is pretty old. No big towns are mentioned which may be due to the rural background of the work. The constant reference to mount Sahya and the river Godā makes it almost certain that the work is the production of the country of Mahārāstra. The winds from Malaya 101 also play their part and the rivers Revā 102 and Tapī103 on the northern side get mentioned. A few names of races and tribes like Pulinda, Palāśa and the Rākṣasas from Lankā occur. Of religious and mythological allusions, we have the praise of Śiva at the beginning and at the end under the title of Pasupati which would indicate that Hāla was a worshipper of Sankara. He is once mentioned by the name of Pramathādhipa104 which also occurs in Varāhamihira. His spouse Gauri and her temple is also referred to, while she is known by her popular names Ajjā 105 and Avanna 106. Kāpālika07 a worshipper of Siva is also described as besmearing his body with ashes. Once Ganapati 108 is referred to, but the reading is doubtful. Visnu 109 also figures and mention is made of his three steps in the heaven while his wife Lakşmi is said to come out of the milkly ocean. The binding of Bali by him is also found in the Mahābhāsya of Patañjali. Krsna 110 often figures. His play with Gopis in the Vraja are described and his beloved Rādhā finds mention. He is called Damodara which epitomises one of the incidents of his childhood. Madana!ll and his five arrows are prominently found. A few epic references occur. The death of the lord of the Kurus112 at the hands of Bhīma in the presence of Mādhava is from the Mahābhārata while Rāma and his brother Laksmana are from the other epic?13. Buddha'14, his Sangha and their formula of greeting siddhirastu show the knowledge of Buddhism and possibly of Jainism.
But many of these references depend upon the different recensions of the work. There are at least six different recensions 115 which differ from each other not only in the number of verses but also in their arrangement Verses common to all the recensions amount to nearly 430 which can, therefore, be regarded as the oldest nucleus of the collection. The first recension is the

Page #126

117
Mähäräştri Language and Literature
common version which lies at the basis of the four commentaries and many Mss. of the work and called the Vulgate. It commented upon by Kulanatha, Gangadhara and Pitambara and there is also an anonymous comment on it. The second version accompanied by another commentary agrees with the first in many respects except in case of the last Sataka where it differs greatly from it. The third one is much nearer the Vulgate but differs in the arrangement of the gathas and the choice of the readings. The fourth. recension on which is found the commentary of Sadharanadeva and the fifth preserved in Telugu Mss. give very different appearance. The sixth one agrees with the first three in all respects as regards arrangement but shows different contents. Even then, from the quotations found in the works of poetics which are absent from all the recensions it appears that the work was also found in other recensions as well. All these divergences can only be explained by supposing that the collection was originally a self-standing but much smaller one and was later expanded and supplemented by successive copyists and redactors who added verses from different sources which they liked and appreciated. But then we have to suppose that the collection got its present name much later.
The third verse at the beginning says that the present collection of the 700 gāthās is taken out of a crore of such verses by Hāla and all the recensions have the same verse. According to Hemacandra Hala is the same as Sātavāhana, Sälähana or Salivahana one of the Andhrabhṛtya kings of Deccan. One tradition says that the goddess of learning Sarasvati once lived in his house for a day and half and inspired all the inhabitants of the town even including the elephants and horses so that they composed Prakrit poems out of which Häla made the present collection of 700 gāthās.
It is obvious that the present selection is taken out of a rich literature lying at its background and current at the time. Unfortunately we have no very authentic tradition preserved to us in giving the names of the poets to whose credit the gathās are to be assigned. The commentators have preserved no doubt names but its value is greatly lessened by their disagreement. The Vulgate mentions 112 poets while Bhuvanapala has as many as 384 names who contributed to the anthology. Even in individual cases the four commentators of the first recension do not agree in attributing the verse to a particular poet which makes their information of no great value if not absolutely worthless, as thought by Keith. Pischel117 has on the other hand taken a sympathetic view in accepting it with some reservations. The names preserved include salivahana, Hāla, Vodisa, Triloka, and a few names of ladies like Revā, Nāthā, Prahată etc. If we believe

Page #127

118
Amrita
a verse in the work to be authentic we have to say that each verse in the collection bears the name of the poet, which, however, is not the case at present. Even then, we need not suppose that the names are pure fiction and invention of the commentators. Of the few we know something from other sources as well. Harivuddha and Pottisa are referred to by Rājasekhara in his Karpūranañjari in close association with Hāla while Bhoja quotes a verse in which Hāla is put as an ancient poet along with Haricandra. Pālia is better known as Padalipta while Aparājita is said to be the author of Mrgānkalekhākathā. Pravarasena is famous. Even though these poets are well known and must have written in Prākrit the allotment of the verses from the anthology to them remains doubtful.
It is difficult to evaluate the exact nature of the work of Hāla, the redactor of the collection. It is pointed out that all the verses bear a good deal of similarity with each other which would lead one to think that Hāla played not only the role of a redactor but also something of the nature of a composer and reviser who brought all the previous verses to a particular standard and a definite form!18. This, it is suggested, is the import of Bāna's statement and the meaning of the third verse. Hāla not only collected these verses but gave them a literary form. From this fact and from the evidence that only a small number of verses are common to all the recensions, Keith suggests that Sattasai was originally no more an anthology but a collection of verses largely his own and refashioned by him which in course of time by being extensively interpolated got its present shape of an anthology and suffered much in its individuality.
The Indian tradition dating as early as the time of Bāna and Dandin attributes it to Hāla who is identified with the king of the Andhrabhrtya dynasty which ruled over Deccan with their capital at Pratīsthāna from the third century B.C. to the middle of the third century A. D. Hāla comes in the middle and is said to be the 17th king and is assigned to the first century. This procedure, Keith calls as mechanical and without any foundation. From the Purānic tradition and the inscriptions at Nāśik we know that Sātavāhana was the name of the family and its greatest prince Gotamīputra Satakarni is assigned to 319-340 A. D. by Bhandarkar. According to these traditions the whole dynasty lasted for 456 years. The Mastya assigns Hāla to 297 years after the beginning of the dynasty. Lassen has put him about 276 A. D. Haraprasād Shastri 119 remarks that Hāla cannot be placed later than the first century A. D. and the mention of a Vikramaditya in the Sattasai should be syncronised with the founder of the Samvat era. There are a few facts which favour this conclusion. From Brhatkathā and other traditions preserved in Kāmasūtra and Kavyamīmāṁsa we know that

Page #128

Māhārāştri Language and Literature
119
the Sātavāhanas were in favour of Prākrit which is corroborated by their inscriptions at Nāśik. The geographical background of the work also points to the country where these kings ruled.
Weber120 however points out that in verse 467 we find king Sālāhana mentioned as a great person which makes it impossible that the work is contemporary with him. It is also doubtful whether both Hāla and Sālāhana are identical. From the internal references Weber thinks it clear that the anthology cannot be earlier than the third century A. D. This conclusion, Weber thinks, is confirmed by the political, geographical, religious and other allusions found in the work and is further corroborated by the evidence of the language. He points out that the two words Vandī:21 and Horā, the word Angāravāra123 to mean Tuesday all showing acquaintance with Greek astronomy show that the work is late. The absence of the name of Kālidāsa in the list of the names of the poets, he is inclined to interpret, as favouring this date. But much of the force of his arguments is lost, when we remember that Sālāhana is the name of the Dynasty and that the anthology has suffered greatly while attaining its present form.
Jacobi, on the other hand, had identified this king Hāla with the Sātavāhana king of Pratisthānapura who, at the advice of the sage Kālika, changed the date of Pajjūsaņā of the Jains in 467 A. D. This date agrees well with all the allusions in the work. But it is very doubtful whether the Jains are right in putting the Sātavāhana king so late and their tradition itself is not of great value. Keith stresses the point that the language of the present anthology is later than the dramas of Aśvaghosa and the inscriptions of the early Christian era, which would induce him to place the production of the poems between 200 and 450 A. D. and their putting together in the form of an anthology somewhat later. From the evidence that is available it is most probable that the work belongs to the first century A. D. and is closely associated with the S5tavāhana kings.
From the majority of the verses in the collection it is evident that Hāla collected his pieces from a literature essentially of the same nature, s verses dealing with complete situations. But even then a few gāthās would indicate that they are taken from popular tales like that of a lady put in captivity but waiting to be released, a woman captured by thieves or the unchaste woman who pretends that she is bitten by a scorpion because she wants to go to the house of the doctor whom she loves. But even this cannot be said to have some context as the situations can be easily imagined for the purpose of understanding the verses.

Page #129

120
Amrita
It is usually difficult to judge the style and poetic excellence of an anthology. The gātās being from very varied hands are of very diverse characteristics. But what strikes one in this case is not their differences as their similarity which is explained on one theory as due to the redaction of Hāla. Speaking generally we find the style of this collection homely and simple and stands in glowing contrast with the style of the later epics like Setubandha and Gaüdavaho. Very scarcely have the authors taken recourse to scholarly devices of making good poetry, such as play on words and recondite allusions. The poets here appear to express their ideas in as direct a manner as possible and this gives their work a peculiar charm. It has a very close connection with the realities of life and a still closer association with natural things which can scarcely be met with in Sanskrit works. They show a certain amount of frankness and rough good sense in dealing with love, while quaint expressions of women would indicate the provincial mode of expressing things. The language is purely Māhārāstri with little mixture of other Prākrits. The work, however, abounds in many Deśī words and the majority of them are found in modern vernaculars especially Marāthī. THE HARIVIJAYA OF SARVASENA
Sarvasena appears to be one of the ancient writers in Māhārāstri and of considerable repute. Unfortunately we know nothing about him or his Prākrit epic Harivijaya. All that we can get to know of him is from later references in the works of Sanskrit rhetoricians. Of them the Dhvanyālokal24 is the oldest to quote from it. A few quotations are also found in the two works of Bhoja, the Srngāraprakāśa 125 and the Sarasvatīkanthābharana 26. In all we have some ten or eleven verses from the work.
The verse from the Dhvanyāloka describes the beginning of the spring in which the god of love takes hold of the face of the goddess of vernal beauty. Another describes the beginning of anger in the heart of Satyabhāmā, one appears to embody the words of Krsna consoling his beloved who had grown angry for the sake of the heavenly flowers, a third describes Rukmini who was greatly delighted at the sight of her husband even though the occasion was one for getting angry. In a verse anger pervading the face of Satyabhāmā is described as charming like the spot on the disc of the moon, while another delineates her face on the verge of getting delighted and the anger passing off being overcome with delight. A few verses give us the reconciliated state of the heroine when her desire of getting the heavenly tree was fulfilled.
From these few quotations it is sufficiently clear that the subject of the epic was the famous episode in Krsna's life, his conquest of Indra for the sake

Page #130

Māhārāștri Language and Literature
121
of the flowers of the heavenly tree Pārijāta to satisfy the desire of his beloved Satyabhāmā who was roused to jealousy on account of his having given her Co-wife Rukmini a garland of the flowers of the same tree. As all the quotations are found in one and the same metre it is but natural to suppose that the author wrote his epic in the Skandhaka as is the case with the work of Pravarasena, even though it is just possible that others were sparingly used. From a remark of Abhinavagupta127 in his comment on the Dhvanyāloka it appears that this epic contained the episode of the taking off of the Pārijāta tree as something that is invented by the poet though not found in the sources, and being made subservient to the purpose of reconciliating his wife. This would make one think that the main theme of the epic was something different and of greater importance.
In the absence of any more evidence it is not possible to determine the date of our author. If we believe the statement of Bhuvanapāla128 that many verses in the Sattasai are from his pen, we will have to place the author fairly early. But much reliance cannot be placed on him unless corroborated by other evidence. At any rate he is earlier than Anandavardhana and must have lived a few centuries before him.
For all that we know Sarvasena appears to have written in a charming style marked with great delicacy as remarked by Kuntala129, his work being an illustration of the Sukumāramārga in poetry and this is fully borne out from the few quotations we now come across. His language is easy and less complicated than that of Pravarasena. But he also shows the use of long compounds and poetic figures. He is also fond of giving gnomic sayings and in all these things there runs a close parallelism between his work and the Setubandha, producing a strong suspicion of one being the model of the other. Yet their precise relation would remain doubtful unless we are able to determine the relative chronology of the two authors which it is not possible with the material at present available. THE SETUBANDHA OF PRAVARASENA*
The Setubandha of Pravarasena 130 is one of the early epics written in this Präkrit. From the text itself we come to know that the author called the work Dahamuhavaho 131 or Rāvanavaho 132 and it is marked by the presence of
* After the article was written, Dr. V. V. Mirashi has convincingly proved that the author of Setubandha was a member of the Vākātaka family of Vidharbha, connected with the Gupta dynasty, and hence the work has not at all connected with the Kashmir.

Page #131

122
Amrita
the word Anurāa at the end of each chapter as do many of the Mahākāvyas in Sanskrit 133. But from the references to our work by such early writers as Bāna 134 and Dandin who lived soon after the poet we know that the epic also had the title Setubandha which was popular enough to admit of a pun. That name is no doubt applicable to the work with sufficient accuracy because the major portion of it is devoted to the description of the building of the Setu or the bridge over the ocean by the monkeys, while the historical allusion to the building of the boat-bridge over the river Jhelam135 by the author must have helped it to get greater currency. The original name Rāvanavaho must have been also current for a long time as can be inferred from the fact that the Prākrit epic of Vākpatirāja styles itself Gaudavaho no doubt after the manner of the present one.
Tradition attributes this work to a king called Pravarasena and this is borne out by many facts. Both Dandin and Bāņa refer to him as its author, and still later Kșemendra136 quotes a verse as being the composition of Pravarasena which is to be found in the Setubandha. At the beginning of the work we find the expression 'begun by a king'137 applied to the book.
But some doubt is cast on this tradition by the curious colophon found at the end of each Āśvāsaka which runs "ia siripavarasenaviraie kālidāskae mahākavve pāncaraho asāsao parisamatto”138, while the commentator Rāmadāsabhūpati who lived in the 16th century refers to this tradition of the composition of Kālidāsa139. But his further remark that Pravarasena may be the same king as Vikramāditya or even Bhoja shows the extremely confused nature of this tradition even at that early time and so appears to be a pure later invention to father one of the greatest Prākrit epics on the famous poet Kālidāsa. It is impossible to explain if the work really belonged to Kālidāsa, how Bāna never knew this fact and how on the contrary he definitely attributed it to king Pravarasena unless there was some genuine tradition current at his time. From verse nine it is just possible to argue that the work was the composition of some court poet of Pravarasena and was made current by the name of the patron 140. This will also explain to a certain extent the fact that the work was later attributed to Kālidāsa as soon as the real author was forgotten. At any rate both the style marked by long compounds, elaborate puns, strained imagery and the form of the work showing no skill in the arrangement and heedless of repetition both of ideas and of situations would make it extremely improbable that it is from the pen of Kālidāsa.
Whoever be the actual writer of the epic, and we have no means to

Page #132

Māhārāștri Language and Literature
123
know him, it is clear that it was composed at the time of Pravarasena with whose name it has become early associated. Naturally the date of the work can be ascertained by the determination of that king's reign. In the early history of India there are four Pravarasenas belonging to two different dynasties. Two of them belong to the list of the Kaśmīrian kings and are mentioned in the Rajatarangini of Kalhana141. The other two belong to the less known Vākātaka family which flourished in Deccan in the 4th and the 5th centuries 142
It appears very probable that the present epic was written in the reign of one of the Kaśmīrian kings and particularly in the reign of the second Pravarasena of that line. According to the calculations of Rājatarangini which Pandit 143 accepts as true and defends at great length the first Pravarasena ruled from A. D. 58 to 88 A. D. and the second from A. D. 125 to 160 A. D. If we are to accept these dates as true the work would fall in the second century A. D. at the latest. But according to the modified calculations of Stein 144 who is followed by many others in this respect the second Pravarasena is to be assigned to the 6th century and consequently that will also be the probable date of the epic. From the style of the poem and its close dependence on Sanskrit literature of the 6th century more than of the second we think the later date to be nearer the truth.
There is another tradition which is referred to by Rāmadāsa at the beginning of his comment, according to which the work is put to the credit of Kālidāsa who wrote it for king Pravarasena at the behests of Vikramāditya Chandragupta the second. In recent times 145 this theory is revived and new evidence to substantiate it is sought for. Great stress is laid on the fact that the curious colophon at the end is found in all the Mss. of the work so far available and therefore must possess some old and genuine tradition. A verse]46 from the first chapter also suggests that the work was not the composition of Pravarasena alone; he only began it and was completed by somebody else. Now in a tradition Preserved in a work called Kuntalesvaradautya 147 of which we have a few quotations only, there is mention of king Vikramāditya who is no other than Candragupta the Second of the imperial Gupta dynasty as being connected with the Vākātaka family as his daughter Prabhāvatīguptā was given in marriage to Prthvīsena the First. It is further reported that kālidāsa was sent by him to this king as an ambassador and must have helped Pravarasena the Second, the sun of Rudrasena in composing the present epic, a fact which would explain the colophon and the tradition preserved by the commentator.

Page #133

124
Amrita
There are real difficulties in accepting this theory of the authorship of Kālidāsa in collaboration with the Vākātaka prince Pravarsena. It is more than doubtful whether the work Kuntalesvaradautya had any historical veracity to base ones conclusions on it. We know too litle of Kālidāsa to judge his character as an emissary in bringing the two royal families together. The identification of Devagupta, the father of Prabhāvatīguptā with Candragupta, though usually accepted, is not beyond doubt and the few glimpses of the lost work do not give us the tradition clearly. The nature of the language and the inordinate desire of the author to form Yamakas and the older tradition preserved by Bāna would force us to reject this suggestion.
The work is handed down to us in three different recensions which do not differ much from each other. The most authentic one is the one preserved by Rāmadāsa who wrote his commentary in A. D. 1596148 at the time of Akbar. Another recension is preserved by the commentator Krsna who is later than Rāmadāsa as is clear from his introductory remarks and his criticisms of his explanations. The difference between the two recensions lies in the fact that while Rāmadāsa has 15 Āśvāsakas, Krsna has 16 which he makes up by splitting the 13th chapter into two at verse 67. This arrangement, however, is not very consistent as the characteristic word Anurāa is not found in that verse. The third is not exactly a version of the text but a free Sanskrit rendering prepared by one Sivanārāyanadāsa at the request of one Rāmasiṁha149. It is not a word-for-word translation but follows the original verse-for-verse150. Even the extra verses found in the different manuscripts come to only 14 and show signs of genuine composition.
The author opens the work with a salutation to Visnu who is described as killing various demons like Madhu, Hiranyakaśipu, Aristha and others and his exploits like the bringing down the Pārijāta tree from the heaven. He also offers his obeisance to siva, his laughter and his dance. A few remarks on poetry and its difficulties are added and the author expresses his intention of composing his epic Dahamuhavaho 151.
The story opens with the description of Rāma whose grief is aggravated by the coming up of the autumnal season and his lamentation for the loss of his beloved wife Sītā, who is taken off by the demon Rāvana. The poet takes this opportunity of describing the Sarad at great length. At that very time Hanumanta comes back successful from his mission of searching Sītā and relates to Rāma the long awaited news about her and gives in token her crest-jewel. Rāma is greatly delighted at its sight but slowly his joy turns

Page #134

Māhārāştri Language and Literature
125
into rage towards his enemy Rāvana, whom he now wishes to conquer by marching against Lankā. He crosses the mountain Vindhya and the army of the monkeys arrives on the shore of the ocean (i).
Rāma now views the vast ocean stretching before him and the poet picks up the occasion to display his power in describing the ocean in an elaborate manner which covers the major portion of the Āśvāsaka 152. The effect of this obstacle in their way is very different on different persons, Rāma looking at it without much concern, Lakşmaņa not losing his courage, and Sugrīva looking at his army of the monkeys. But, the monkeys are disheartened at the unexpected difficulty and look admiringly towards Hanumanta who has crossed and recrossed it (ii).
Sugrīva now views his drooping followers and to put courage in them delivers his fiery speech in which he points out to them their duty and shows them the necessity of overcoming the obstacle, as death is far more preferable to infamy. This speech, however, has no effect on the army. Another speech is made in which he promises to overcome the difficulty himself if his followers are not willing to take up the risk which enkindles the fire of heroism in the army. The whole of the chapter abounds in good many moral sayings 153 and maxims with appropriate illustrations from nature, which does credit to the author and more than justifies the remark of Dandin that it is a veritable ocean of good sayings (iii). Now gets up Jāmbavat, the oldest in the army, with a long line of experience behind him, on which he harps so often, and who now tries to combat the agitation in the army with sober words, importing a slow and thoughtful action on their and on Sugrīva's part and gives a real appreciation of the magnitude of the danger. He hints that Rāma is fully capable of subduing the ocean and suggests that the ocean itself should be requested to give a free path on its waters of his own accord. At this moment there arrives Bibhisana from the sky, whom Māruti easily recognizes to be a friend and carries him to Rāma who greets him with praise and suggests that he will be made the lord of Lankā (iv).
The fifth chapter opens with the description of Rāma again suffering from the pangs of separation in the moon-light night, and early in the morning he makes his mind to chastise the ocean. He then hits it with his arrow which emits fire and thousand of other arrows from itself and plays havoc on the ocean and its acquatic animals. The author devotes the whole of this chapter to the graphic and minute description of Rāma's shooting the arrow and the plight of fishes, serpents and the surging up of the waves. The

Page #135

126
Amrita
minuteness of the picture can be well imagined from the fact that 15 verses 154 are taken up by the mere shooting of the arrow (v). Being hit strongly the ocean comes up in a human form and along with his spouse Gangā falls at the feet of Rāma whom he praises and argues that it is but at his bidding that he has kept the boundaries. He further suggests Rāma to build a Setu over its waters to cross them with ease. Rāma accordingly gives orders to build a bridge. The monkeys begin to pick up mountains to fill in the cavity of the ocean and a detailed description of it occupies the remaining Āśvāsaka (vi). Now the monkeys begin to work out the idea and throw big mountains in the ocean which make the water surge up high and disturb the ocean enormously. But after a long and strenuous exertion the ocean remains unfilled and the monkeys begin to feel a little exhausted (vii). They now stop their work for a moment as it is of no avail, and heap the mountains on the shore. Meanwhile the ocean regains its calmness. Sugrīva now requests Nala to build the Setu so that Rāma should not find it necessary to string his bow once more and use his arrow for a second time. Nala promises to do the work and with the help of the monkeys succeeds in constructing the bridge which is now described at great length. The army now passes over the ocean with ease and arrives on mount Suvela on the other side. At the news of this wonderous feat of the enemy the demons lose their faith in Rāvana (viii). The next āśvāsaka is devoted to the description of Suvela in which the author shows his skill in the use of figures of word like Yamaka and Anuprāsa 155 which are taken to be signs of the poet's ability of wielding the language as he wishes (ix). Now follow the usual things to be described in a Mahākāvya like the setting of the sun, the approach of the evening, the spreading of the darkness, the rise of the moon and the sports of the ladies, love and its enjoyments (x).
The scene is now transferred to Lankā where we see Rāvana suffering from the fever of love towards Sītā. He thinks of all possible ways of winning her over and finds no suitable means. At the end he hits upon a plan and calls his servants. He orders them to show to Sītā a false head of Rāma which they produce by their magical power. Sītā is described in her pitiable condition when they show her the false head of Rāma and swoons immediately from which she recovers only to fall into another. Then follows her pathetic lamentation and the consolation offered to her by her demon - attendant Trijatā who points her out the falsity of the situation. This forms the best passage 156 in the whole epic (xi). The next āśvāsaka opens with the description of the dawn when Räma gets up and the army of the monkeys

Page #136

Māhārāștri Language and Literature
127
is prepared to fight. Rāvana also is ready and his army is arrayed. While the demons are girting themselves up the monkeys lay a siege to Lankā. Now follows the encounter of the two armies on the battle-field (xii). The battle continues for a long time with victory inclining towards one side or the other in turn and the heroes of both the parties meet in individual duals and Angada defeats Indrajit (xiii).
. On the field now come both Rāma and Laksmana and there arises a fierce combat between them and Indrajit, who however, overcomes them for a short time with the missile of the serpents. Both of them fall in a swoon to great grief in the army of the monkeys. But they soon recover and Rāma invokes Garuda to ward off the serpents. Individual combats between the heroes of both the armies go on in the meanwhile (xiv). One by one the leaders of the demons begin to fall. First Kumbhakarna is killed, Indrajit follows him and Rāvana, enraged at the loss of his brother and son, hits Laksmana with his irresistible power. Laksmana falls in a swoon but soon recovers. Now follows a fight between Rāma and Rāvana wherein the latter is defeated and killed. Bibhisana laments the loss of his brother. After the battle Rāma returns, back to Ayodhyā along with Sītā who is purified in fire to satisfy the love of Bharata. The author then closes the epic which he names Rāvanavaho and in which he has used the word Anurās as a characteristic mark (xv).
In this epic Pravarasena shows all the equipments of a classical scholar composing a Mahākāvya. In the third and fourth chapters he shows his knowledge of the royal policy and depicts a lovely controversy about the respective merits of following the direct policy of proceeding against the difficulty and the sober method of inventing some device to overcome the same. Sugrīva asks the monkeys not to wipe off their fame which he compares to a good person who is somehow come to make a request, both being difficult to be obtained once more 157. It is the duty of the lord to order while the real burden of the work falls on the servants158 Battle and its hardships alone are able to distinguish between true heroes and those who pretend to be so 159. The words of Jāmbavat are equally weighty. He points out that dejection carries off fortitude, youthfulness removes modesty, and love takes off the sense of shame 160 What warriors in union can do can never be done by them individually which he illustrates by the fact that one ray of the sun may torment at the most, but all of them are able to burn the three worlds 161 Energy misplaced proves neither effective nor dreads the enemy i the mistaken arrow shot in rage neither hits the mark nor produces fear162.

Page #137

128
Amrita
He is equally happy in describing the state of love in separation. The lamentation of Sītā at the sight of Rāma's head reaches the high water mark of the poet's ability in producing the sentiment of pathos. He is equally clever in giving a situation in as few words as possible. Sītā was downcast when she looked at the false head, began to tremble when the demons pushed it before her, and instantly swooned when they said it to be Rāma's head 163. Rāma's love towards Sītā is finely given 164. Rāma valued the southern quarter because she lived in that direction, the moon because she admonished her in her separation, the ground because she sat on it, and the sky because she was carried through it. A fine contrast is effected in the description of the words of the ocean265 which were soft but able to bear the weight of their import, few but full of meaning, lowly but lofty with courage, obliging but true in sense. Equally concise is the description of the meeting between Rāma and, Hanumanta, when told that he has seen her (Sītā) he did not believe it, when told that she was emaciated he sighed slowly with tears, when told that she laments he cried, and when told that she lives he embraced Māruti 166 Sugrīva, when he sees that the bridge is not at all visible even when so many mountains are thrown in it, requests Nala in plaintive terms 'the monkeys are exhausted, the earth has only a few and far off mountains, nor is any trace of the bridge to be seen, but let not the mighty bow of Rāma be bent once more'167
The work is divided in 15 Āśvāsakas and contains on the whole 1362 verses. The prevailing metre is Skandhaka, a mātrāvrtta having 32 syllables in each of its two feet and divided into four feet of four mores each. At the end of an Aśvāsaka and sometimes even in the middle other metres are found used most of them found in Prākrit poetry alone. Next to Skandhaka comes the Galitaka168 occurring 32 times and Anusthubh 169 only four times, while there are three metres 170 unknown from other sources.
The work is beyond doubt written in the Māhārāstrī Prākrit and quotations from it in the grammars of Hemacandra 171 and others go to support it. Dandin explicitly says that the Setubandha is written in it. This is also in conformity with the rules of the rhetoricians who assign this Prākrit for the writing of the epics. In one point however, the text shows forms which are not in agreement with the rules of the grammarians. Goldschmidt says 172 that here and in cases like them we should give greater weight to the authority of the Mss. than the rules of the grammarians. It is true that Hemacandra's grammar is older than all the Mss. of the work so far available to us, and more so in case of his predecessor, but it is doubtful whether we

Page #138

Māhārāștri Language and Literature
129
should follow them completely in restoring texts older than themselves. In many cases where there is a discrepancy between the readings of the texts and the rules of the grammarians it is usually the mistake of the copyist but in few the Mss. tradition is better and such appears to be the case in forms like udu, dāva, mailida, vivannada, and Rāmādo. It is worth noting that in all these cases Vararuci173 gives these forms as correct even though Hemacandra and others would bar them.
As remarked by Winternitz 174 the present work is written for a public which was well versed in Sanskrit. Naturally the language is greatly influenced by the Sanskrit style and conventions. Out of the three-fold division of the Prākrit vocabulary the major portion of our work is made up of Tatsama and Tadbhava words, the latter far exceeding the first. But the author is also influenced by the popular literature before him and so a small number of Desi words is found in the epic. If we accept further the fine distinction between the Dhātvādeśas and the Deśī words which consists in the fact that while the former can be traced to a great extent in the vocabulary of the modern Indian languages the latter do not admit of such a ready recognition, we have în Setubandha the words of the former class only 175.
The style of the epic is strongly influenced by the style of the famous Sanskrit epics. As a result we find the use of long compounds and elaborate puns to be met with on every page. The author cares more for the Yamakas and Anuprāsas and figures like Upamā and Utpreksā are based on Ślesa 176. Much need not be made of such defects as almost all of them were highly valued at that time and it is but natural that the writer should be guided by the taste of his days and should try to show his skill in excelling in them. Even then the epic shows considerable power of poetic merit and comes up to the rigorous standard of a Mahākāvya. Pravarasena's style is full of beautiful expressions, charming alliterations and strikingness of thought and imagery. With all the artistic pomp in style and language we find real sentiment running as an undercurrent throughout the book.
There arises the question of the use of prākrit in a work of this nature. It has the same style and artistic nature as that of Sanskrit epics and it is difficult to believe that it was written for some other public than the one to which Sanskrit epics were addressed or to regard it as strictly popular works. It is just possible to think that in the court of king Pravarasena Sanskrit was of less concern than Prākrit. But a truer explanation would be to suppose that the author thought of showing that one can overcome all the

Page #139

130
Amrita
difficulties of a different language and is able to produce a work claiming comparison with famous Sanskrit epics. The epic got new interest in it at the time of emperor Akbar and his son Jehangir. THE GAUDAVADHA OF VĀKPATIRĀJA*
The Gaudavaha 177 of the poet Vākpatirāja is an epic of some historical importance. Vākpatirāja was a court poet of king Yaśovarman of whose conquest the poet supplies us information, though meagre. The work is a semi-historical poem written in the gāthā metre numbering over 1200. It has no divisions and is written in a continuous form. As the name indicates the subject matter of the poem is the defeat and death of king of Gauda at the hands of the hero, the king of Kanauj. Unfortunately the name of this Gaudian king is not preserved to us.
The poet begins with his salutations to the deities of the Hindu pantheon which includes Brahmā, Hari, his various incarnations, Siva, his spouse, Sarasvati and others 178. Then follows a group of verses dealing with general topics like poetry, former poets, wherein the Prākrit language receives high praise and some good remarks on poetry are given 179. Then begins the real theme of the work and king Yasovarman is praised with the usual lavishness of a court poet but with real poetic flight of his imagination 180. Then follows the description of the effects of this king on the minds of the ladies of the town who flock to see him issue forth out of his capital 181, his greatness leading to the astonishment of Indra 182, and his sports in water183. This portion of the poem contains the digression of the exploits of Indra in cutting the wings of the mountains which formerly flew through the sky to great inconvenience of all. The author has further introduced the description of the time of universal destruction of the world with the simple device of identifying the king with god Hari 184
Now in winter Yasovarman starts on his expedition of the conquest of the world. His sight affects the ladies of the town185 and even the damsels of the heaven with the passion of love. Then follows a long but beautiful praise 186 put in the mouth of his bards which once more repeats the incidents of cutting the wings of the mountains. His army of the horses and elephants is described at great length 187. In the Hemanta season he arrives on the banks of the river Sona, and his warriors wander at pleasure in the fields which are painted with a keen observation of nature and a first hand knowledge of the
* About the incomplete nature of the work Gaudavaho by Vākpati, it has been argued by the Dutch scholar Daalen has argued in a number of articles for regarding it as a complete work and not merely a beginning of a larger work now lost or never completed : A good summary of his arguments is given by H. W. Bodaewitg in Wiener Zeitschrift Vot. XLII 1998 pp. 41-66.

Page #140

Māhārāștri Language and Literature
131
rural life188 He then arrives on mount Vindhya where he is led by a forester to the temple of the goddess Durgā whose praise 189 forms a good hymn in the mouth of a king, and who receives a long eulogy. Then follows a very slight reference to the flight of king of Magadha who is afraid of Yaśovarman. His followers, however, offer a fight and are easily defeated 190
Then follows a description of summer191 and a brief statement that the Gaudian king is killed by Yasovarman192. He then proceeds to the shore of the ocean. He defeats the king of Vanga193, receives submission from the king of the southern country, and goes to the south ocean. At this point the author gives the description of the incidents in which Vāli takes hold of Rāvana194. Yasovarman then conquers the Parasikas in a hard-fought battle 195
The author again introduces the description of the acts of king Prthu who pushed the mountains on both the sides with the bow196. Now the king arrives on the bank of Narmadā which gives the opportunity of describing the love of the river for Kārtivīryārjuna 197. He then wanders in the deserts of Maru. He arrives to the country of Śrīkantha and a description of the sacrifice of the serpents is given 198. He laments over the hard fate of Duryodhana when he arrives on the battle-field of Kuru199.
After this long march of the conquest he returns back to his capital. This opportunity is utilised by the poet to give a long and brilliant picture of Yaśovarman's sports in water and the beauty of the ladies 200
From this main theme of the work the author again turns to the praise of a few poets of old times and the occasion of the composition of the present work201. The learned people wish to hear Vākpati compose a poem dealing with the exploits of king Yaśovarman who is no other than an incarnation of god Hari. The occasion of Yaśovarman's coronation gives rise to the description of incidents like Pịthu's stabilising the earth202, the churning of the ocean203, and many other mythological incidents. The poet promises to give out his poem in the next morning which is followed by a glowing picture of the rising sun204. The work is brought to a close with the remarks that the poet begins to give to the audience his big work dealing with the life and the conquests of the king which he requests the people to hear205.
From this brief review of its contents it is clear that the work has little historical interest and throughtout are scattered the descriptions of mythological incidents, which, on the whole, out-weigh the frame-work of the poem.
As is to be expected, the style of the work is modelled on the style of Sanskrit epics with their merits and defects. But unlike his predecessor,

Page #141

132
Amrita
Pravarasena, whose epic serves our author to be his model in naming the work, he avoids all extravagant use of puns and word-plays and all purely mechanical devices. The favourite figures of our author are Utpreksā206 and Upamā, both of which are used in a very masterly and beautiful manner, each one of which gives us a new stroke of his imagination. Another peculiarity of our author is his accurate knowledge of natural scenes and the behaviour of animals and the life of the simple country folk in which he differs greatly from the practice of Sanskrit poets 207. Moreover, the ideas of our author though based on conventional facts, show a new liveliness and beauty. In fact, as a poetic composition, the work deserves our greatest admiration and appreciation than is usually given to it.
The form of the work, however, gives great difficulty in knowing its real nature. Pandit suggested that the work was only a prelude to a more extensive scheme which the author planned to write but which he may or may not have composed208. He points out that if the present work formed the whole of what the writer intended to write, the title becomes a misnomer, because the killing of the Gaudian king is alluded to in a pair of stanzas only and is not enough to give the name to the work. Nor will it explain the expectation aroused by the end of the present fragment where the writer requests the audience to hear the great work which is yet to come. Bühler209, on the whole, agrees with Pandit and adds the argument that the present work is styled in the manuscripts gāhāvīdha which usually signifies the introduction of a work as in the case of the famous Brhatkathā of Gunādhya.
Jacobi210 and others, on the other hand, regard the work available, to be all that Vākpati wrote and intended to write. They explain the curious form of the book as due to the epitomisation probably done by the commentator who purged away all the historical matter and kept what was of permanent interest, the description of nature and the mythological incidents, which, according to him, were of greater value than the narration of the story. This will also explain the differences between the various manuscripts which differ considerably as regards their arrangement and number of verses.
The present work refers to another of the poet's composition called by him Madhumathavijaya211 or the victory of the enemy of the demon Madhu. Vākpati has himself remarked that this earlier production was the better of the two.
Unfortunately, we do not possess the poem. A single quotation from

Page #142

Māhārāştri Language and Literature
133
the work is preserved to us in the Locana212 of Abhinavagupta where we see Pāñcajanya taunting Krsna at his love-lorn condition when he finds the lotus stalks too heavy when he is able to bear the weight of the earth in his boar incarnation. Pischel213 remarks in his grammar that two other verses214 are quoted from this work in the Sarasvatīkanthābharana, which, however, is doubtful. By a look at these two verses one can see that they are not quotations from this book but from the Harivijaya as is proved by their metre. From the verse of Vākpati given by Abhinava it is evident that the lost epic was also written in the gāthā metre while these two are in the Skandhaka in which Harivijaya was composed. There are two more quotations of Vākpati in Mārkendeya's grammar215 which are not found in the Gaudavaho and may be from his other epic.
We know very little of the personal history of Vākpati himself except that he was a favourite of king Yaśovarman and patronised by him, and possibly a student of the famous dramatist Bhavabhūti. His date, of course, depends on that of his patron whose date in its turn depends upon his conqueror Lalitāditya216 of Kaśmīr. From the chronology of Rājatarangini which is modified a little by modern scholars it is clear that Lalitāditya's reign extended from A. D. 724 to 760 A. D. His conquest of king Yasovarman is more accurately determined by Jacobi to fall in the year 733 A. D. Therefore the major portion of the life of king Yaśovarman falls in the first part of the eighth century and in the same period must have been the time of Vākpati's literary activity.
THE VIŞAMABĀŅALĪLĀ OF ANANDAVARDHANA,
Anandavardhana was a court poet of king Avantivarman and is famous for his work on poetics, the Dhvanyāloka. He was also a poet writing in both Sanskrit and Prākrit. The only Prākrit work of his that we know is the Visamabanalīlā217 which appears to be a work dealing with the subject of love on the model of Sattasai. Konow218 thinks that it must have been an anthology and he is right if we take it to be wholly composed by him alone and not a mere collection. A verse219 often quoted appears to be a popular saying. From another reference 220 it appears that the work contained speeches of Cupid and his companions Youth, Spring, Malaya winds and others. A third quotation221 describes the power of good poets which knows no limit and are never repeated as are the sports of good ladies. In one verse222 Kāma is said, to conquer the three worlds and the defeat of the Asuras.

Page #143

134
Amrita
THE RĀVANAVIJAYA
There appears to be another Prākrit epic called Rāvanavijaya or the conquests of Rāvana. It is referred to by Hemacandra in his Viveka223 on the Kävyänušāsana where he also quotes a verse from it in the Skandhaka metre stating that the heart of a poet is known only by another poet. From other references there, it is evident that the work began with the praise of good poets, contained the descriptions of town, sunset, the praise of the hero, horses, elephants, and other kindred things which are the usual requirements of a Mahākāvya. The author of the epic is, however, unknown, nor is the work known from any other source. THE LĪLĀVATI
Līlāvatī224 is a short romance in pure Māhārāstrī of an unknown author. From the introductory portion of the work we come to know that he was a son of Bhūsanabhatta and grand-son of Bahulāditya, but does not give his own name anywhere in the text.
It is difficult to know the date of the author. The single Ms. of the work is dated A. D. 1208 and so must be earlier to it. Hemacandra refers in his Kavyānusāsana225 to a work called Lītāvati but there is nothing to show that the reference is to the present work. The style and the introductory verses of the work, however, make it probable that the writer appears to follow the lead of Vākpati and must have lived later than him.
The romance deals with the love story of king Sātavāhana and Līlāvatī, a princess of the Simhaladvīpa. At the beginning of the work the writer offers salutations to the Brahmanic deities like Hari, Madhumathana, Gaurī, Candi and others and then proceeds to give the personal information.
The Ms. of the work is preserved in the Jain Bhandar and naturally their orthography is found to affect the work as in the case of Gaudavaho. But there is nothing in the language to show that it is written in the Jain Māhārāstri. The poetic value of the romance is certainly high and the work deserves to be soon published 226. MĀHĀRĀȘTRĪ VERSES IN ALANKĀRA WORKS
The early extent of the Māhārāstri literature can be easily soon from the verses quoted in works on Sanskrit poetics which show abundantly the richness of this literature for many centuries. Even though we do not find Prākrit verses in the early works of Bhāmaha and Dandin they clearly state

Page #144

Māhārāştri Language and Literature
135
the existence of a vast Prākrit literature, and the absence of Prākrit verses in their works due to the scheme which admitted no quotations. The few Prākrit verses of Rudrata 228 are his own compositions. But from Anandavardhana onwards Māhārāștri verses are found in great number, the Sarasvatikanthābharana and the Kavyānuśāsana being particularly rich in them. It is also inconceivable that works like the Sattasai, Rāvanavaho and Gaudavaho should have been left without other imitations in later days, particularly so when they were deemed worthy of imitation by Sanskrit poet like Govardhana229 and others. Besides the meagre information about the lost Prakrit epics like the Harivijaya of Sarvasena and such anthologies as the Visamabānalīlā of Anandavardhana we meet with names of works and their authors to some extent. Thus Viśvanātha, the famous author of the Sāhityadarpana who lived in the 14th century, appears to have composed a Prākrit epic called Kuvalayāśvacarita of which a single verse is quoted by him. This verse describes a pair of young persons looking at each other. Another older work of the same title and probably of the same theme was known to Hemacandra230 who refers to it in his Viveka. Abhinavagupta231 quotes a Prākrit verse and attributes it to one of his teachers Bhattendurāja who thus appears to have composed some Prakrit work. Abhinava himself quotes one verse as his own composition in his Locana. Dhanika in his Avaloka on the Dasarūpa quotes many Prākrit verses of which four are his own. One verse points out that while the face, the eyes, and the youth remain the same the beauty of love produces quite a different effect232. Another verse asks a friend to look at the girl gazing at something and beautiful with affection233. At the swelling of the breasts the charming movements of a girl captured all her limbs so that her friends were doubtful of her childhood234. Another verse points out that the charming talk, sight, and all other activities of a girl excite curiosity in the minds of the co-wives 235. From all these verses, it appears, that Dhanika also composed some Prākrit anthology of erotic contents.
Anandavardhana quotes some 45 Prakrit verses in his Dhvanyāloka out of which 19 are not to be found anywhere else and so remain untraced. Many of them are stray verses and are probably quoted from some anthology or collection of moral sayings. Thus, one verse reports the words of the lover to his beloved to return back from going outside, for the lustre of her face produces obstacles in the way of other ladies going to their lovers 236. Another fine gāthā points out that love-sports lose nothing by repetition 237. A third verse says that the stroke given by the lover on the breasts of the beloved pains the heart of the co-wives238. In another verse a warrior is described as

Page #145

136
Amrita
having arms like the elephants who have broken the golden lotuses of the wishes and thus have spread fragrance of fame and which are constantly giving out ichor in the form of gifts239 A fine illustration of the figure Dīpaka is found in a verse which points out that value is imparted to night by the rays of the moon, to the pond by the lotuses, to the creaper by flowers, to the autumnal beauty by swans, and to poetry by the appreciative people240. Time is described as either full of poison or nectar or of both or without both for different people241
There are nearly 350 Prākrit verses in the Sarasvatīkanthābharana, out of which something like 170 remain untraced. In many cases we lack the exact context out of which they are taken and so they remain somewhat obscure to us. In a verse we find the arrows of Rāma producing horripilations, on the body of Rāvana, as if, the poet imagines, they were touched by the hands of Sītā242. Another verse describes the fading away of the natural unsteadiness of the defeated army with their lustre gone, and the sparks in the eyes steady and thus presenting the appearance of lamps drawn in pictures243. The anger of the proud woman vanishes slowly with the happiness of the new reconciliation and with pain at the recollection of the fault of the lover244. A verse describes how one placed ones eyes oh the breasts of the heavenly damsels with anger when he got out of his swoon with the mistaken idea that they are the temples of elephants, evidently describing a fiery hero who has newly awakened in the heaven after his death on the battle-field 245 In another verse we find Sātyakī, the charioteer of Indra, expressing the intention of his master in the words, 'Oh, Lord of the gods, if you value the friendship of Mahumaha hand over the heavenly tree, cast off the feigned ignorance and honour the Yādava people246. The names and incidents that occur in verses like these go to show that they dealt with some famous episodes in the epics and are probably taken from Kāvyas of great extent.
But, besides these, the greater amount of quotations that we meet with here are stray verses which are self-sufficient and require no context for their understanding. Their scope is very wide and they touch upon diverse topics. But the prevailing tone appears to be the delineation of the sentiment of love of the lower classes of the society in all its aspects. 'I die without my wishes unfulfilled' cries a young maiden, 'as there is not even a rumour of my intimacy with the Hālika youth.247 'What is the use of knowing whether it is true or false', cries another, there is a rumour that the daughter of the Hālika is in intimacy with the son of the house-holder.'248 The breasts of the Hālika girl are seen by her friends to be like the bunch of the Kadamba flowers full

Page #146

Māhārāştri Language and Literature
137
of pollens thrown by the lover and heaving with emotion249 A wily girl wishing to be touched by the youth cries falsely that her hand is burnt by fire 250
There are many verses giving interesting situations. While the traveller drinks water for a long time by letting drop it through his fingers, the girl also makes the jet of water very thin, both with the desire of looking at each other for a long time251. The travellers only see the eyes of the woman beautiful in all her limbs, who guards the rice-field and who wards off the birds252. Here and there some advice is given, while descriptions of nature are often to be found. The advent of the rainy season is described as, the clouds rumbling, the Nīpa buds blooming, the peacocks dancing and the moonlight obscured253. The beauty of a lady is said to put forth sprouts in the foliagelike-hands, as blooming with eyes and giving out fruits with the breasts. A blunt way of expressing love is found in a verse when the lady asks the boy, ill-taught as he is in the use of black letters, to embrace her neck and then she never minds if both go to hell254. Curious and out of the way ideas are also to be met with as when the poet is compared to a thief, with the help of Slesa; the poet puts his words with caution, the thief places his foot with care; the poet takes care of the style, the thief often looks at the way; the poet often finds it difficult to get at the desired meaning, the thief finds it hard to get wealth255.
In the Kävyānuśāsana, Hemacandra quotes some 60 verses in the text and some 20 in the commentary. Most of them are quoted in earlier works and only few new ones are found. In one we find a bull addressed to remain silent, it being a suggestion to a paramour to be satisfied 256. In another a traveller is advised to go another way as the one he is taking is infested by a woman whose snare is difficult to be broken through257. A verse gives an address of a friend to a lady to walk very carefully, lest she break her slender waist which was produced by the creator with so much exertion258 A woman says to her lover that it is not his body marked with nails nor his eyes rolling with sleep that trouble her heart as much as his lower lip unbitten 259. An ironic expression is given in a verse where the poet offers his salutation to the moon who makes the lotuses which are soft, beautiful and of pure fibre and opening at the touch of the rays of the sun, devoid of lustre 260. A lady points out that her beauty is fine, the lover's affection is strong and the circle of friends is clever, then no use to paint the feet261.
Dhanika quotes 26 Präkrit verses in his Avaloka of which 10 remain

Page #147

138
Amrita
untraced. In one we find the description of the hero whose mind wavers between two emotions of love and heroism as his wife is crying on one side and the war drums are sounding on the other262. The poet describes the love of a girl born in a noble family by pointing out that all her sports depart with the departing lover and come back when he returns263,
All such stray verses quoted in the Alankāra works produce a strong impression that they are essentially the poetry of the populace with all its : frankness and rough good common sense, lack of courtly pomp and delicacy. They lack the usual formal phrases and expressions which are so prominent in the Sanskrit literature and a kind of naturalness imparts them a peculiar charm and simplicity.
Annotations :
1. Cp. Vararuci, Prakrta-Prakāśa chapters 1-9; Hem. Prākrit Grammar VIII, 1-4,
Sutra 259; Lakşmīdhara pp. 1-246; Mārkandeya 1-8. 2. Kāvyādarśa 1, 37.
3. Besides Dandin an explicit statement to this effect is given by Ratneśvara on
Sarasvatīkanthābharana p. 227 "divyā sanskrtā/praktä mähārāstri 4. Rudrața, Kāvyālankāra, 2, 12; Laksmīdhara p. 1-2. p. 4, p. 26. 5. Var. 12, 32. 6. Hem. VIII, 4,286. Lakşm. pp. 252, 256. 7. A few remarks of Hem. make it clear that the basis of his Ārsa is both Sanskrit
and Māhārāstrī. Cp. 2,146; 3,162. 8. Ghosh, Māhārästri a later phase of Sauraseni, Journal of the department of
Letters, University of Calcutta, XXIII, 1933. 9. Compare Var. 1, 2; Hem. 1, 44. Var. 1, 3; Hem. 1, 46; Var. 1, 4; Hem. 1, 66.
Var. 1, 5; Hem. 1, 57. Var. 1, 10; Hem. 1, 67. Var. 1, 11; Hem. 1, 72. Var.
1, 12; Hem. 1, 85 and so on. 10. Compare Var. 12, 13; Hem. 4, 260; Var. 12, 9; Hem. 4, 271; Var. 12, 19; Hem.
4, 272. 11. Cp. Var. 2, 8; 12, 4; Hem. 1, 206; 1, 207. 12. Var. 12, 6; Hem. 1, 128. 13. Var. 3, 5; 12, 6; Hem. 1, 56; 2, 83.

Page #148

Māhārāștri Language and Literature
139
14. Var. 12, 30; 15. 17. 18. 8, 69. Hem. has no rules for sau.
15. Ghosh The Date of Nātyaśāstra. Jou, of the Dept. of Lett. 1935. 16. Nātyaśāstra K. S. S. p. 214. 17. Kāvyālaňkāra 4; 11, 14, 17, 20. 18. Dasarūpa 2, 65. 19. Kävyamīmāṁsa G. O. S. No. 1, Introduction p. xxv. 20. Kavyamīmāṁsa p. 6. 21. Ibid p. 51. 22. Ibid p. 51. 23. Ibid p. 33. 24. Ibid p.p. 54-55. 25. 1, 10. 26. Act i, 4
27. It is into Bhāsā, Vibhāsā, Apabhramsa and Paicāci. 28. Kāvyādarśa w. 34, 38. 29. Ibid v. 32. 30. Kävyamīmāṁsa p. 66. 31. Hem. 4, 286: 32. Hem. 4, 302. 33. Kavyānusāsana p. 228.
34. Ibid p. 337. 35. Sattasai 1, 3; Gaudavaho 188, 36, 319; Setubandha 1; 2, 12, 41; 2; 1, 24, 44;
3; 6, 7, 57, 4; 23; etc. Visamabānalīlā; tālā jāanti, on 3, 65. 36. Sarasvatīkanthābharana KM. 94, pp. 140-149. w. 13, 14, 16 are copied from
Rājasekhara. 37. Ibid p. 145. 37a. Linguistic Speculations of the Hindus. 38. Pischel Gra. Pra. Spra. Set. 9; Jacobi and Gune in their introductions to the
editions of Bhavisattakahā.

Page #149

140
Amrita
39. Var. 4, 18-20; Hem. VIII, 1, 31-36 : Laksm. 1, 1, 49-53. 40. Chakravarti; Linguistic Speculations of the Hindus p. 455. 41. Var. 1, 18; 1, 2; Hem. 1, 101; 1, 44. 42. Hem. 2, 98-99. 43. Var. 2, 2; Hem. 1, 177. 44. Var. 2, 27; Hem. 1, 187. 45. Var. 2, 42; Hem. 1, 228-229. 46. Katre; Treatment of R in Pāli. Annals BORI. XVI, pp. 200-201. 47. Hem. 1, 1. 48. Hem. 1, 182; 1, 57; 1, 188; 1, 264; 2, 260; Var. 2, 7. 49. Hem. 1, 254; 1, 96. 50. Hem. 1, 243. 51. Hem. 1, 239; 1, 259; 1, 234; Var. 11, 16; 11, 19. 52. Hem. 1, 178; 1, 46. 53. Hem. 1, 53; for Vrācada Ap. He. 2, 80. 54. Hem. 1, 171; 1, 262; 1, 263; 1, 242. 55. Jacobi; Bhavisattakahā p. 62 of the introduction. 56. Cp. Sauraseni Prākrit J. U. B. III, Part VI, pp. 47-52. 57. Brugmann, Vorgleichende Grammatic pp. 413-417. 58. Hem. 3, 90-117. 59. Bhandarkar; Wilson Philological Lecturers. 60. Pischel; Gra. Pra. Spr. Set. 6. 61. Hornle; Introduction to Canda's Prākrta Laksana B. I. 62. Grierson; Linguistic Survey Vol. VII, pp. 5-7. 63. The big list given by Hem. of the Dhātvādeśas in his 4th chapter reveals many
Marathi words. Dr. Vaidya has collected them in one of his articles. 64. Bhandarkar; Early History of Deccan. 65. Pāṇini 4-2-133. 66. Pāņi. 4-1-176. 67. Pāni. 4-1-171. 68. Pāņi. 4-1-170.

Page #150

Māhārāştri Language and Literature
69. "Pāņi. 4-1-173. 70. VII, 34, 9. 71. Ait. Brah. VII, 17, 19. 72. Suttanipāta V. 73. On pānini 4-1-108, 175. 74. Rock Inscription No. 2. 75. Carmichael Lectures 1918 pp. 1-41. 76. Ancient India p. 5.. 77. Shorter Cambridge History of India p. 90. 78. Ed. by Kern 10, 8. 79. Ed. by Durgāprasās p. 131, 154. 80. Itihasa ani Aitihăsika lekha p. 290 foll. Introduction to the
Rādhamadhavavilāsacampū set. 84 foll. 81. Ancient Geography and Civilisation of Mahārāstra JBBRAS. Vol.XXI. 82. Prācina Mahārāstra, śātavāhanaparva, Summary pp. 12-16. 83. For the the whole problem cp. Grierson, Ling. Sur. Vol. VII, p. 15; D. R.
Bhandarkar, Carmichael Lectures 1918 Lect. 1, pp. 1-41; Pendshe, Mahārāstrācā
Sānskrtika itihāsa; Chap. I. 84. Kāvyamīmāṁsa p. 50, Sarasvatīkanthābharana p. 143, and the introductiry of
the Brhatkathā. 85. Hillebrandt and others. 86. History of Sanskrit Literature. Part 1, Classical Sanskrit Literature pp. 20-21. 87. Ed. B. S. S. p. 2 88. Introduction to Prākrit p. 189. 89. Sāhityadarpaņa p. 173. 90. Ling. Surv. Vol. VII, p. 15. 91. Kävyādarśa 1, 38. 92. Kävyālankāra p. 171. 93. Sarasvati p. 144. 94. Journal Asiatique 1872. 95. Indian Antiquary Vol. XXX, 1901.

Page #151

142
Amrita
96. Ling. Surv. Introductory.
97. Weber first edited the 370 verses under the title Uber das Saptasatakam des
Hala. Leipzig 1870; Some additions were made in Z. D. M. G. 26, pp. 735 foll. Later on a complete edition was given as The Saptasataka des Häla. Leipzig 1881. About Bhuvanapala's recension Indi. Stu. 16. With the commentary of Gangadharabhatta edited in KM. 21. A new edition by Mathuranatha with his own Tīkā.
98. Harṣacarita v. 13.
99. Geschichte der indischen Literatur pp. 97-104 of Vol. III.
100. Keith, History of Sanskrit Literature pp. 223-235.
101. 97,443.
102. 579.
103. 239.
104. 448.
105. 172.
106. 467.
107. 408.
108. 403.
109. 411, 338, 406, 425.
110. 657, 112, 87.
111. 545, 549.
112. 443.
113. 35.
114. 308.
115. Weber p. XXVII.
116. Abhidhanamālā 712 Desīnāmamälä 294.
117. Gra. Pra. Spr. Sect. 13.
118. Winternitz. p. 101.
119. Epigraphica Indica. XII, 320.
120. p. XXIII.
121. 435.
123. 267.

Page #152

Mähäräștri Language and Literature
143
124. KM. p. 127, 148. 125. 22, 43; 22, 79; 23, 73. 126. pp. 647, 655, 673, 678, 724, 727. 127. Dhvanyāloka p. 148 Com. 128. Weber's edition of Sattasai. 129. Kane; Introduction to Sāhityadarpana. p. LXXXVI. 130. The first two chapters were edited by Paul Goldschmidt. A complete edition
was brought forth by Siegfried Goldschmidt with a German translation and a word index in collaboration with Paul Goldschmidt. Strassburg 1880-1884. It is edited in Kavyamālā No. 47 with the Sanskrit commentary of Rāmadāsa and a
Sanskrit rendering by Sivadatta and Parab, Bombay 1895. 131. 1, 12. 132. XV, 95. 133. cf. XV, 94-95. Kirāta has the word Laxmi and Māgha the word Śrī. 134. Harsacarita, Introductory verse 15. Cp. Peterson, Kadambari Introduction pp. 77
ff. 135. Rājatarangini III, 97 ff. 136. In his Aucityavicāracarca he quotes Setubandha IV, 20. Kävyamālā Part I, p. 135. 137. 'ahinavarāāraddha' 1, 9. 138. Kävyamālā ed. p. 497. 139. Ibid pp. 3, 10. 140. Cp. Winternitz, Gesch. Ind. Lit. p. 63 n. 3. 141. Chapter III. 142. cf. The inscription of the Vākātāka kings. IA. XII, p. 243. 143. Introduction to Gaudavaho. pp. clxi-ccv. 144. M. A. Stein, Rajatarangini. Tr. Vol. I, pp. 66, 84ff. 145. Cp. Rāmasvāmi Shastri, 7th Oriental Conference. Baroda. 146. 1, 9. 147. Cp. Aucityavicāracarcā. p. 139 where the verse is attributed to Kālidāsa, also
Kavyamīmāṁsa, p. 60-61. 148. Cp. his concluding verses p. 497. 149. Only the third chapter of this is published by Goldschmidt in his edition.

Page #153

144
Amrita
150. But in different metres. 151. I, 1-12. 152. II, 1-36 forming one big sentence. 153. Cp. III, 6, 8, 9, 10, 12, 13, etc. 154. V, 17-31. 155. IX, 18, 43, 44, 47, 50, 82. 156. XI, 75-131. 157. III, 4. 158. III, 6. 159. III, 8. 160. IV, 23. 161. IV, 28. 162. IV, 29. 163. XI, 52. 164. V, 6. 165. VI. 9. 166. I, 38. 167. VIII, 14. 168. II, 24, 27, 31, 33, III, 45-47, VI, 48, 53 etc. 169. IX, 43, 44, 47, 50. 170. VI, 65, VII, 59, 61. 171. Cp. above not 35. 172. His preface to his edition. 173. Vararuci II. 7. 174. Gesch. Ind. Lit. p. 64. 175. Cp. guppanti I, 2; volīnam I, 3; khudio I, 4 etc. 176. Cp. II, 10; III, 5 etc. 177. Ed. by S. P. Pandit in BSS. 34. Second Ed. by N. B. Utgikar, 1927. The edition
contains the comment of Haripäla. 178. Gāthās 1-61. 179. 62-98.

Page #154

Māhārāștrī Language and Literature
145
180. 99-108.
181. 109-113.
182. 114-160.
183. 161-166. 184. 167-181. 185. 202-207. 186. 212-254. 187. 255-262, 263-269. 188. 270-279. 189. 285-338. 190. 348-354. 191. 355-382. 192. 414-417. 193. 419-421. 194. 424-430. 195. 431-439. 196. 440-459. 197. 460-465. 198. 472-484. 199. 487-497. 200. 694-796. 201. 797-800.
202. 848-856.
203. 1016-1039 204. 1185-1193. 205. 1207-1209. 206. As pointed out by Pandit almost every verse contains the fact and the fancied
idea of the poet forming the Utpreksā. 207. Cp. particularly w. 513-658. 208. Introduction pp. xl-xlviii. 209. Vienna Oriental Journal. Vol. I pp. 324-330. Vol. II, pp. 328-340.

Page #155

146
Amrita
210. Gottingische Geleherte Anzeigen, 1888 pp. 61-75. 211. Gāthā 79. 212. Dhvanyaloka p. 152. The verse should be read as follows : -
Līlādādhagguvvūdhasayalamahimandalassa via jassa
Kısa munālāharaņāi tujjha garuāi angammi| 213. Gram. Prk. Spra. p. 11. 214. Cp. Kāvyamālā ed. pp. 647, 655. 215. Prākrtasarvasva, pp. 50, 61. 216. Rājatarangini IV, 126ff. 217. The work was also called Anandakathā cf. his Devīšataka v. 104 Kavyamālā Pact
IX, p. 30. 218. His essay on Rājasekhara H. 0. S. IV. 219. Tālā jāanti guņā jālā te sahiaehi gheppanti
Raikiranānuggahiāi honti Kamalāi kamālai || Dhvanyaloka p. 62. 220. Ibid p. 152 Cp. Abhinava's comment. 221. Ibid p. 241.
222. Tam tāņa sirisahoara etc. 223. Kävyānuśāsana pp. 334-335. 224. Jesalmer Catalogue G. O. S. 225. p. 338. 226. A later work, the Kamsavadha of Rāmapāņi in 4 cantoes will be soon published
by Prof. A. N. Upadhye, M. A. 228. Kāvyālankāra IV, 11-15. 229. His Aryāsaptaśati is an imitation of Sattasai. 230. Kavyanusāsana p. 336. 231. Locana p. 223 Read the verse as follows :
Langhiagaanā phalahīlaão hontu tti vaddhaāsīe
Haliassa a āsīsam pādivesavam tu nivvavio || 232. p. 52. 233. p. 52. 234. p. 52.

Page #156

Māhārāștri Language and Literature
147
235. p. 54. 236. Dhvanyāloka p. 22. 237. Ibid. p. 52. 238. p. 52. 239. p. 97. 240. 109. 241. p. 126. 242. Sarasvatīkanthābharana p. 576. 243. p. 580. 244. p. 651 245. p. 363. 246. p. 550. 247. p. 615. 248. p. 645 249. p. 667. 250. p. 669. . 251. p. 346. 252. p. 378. 253. p. 383. 254. 471. 255. p: 520 256. Kävyānušāsana p. 80. Read
Āsāiam aņāena jettieņa tettiena bandha dhiim
Uvaramasu vasaha enhim rakkhijjai gahavaichetam || 257. p. 31. 258. p. 32. Read :
Saniam vacca kisoari paena jattena thavisu mahivatthe
Bhajjihisi vitthayacchiņi vihiņā dukkheņa ņimmaviä || 259. p. 33 Read line second as follows :
Jaha nivvanāharo sāmalanga dūmesi maha hiaam 260. p. 43, Read :

Page #157

148
Sarasam mauasahāvam vimalagunam mittasangamollasiam | Kamalam natthacchaam kunanta dosãara namo de 261. p. 107. Read:
Ahaam ujjusurūā tassa vi ummantharai pemmai | Sahiãano a niuno alāhi kim päarãeṇa ||
262. Dasarupa. p. 91 263. Ibid. p. 43.
☐☐☐
Mähäräştri Language and Literature
Amrita
JUB Vol. IV 1936

Page #158

Śaurasenī Prākrit
THE NATURE OF THE PROBLEM
The linguistic nature of the various Prākrit dialects is a point of much confusion and uncertainty. Very often we are at a loss to know whether a particular feature is a characteristic of a particular Prākrit language or not. This confusion is aggravated by the lack of unanimity on the part of the Prākrit grammarians who more often differ than agree among themselves in giving the features of the different Prakrits. Again where they are found to agree, the usage of the writers is not in agreement with them. Nor is the condition of the Mss. of the Sanskrit dramas in any way satisfactory. They present to us a bewildering multiplicity of forms from which it is difficult to reconstruct the original forms and to determine the nature of the language in which the author wrote them. This uncertainty is so great that scholars differ among themselves as regards the language of the verses found in the Sanskrit dramas. Following the authority of the rules of the Sanskrit rhetoricians which lay down that Māhārāstrī should be the language used in the songs of the ladies, some regard them as written in that Prākrit', while others rely more on the forms found in the Mss. and contend that they should be regarded as written in Saurasenī?. The forms found in the Mss. are so varied that scholars can choose forms which will suit their hypothesis with equal facility. Therefore it is necessary to view the whole problem from a historical point of view and try to decide the linguistic nature of Saurasenī with the help that can be derived both from the grammarians and the existing literature found in it.
THE GRAMMATICAL AND LITERARY SOURCES
It is usual to divide the Prākrit grammarians into two schools, the eastern one represented by Vararuci, Kramadīśvara, Rāma Tarkavāgisa and Mārkandeya, and the western one comprising the works of Hemacandra,

Page #159

150
Amrita
Trivikrama, Laksmīdhara and others. Even though the account of these schools is based on inadequate grounds, we do find a marked difference between the treatment of these two sets of writers. Among the writers of each school there is a good deal of similarity both in the manner and the material of their grammars, while they differ equally from the writers of the other school. For Saurasenī, Mārkandeya is the best representative of the eastern school. In fact he appears to be the only writer who has dealt with this language with anything like fullness. Varacuri often agrees with him, but his treatment is very meagre, and the absence of Bhāmaha's commentary on the 12th chapter makes this portion of his Prākrita-Prakāśa? obscure and its rules uncertain. Hemacandra is, of course, the best representative of the western school. All later writers draw on him and add nothing new and important to his information.
The oldest available dramas are those of the Buddhist poet Aśvaghosa. They are found in a few fragments only. They are usually assigned to the first century A.D. In these dramas there are a few passages which are written in Śaurasenī, and are found in the mouths of the Vidūsaka and the courtezan. They are of great value as giving us some insight into the Prākrits at a very early stage of their development. Next, we have the dramas of Bhāsa about whose date there is much uncertainty even after so much discussion. The archaic nature of his Prākrits' is often doubted and the peculiarities found therein are put down as due to the method of writing the Mss. of his works. In spite of all the prolonged discussion, it appears much more reasonable to assign Bhāsa to the third century and regard his Prākrits as representing a stage later than the dramas of Ašvaghosa and earlier than that' of the later dramatists. The date of Mrcchakatika10 is also uncertain, but the drama is of great value because of the great number of Prākrit passages in it and the variety found therein. So also the dramas of later writers like Kālidāsa, the Mudrārāksasa" of Viśākhadatta, the Karpūramañjarī of Rājasekhara are of some value in understanding the present language. The Prākrit verses found in the Nātyaśāstral2 of Bharata are also written in Saurasenī. Jacobi, however, prefers to call it the Pre-classical Prākrit13. We have in the Nātyaśāstra itself the statement that the verses illustrating the Dhruvas should be written in Saurasenī?4. Besides this, most of the prose passages in the dramas are in this language, both according to the theory of the science of dramaturgy and the practice of the dramatists.
There is, however, another group of Prākrit works which shows some signs of the Saurasenī language. These are the Prākrit works of the Digambara

Page #160

Śauraseni Prākrit
151
Jains. They are written mostly by the southern writers and are found from the beginning of the Christian era. Even though they cannot be said to be written in Sauraseni proper, their language shows the phonetic change of t to d to a considerable extent, and on that account is called Jain Saurasenī by Pischel and others. This extensive literature affords us some help in determining the features of this Prākrit.
MODERN VERNACULARS
The present vernaculars of India are, no doubt, the legal descendants of the middle Indian Prākrits. They are, therefore, expected to help us a little in affording some insight into the nature of their mother-languages from which they must have drawn to a considerable extent and must have preserved most of their peculiarities. But the exact nature of their relation with the various Prākrits is doubtful. Grierson has connected them, one and all, through an intermediate stage of Apabhraṁsa dialects16. Others with greater probability, want to dispense with this stage of the Apabhramśa in case of Márāthi and some other vernaculars. All the same, we have no doubt that Marāthī is immediately connected with Māhārāstrī, while Gujarātī and Western Hindī are much more akin to the Sauraseni Prākrit. As pointed out by Grierson W. Hindi shows some signs which disagree with the features of Saurasenī and this is to be explained by assuming that it is derived from various sources??. By a comparison of the various forms of common words in these languages we can know, in a few cases, the tendencies that were present in the languages which gave them birth. This will be a good help in checking the conclusions of the Prākrit grammarians. But the use of such conclusions must be made with great caution. It is just possible that a phonetic tendency that was present in the Prākrit stage might have dropped itself in later times as these languages developed, and new ones might have made themselves felt. Moreover, in later times, these vernaculars have drawn from Sanskrit a good deal of their vocabulary without subjecting it to the natural changes through which their original stock of words has passed. Further, it must be noted that such a help is mostly to be found in phonetics. In questions of morphology, we have to do without it, as most of the vernaculars have developed an analytical structure, a thing unknown to the Prākrit stage.
PHONOLOGY
Consonantal changes : In determining the consonantal changes in

Page #161

152
Amrita
Sauraseni we have first to deal with the important rule of the change of t to d, which according to the grammarians is the chief distinguishing feature of this language. (Va. 12, 3. Hem. 4, 260. Mār. 9, 20. Rt. 2, 1, 5.) This supposition of the grammarians, however, is called in question by modern scholars like Jacobi and Grierson, the former contending that this change of t to d is not a peculiarity of Sauraseni as such, but is a feature found in the Prākrit which was formerly used in place of later Māhārāstrī, and the latter thinking it not a dialectal feature but one due to the natural development in a language. It is necessary at first, to examine the grounds on which the supposition of the Pre-classical Prākrit of Jacobi is based. He appears to base his conclusion on a line of argument something like the following. He thinks that the rules of the Prākrit grammarians dealing with the Saurasenī. Prākrit are correct and therefore whatever goes against them is not Sauraseni: He, further, thinks that the rule of the Sanskrit rhetoricians that the language of the verse both in the dramas and everywhere else is always Māhārāstrī is of unimpeachable validity. Naturally, as he finds that in works like the Nātyaśāstra there are verses which show some signs of the Sauraseni language as described by the grammarians, he is compelled to disregard the suggestion of thinking them to be written in that language, because it comes into conflict with the rule of the writers on poetics. Nor can it be supposed that they are in Māhārāstrī, as they show some features which the grammarians do not allow in that language. This has forced Dr. Jacobi to consider these verses to be written in a language which he called the Pre-classical Prākrit and the main features of which he collected from the verses in the Natyaśāstra. But both these rules are not sufficiently proved. The grammarians, as Grierson says, often seem to have constructed general rules out of stray occurrences or phonetic tendencies. As regards the verses of the dramas, Hillebrandt" has conclusively proved that the verses in Mudrārāksasa are not in Māhārāstrī but in Saurasenī. The same can be said of the verses in Mrcchakatika and other later dramas. The early dramas of Ašvaghosa and Bhāsa also show no traces of Māhārāstrī in the verses found in those plays. For the first time, we meet with Māhārāstrī verses in the dramas of Kālidāsa. As we have already seen there need be no hesitation in accepting the view that the verses in the Nātyaśāstra are in Saurasenī. In the Karpūramañjarī of Rājasekhara there are many indications of Sauraseni features even in verses, a fact which is also noted by Markandeya (3. 77).
At the earliest stage of Prākrit, as seen in the inscriptions of emperor Asoka, we have no trace of this phonetic change, in any of his dialects.

Page #162

Sauraseni Prakrit
153
Therein t is preserved intact. But later on in the inscriptions of Khāravela20 we meet with the form edāni = etāni (line 13), and in the Našīka inscriptions the form sādavāhana = sātavāhana (1). Besides these solitary instances, the inscriptions show no trace of this change. Everywhere else the retention of t is the rule. The Mayidavolu grant of the Pallava king Sivaskandavarman has no form containing this change. • It is in the literary Pali that we find this change for the first time. It is found in words like udaho = utāho, sugado = sugata, niyyādeti = niryātayati, patiyadeti = pratiyātayati, pasada = prasrta, ruda = ruta, vidatthi = vitasti and others21. The contrary change of d to t is also to be met with in a few words like kusīta and others. In some places the intervocalic d is dropped. From all these instances it will be very easily seen that this change is found in Pāli to a limited extent. It affects only the body of the word and not the system of morphology. S. Levi22 has seen this change in a few technical terms of the Pāli Vinaya like ekodi (explained as eka-ūti) and sanghādisesa (explained as sangha-atiśesa). He tries to explain the presence of this change on the supposition that this feature must have been the characteristic of the original language in which the words of Buddha were composed. This is, of course, doubted by others and that also for good reasons23
Sauraseni is taken to be the normal Prākrit of the Sanskrit dramas and it is but natural that we should find this change to a greater extent in the dramatic Prākrits than in others. The fragments of the dramas of Aśvaghosa have passages which Lüders takes to be written in the older form of Sauraseni24. Here also we find nowhere that change of t to d except in the solitary case of surada = surata which may have been a mistake. In other respects the language in which the vidūsaka and the courtezan speak shows marked similarity with the Sauraseni Prākrit of the later dramatists and the grammarians. From this it appears that this change is not the characteristic of this Prākrit in its earliest stage.
In the dramas of Bhāsa we meet with this change to a greater extenta. There are forms like bhodi = bhavati, hodu = bhavatu, ganhādi = grnhāti, kädavva = kartavya, kādum = kartum, and others in which t is changed to d. We further note that this change has now begun to affect the formation of the nouns and verbs. In forms like gaanaalam = gaganatalam, gai = gati, duhiä = duhitā, bahumadaara = bahumatatara, sia = śīta, vāa = vāta26 and others t is dropped. It is very rarely retained. In other respects the language

Page #163

154
Amrita
stands midway between the dramas of Aśvaghosa and the later dramas of Kālidāsa and others, preserving some old forms and showing some new ones side by side27
.
The position of the Prākrit verses of the Natyaśāstra is the same as that of the dramas of Bhāsa in all respects. As a rule they show the change of t to d28 in forms like vādo = vāta (1), vidārida = vidārita (11), hada = hata (6), vīsadi = visati (10), vicaradi = vicarati (12), āgadam = āgatam (24), juda = yuta, jadā = yadā, kusumavadi = kusumavati (26), and a host of others. In very few cases like sahiä = sahitā t is dropped, and in some others t is retained as in bhūsitao = bhūsita, gaanatalam = gaganatalam, kusumavati, rajata. Particularly in the case of the third person singular forms of the verbst is dropped and therefore we have forms like vãi = vāti (1), bhamai = bhramati (6), pavisai = pravišati (13), etc.
In later dramas this change is consistently followed. This will indicate that by that time the rules of the Prākrit grammarians have once for all decided that this change was to be the special feature of Sauraseni Prākrit and was followed by the dramatists with great care. As seen above this is not, however, the case in the early formative period of that language. It arose after Aśvaghosa and slowly established itself in the language and at the end was taken to be the distinguishing feature of that Prākrit.
Another perplexing point about this change is that it occurs where it should not be found according to the rules of the Prākrit grammarians. In the verses of Mrcchakatika of Sūdraka which are usually supposed to be written in Māhārāstri and which are required by the rules of Sanskrit dramaturgy to be so written, we find traces of this change along with the dropping of t. So we have rakhidum = raksitum (2,3), turidam = tvaritam (2, 18), Pasadam edi (2,18), and so on. This fact may be explained on the supposition that these verses are in reality written in Sauraseni and not in Māhārāstrī as is usually supposed. Much more difficult it is to explain its presence in works like Setubandha. In that Prākrit epic of Pravarasena we find forms like dā (3, 10), dāva (3, 26), udu (3, 26) and others29. These forms are not admissible in the Māhārāstrī language. We have, on the other hand, the early testimony of Dandin to the effect that this epic was written in the Prākrit called Māhārāstrī. Dr. Jacobi has pointed out traces of this phonetic change in the Prākrit romance of Haribhadra called the Samarādityakathā30. This work is supposed to be written in the language called Jain Māhārāstrī. In another early work, the Vasudevahindī of Sanghadāsa and Dharmadāsa, we can trace

Page #164

Sauraseni Prākrit
155
this change in words like udāhu = utāho (p. 12), vipathādio = vipathātīto (p. 19), and padiņā = patinā (p. 70). In few cases t is kept, viditam (p. 6), naditam (p. 34) while in some places original d is preserved, āpadam (p. 12), bhayada (p. 2), pädava (p. 5), pamādi (p. 8) and others. These references to this change in Māhārāstrī literature show that in the early stages of that language there must have been some effect of this change of which these few traces are only remnants lagging behind. In later times the other change of dropping the intervocalic consonants completely effaced this change from Māhārāstri literature.
As a rule 32, in Ardha-Magadhi the language of the Svetāmbara canon, both t and d should be elided. So we have forms like payāvai = prajāpati, dharanīyala = dharanītala, mayana = madana, and others which follow th rule. But sometimes we find forms like udāhu (Vi33. 28), vidiya (Vi. 56), adūra (Ni34. 3; Anu 35. 56), Samudaya (Ni. 38), videha (Vi. 31; Ni. 36), and nidāna (Vi. 23) in which t is changed to d and the original d is retained. T is preserved in atihi (Ni. 93), avitaha (Ni. 16), royātanka (Ni. 119) and others. In the MSS. t is very often found used in place of the various single consonants that are dropped, and this is taken by some 36 to be the characteristic of Ardha-Māgadhi. But really speaking it is a peculiarity of Jain MSS. and its arbitrary use shows that it has nothing to do with the retention of t.
This change also plays an important part in the language of the Digambara Jains. Pischel7 called that language Jain Sauraseni on the basis of this rule being found therein to a considerable extent. Normally t is changed to d in this Prakrit. So iti becomes idi, ghāti becomes ghādi, devatā devada pati padi and so on. In words like ayisaya = atiśaya, eyam = etad, ghāi = ghāti, pariņai = pariņati, suya = śruta t is dropped. Original d is almost always retained while t is preserved nowhere. It is either changed into d or elided. This will indicate that by this time the change affecting t has worked itself out completely.
The presence of this change in other Prākrits is of no great value. It is found in Māgadhi where it is usually in correspondence with its presence in Saurasenī, if it has not undergone the further change of cerebralisation which is very common in that language. In Paiśācī and its sub-dialect CulikāPaiśācī we find the contrary change of d to t which is the result of the tendency of hardening the soft consonants found in that Prākrit. In Apabhraíía, however, the presence of this change is of some importance. The

Page #165

156
Amrita
normal Apabhramsa, in which the majority of the existing works are written, is much more akin to Māhārāstrī and shows no trace of this change. But its oldest dialect the Vrācada Apabhramśa is found to contain this change to a considerable extent. It is true, that Hemacandra nowhere makes a distinction between the various dialects of Apabhramśa, but his treatment shows unmistakable signs of the presence of this dialect. Further he does admit that there are dialects in Apabhramśa as is evident from the remarks in his Kāvyānuśāsana. Some of his Sūtras and the verses quoted herein illustrate the Vrācada dialect. Rudrata's Apabhramśa verses3 of this dialect also contain this change. From this fact it appears that this change has affected the Apabhramsa language in its early period.
From all the evidence adduced it is very difficult to say that this. change is the peculiar and distinctive feature of Sauraseni Prākrit. In fact, in the early stages of all the Prākrits we have no trace of it. It must have originated after the date of Aśvaghosa and the Prākrit inscriptions of the second century A.D. This change is a natural result of the wider change of softening the hard consonants, and it must have been working in all the Prākrits which have a tendency to soften. It is later than the origin of Sauraseni and affected it at a later stage in its development. Soon after, the development of Saurasenī was stopped and the grammarians took it to be the special feature of that language. It has also affected the Māhārāstrī Prākrit in its primitive stage, but for various reasons that Prākrit was not stereotyped until the further change of dropping the intervocalic consonants has affected it. Naturally the grammarians postulated the rule of dropping t and other stops for Māhārāstrī. The few traces that are found in that language are to be taken as the remnants of the earlier stage of that language. Its continuance in a particular locality where Saurasenī was current has further affected other languages which came in contact with it.
The Change of th to dh :- Another phonetic change in this language is that of th to dh (Va. 12, 3; Hem. 4, 267; Mār. 9, 24; Rt. 2, 1, 5). The history of this change is practically the same as that of the previous one except that in its later development th was changed to h. This change is further not found in Māhārāstrī at any stage in its development. Another question that arises in this connection is about the treatment of d and dh which are found in the original words in Sanskrit. Hemacandra, if interpreted correctly, appears not to allow them, and rules that they should be treated exactly as in Māhārāstrī. This should make us drop d and change dh to h. Most of the modern scholars have taken this implication of Hemacandra as

Page #166

Sauraseni Prăkrit
157
correct. Accordingly Konow gives in his edition of Karpūramanjari forms uara = udāra (1, 19, 2), kaali = kadali (4, 18, 3) where d is dropped, and accahama = atyadhama (1, 20, 4), ahiva = adhipa (3, 5, 2) and others where dh is changed into h. Prof. Suru gives in his Priyadarsikā forms like viviha = vividha (p. 10), mahuara = madhukara (p. 17), sāhu = sadhu (p. 23), and others where dh is reduced to h, while he appears to keep d in words like Kadali (p. 19), nivedia = nivedita (p. 25), etc. R. Schmidt3% makes it a rule that both d and dh should be so treated (5). This rigorous interpretation of Hemacandra's rule does not appear to be correct. In his Kavyānusāsana 40 he gives us a verse which can be supposed to be written both in Sanskrit and Saurasenī. There we find d kept in words dalam, mada, madirā and sāmoda, while dh is found in adhara, madhura, sādhu, and payodhara. As the verse is intended to serve both as Sanskrit and Saurasenī there cannot be any doubt of the proper preservation of the readings. This verse will lead us to conclude that Hemacandra never intended the treatment which is given to these two consonants by modern scholars. Mārkandeya lays down the rule that original d should not be dropped in Sauraseni (9, 26). Rt. further adds that original dh is not to be changed to h (2, 1, 6-7). In Aśvaghosa d is preserved in hidayena, komudagandha, ādamso, and dh in dadhi, sadhu, and others. In the verses of the Natyaśāstra d is preserved in many places like chādantā (4), rudadi (9), vidāride (11), jaladā (14), pādavam (16), vadana (22), ninada (30), pamadavana (36), kumuda (38) and so on. dh is preserved in cases like vadhu (19), madhava (24), vividha (38), and a few others. In many cases it is reduced to h; salilahara (6), jalahara (29), mahihara, etc. These cases must be explained as due to the influence of Māhārāstrī. From the point of view of phonology it appears inconsistent to drop d when t is changed to d, and to reduce dh to h when this changed into dh. In this respect I think that Mārkandeya and Rāma Tarkavāgīša are right, in spite of the fact that they come so late in time, in relation with other grammarians. Even Hemacandra appears not to have intended it. There are, however, a few exceptions to this rule. Var. gives vyāprta and Mār. adds patākā and garbhita where t is changed to d. (Va. 12, 4; Mār. 9, 20). He further adds that prathama and Prthavi do not change their th to dh (9, 24).
The Intervocalic Stops :- The question of dropping the intervocalic consonants like k, g, c, j, and others is much more clear and easy to follow in its development. In Asvaghosa we do not find this elision of these stops, and we find them preserved in rucakena, duguna, Magadhavati, etc. In Bhāsa their treatment is uncertain; k is preserved in ekka; kokanada, bahuka, and

Page #167

158
Amrita
dropped in aidi = akrti, āvantia = āvantikā, and bahua = bahuka; g is found in bhāgadeya, māgadhiā, vega, etc., but is dropped in bhāadhea, māadhiā, veena, āada; c is retained in ucida, vācaissadi, vācāda, etc., but is dropped in uida and others; i is preserved in abhijana, parittajia, etc., pis dropped in anteura, kannaūra, niuna, and also changed to v. The Natyaśāstra also shows the same condition of either preserving or dropping the stops; k is dropped in valaya, loya, ūsasiya. It is preserved in madhukarikā, sandakam; c is dropped in sahayara, viaradi, and preserved in vicaradi; g is dropped in gayana, suandha, and is found in nage; j is elided in padigaya, raani, and found preserved in manasija, teja, etc. In later times the dropping of these consonants was a rule. This was either due to the hasty generalisations on the part of the grammarians or it was the effect of the strong influence of Māhārāstrī which was affecting all the Prākrits in later times because of its - dominating position. In Māhārāstrī itself it may be the effect of its exclusive use in songs and lyrics where vowels count for all and the consonants for nothing as being of no great use to the melody.
In the language of the Digambara works the treatment of these consonants is equally uncertain. K is generally softened into g, as in adhiga = adhika, khāiga = ksāyika, gunappagāni = gunātmakāni, pattega = pratyeka, loga = loka, possibly the effect of the influence of Ardha-Māgadhi the language of the Jain canon. Sometimes k is elided, ajjhāvaya = adhyāpaka, ahiyam = adhikam, titthayara = tīrthakara, sayala = sakala, and scarcely it is retained, adhika; g is retained in āgama, bhogehim, roga, rāgo; c is also often preserved, ayadācāra = ayatācāra, ālocittā = alocya, vimocido = vimocita; it is dropped in āloyana = ālocana and pavayana = ,pravacana; i also stands on the same footing.
The Intervocalic Aspirates :- The aspirates kh, gh, ph, bh, are found to show a similar development. In Aśvaghosa we find them kept in sakhi, dullabha, bubhukkhā. In Bhāsa we find parikheda, agha, ogha, silāghania, oghattadi, megha, adhi, abhippeda, mandabhāga, and abhijana in which they are retained, while they are elided in bhaddamuha, suha, meha, lahu, ahi, sahala, and anuhūda. In the Nātyaśāstra kh is dropped in muhe, muhala, sahala, but is kept in sukha, mukhi, etc. Both megha and meha are found; bh becomes h in soha, sohida, and is retained in nabhaiala, samnibha and others. In the Pravacanasāra we find kh reduced to h in suha, suhida, and bh in lahadi, vasaho, vihavo, sahāvo, suhena; sometimes it is retained as in anubhāgo, abhibhūya, nabho, and sabhāva.

Page #168

Sauraseni Prākrit
159
Besides these rules about consonants the eastern grammarians add many rules which are mostly restrictive in that they do not allow some forms in Saurasenī which are otherwise correct in Māhārāstrī. It is very difficult to determine how far they are acceptable as there is no evidence to decide the point one way or the other. In Saurasenī śikhara does not change its kh to bh (Mār. 9, 19; Rt. 2, 1, 6); candrikā does not change its k to m (Mār. 9, 19; Rt. 2, 1, 6; cp. Mudrārāksasa, 6, 1. candimā); apūrva becomes either auvva or avarūva (Mār. 9, 23; Rt. 2, 1, 6) while according to Hem. apurava; bharata becomes bharadha (Mār. 9, 25); chāva becomes sāva and not chāva (Mār. 9, 32; Rt. 2, 1, 8); yasti does not become latthi (Mār. 9, 35; Rt. 2, 1, 8; cp. Marathi, lāthi.); kirāta does not become cilāta (Mār. 9, 36; Rt. 2, 1, 8); haridrā does not become haliddā (Mār. 9, 28; Rt. 2, 1, 7; cp. Marathi halada); in words das and others Ś is not changed to h (Mār. 9, 31; Rt. 2, 1, 7; cp. Marathi dahā, Hindi das); in words like dolā, danda, daśana, dahana, Sauraseni does not admit cerebralisation (Mār. 9, 37, Rt. 2, 1, 8); in sammarda and gardhabha there is no cerebralisation (Mār. 9, 42; Rt. 2, 1, 9; similarly Aśvaghoṣa); idānim becomes idāņi or idāņiṁ (Mār. 9, 52; Rt. 2, 1, 13) but in Asv. we find dāni and idāni and Hem. gives dāņim. For all such changes and restrictions it is not possible to determine their validity. As early as Rājasekhara many of them are violated while others are scrupulously followed by others.
Conjunct Consonants : The three conjuncts ny, ny, jñ fare differently at different hands. Hemacandra gives them the same treatment as one received in Māhārāstrī, and so they should be reduced to nn. In this he is followed by most of the modern scholars. (Schmidt. 6-7). Vararuci gives the substitute ñj optionally in words like brāhmanya, vijña, yajña, and kanyakā (12,7), but in sarvajña, and ingitajña it is nn (12, 8). Kramadīśvara gives ñj for vijña, yajña, adhijñā and pratijñā (5, 76). Pischel41 thinks that both these rules are doubtful because no other grammarian has given them and no MS. shows them. He would like to attribute it to some scribal mistake. Mārkandeya gives both jj and nn for these conjuncts (9, 46).
In Aśvaghosa we find that all these three conjuncts are reduced to one ññ, akitañña = akrtajña, añña = anya. Bhāsa shows both ññ and nn with equal frequency; jñ is reduced to ññ in akkharañño = aksarajña, atthañño = arthajña, adesakālaññadā = adeśakālajñatā, akidañña = akrtajña, bhāvañña = bhāvajña, rūvañña = rūpajña, it becomes nn in annadā = ajñatā, parinnāna = parijñāna, janna = yajña, etc; ny becomes ññ in abhimaññu = abhimanyu and nn in anna = anya, kannaa = kanyakā, ny becomes ññ in dakkhiññadā

Page #169

160
Amrita
= dāksinyatā. From this it appears that ññ is the older and nn the younger substitute42. In later dramas nn alone is found. Hemacandra43 admits ññ as the proper substitute for Māgadhi and Paiśācī. In Jain Sauraseni nn alone prevails44. In this confusion it is not possible to decide definitely the correct substitute but if we are to value the evidence of the ancient writers as of greater weight we will have to say that all the Prakrits except Māhārāstrī agreed in reducing these three conjuncts to one and the same form ññ, but later on it was changed to nn.
The conjuncts dy and ry :- The conjunct ry has received different treatment. Hemacandra says that it is substituted by either yy or ij (4, 266). In Bhāsa ry is reduced to yy in ayya = arya, kāyya = karya, payyatta = paryāpta, payyatcha = paryasta, bhayya = bhāryā, etc., and to ij in ajjua = äryakā, kajja = kārya. In later dramas jj is found with greater frequency. In Mudrārāksasa we find forms like kajja = karya (1, 16), ajja = arya (6, 1), and in Mrcchakatika ajja = arya (6, 14) in the verses. The usage of Natyaśāstra wavers between yy and ij.
Dy is reduced to yy by Aśvaghosa in jinnuyyāna = jirnodyāna. In Bhāsa dy becomes yy in uyyāna = udyāna, and ij in ajja = adya, pajjoda = pradyota. Later dy is regularly reduced to ij. Here also it appears that originally both these conjuncts were reduced to yy alone and only in later times ij took its place possibly under the influence of Māhārāstrī.
The conjunct ks :- In the treatment of the conjunct ks we have a real distinction based on dialectal variation. The eastern grammarians are perfectly right in assigning kkh in Sauraseni and cch in Māhārāstrī, a distinction unknown to the western school. Both in Aśvaghosa and Bhāsa kkh is found in most cases, akkhi = akşi, antarikkha = antariksa, kukkhi = kuksi, khana = ksana, dakkha = daksa and others. This distinction is handed down even in modern vernaculars and can be seen with clarity in the treatment which the Sanskrit words receive in Marathi on one hand and Gujarāti and Hindi on the other. So ksetra Guj. Hin. khet, Mar. set; Sk. iksu, Hin. ikh, ükh, Mar. üs; Sk. paksa, Hin. pankh, Mar. pisa; Sk. kuksi, Hin. kokh, Guj. kūkh, Mar. kus; Sk. maksikā, Hin. Guj. mākhi, Mar. māsi; and so on.
The Vowels :- Of the vowels the treatment given to r is of a very confused nature in nearly all the Prākrits. It is changed to a, į or u in a. particular set of words. U is mostly found when it is either preceded or followed by a labial sound. In case of the other two substitutes, however, there appears to be a dialectal difference underlying it. Māhārāstrī appears to

Page #170

Sauraseni Prākrit
161
favour a where Saurasenī would prefer i. This fact is noted by the eastern school of the grammarians when they put restriction on the use of a and extend the scope of i in the changes of original r. So Vararuci says that i is used in words like grdhra and others, (12, 6); Markandeya extends its scope to a considerable extent (9, 16). In Asvaghosa we find i in words like sadisa = sadrśa, hidaya = hrdaya. In Bhāsa we find forms like aidi = akrti, samiddhi = samrddhi, bhattidária = bhartsdārikā, pakida = prakrta, pavitta = pravrtta, riddhi = rddhi, jaicchā = yadrcchā, etc., where i is used. Before the suffix k, r is usually changed to u as in jāmādua = jāmātrka, bhādua = bhrātrka and others. This distinction is found in the two main dialects of the inscriptions of king Aśoka and appears again in the modern vernaculars, Sk. trna, Guj. Hin. tiņa, Mār. tana; Sk. krta, Guj. Hin. kido, kiya, Mar. kele; Sk. bhrāts, Guj. bhāi, Mar. bhāū; Sk. ghrta, Guj. ghi.
The Diphthongs :- As regards the changes of au and ai to aü and ai Mārkandeya refuses to admit them in Sauraseni (9, 13, 15) and herein agrees with the practice of Aśvaghosa where only e and o are found. In the Pallava grant only e is found in vijayavejayike. According to the eastern school of Prākrit grammarians in words like kaitava, viaiakha, svaira, vaira, etc. e alone should be used and in others like kaula, gauda, mauna, mauli and saudha o should be used. In these words Māhārāstrī, however, uses aï and aü necessarily. It is very difficult to find the trace of this distinction in the modern vernaculars, but a comparison of forms like "Guj. sohāg, Mar. saubhāgya, Guj. besavum, Mar. baisane, Guj. codā and Mar. caudā would indicate it to some extent 46.
There are a few other vowel changes operating in particular words. The grammarians give that vetasa and angāra do not change their a to i as in Māhārāstri (cp. Ma. ing!a); caturthi and badara have no o (Mar. 9, 3, Rt. 2, 1, 1; cp. Hin. ber, Ma. bor); in words like yathā, tathā, kumāra, mārjāra a is not shortened (Mar. 9, 4, Rt. 2, 1, 2; cp. Ma. kuvār); in pinda, tunda, dhammilla and others there is no o taking the place of u (Mār. 9, 5-6, Rt. 2, 1, 2; cp. Guj. pind, Ma. pend Guj. sümdh, Hin. südh, Ma. somd); words like idrśa, kīdṛśa do not change their i to e, a fact proved by the usage of the best MSS. (Mār. 9, 8, Rt. 2, 1, 2); the word puruşa does not become purisa a thing attested to by Bhāsa (Mār. 9, 9) and a few other unimportant rules.
MORPHOLOGY
Declension :- In the declension of nouns we have the debatable

Page #171

162
Amrita
question of the correct forms in the Loc. Sing. of a-stems. The eastern school (Mār. 9, 62; Rt. 2, 1, 14) explicitly states that in Sauraseni forms only in-e are admissible. The eastern school is not definite on this point but it appears that they would allow forms in-mmi as well, as equally correct..
Pischel47 has taken the view of the eastern school as the correct one. He naturally regards all forms in -mmi as incorrect in this language. Konow has also expressed the same opinion in the introduction of this edition of Karpūramañjarī where he criticises Rājasekhara for failing to distinguish between Māhārāstrī and Saurasen148 Rājasekhara uses such forms as majjhammi (Krp. 16, 8), kavvammi (Krp. 16, 8), rāmammi, sedusīmantammi in the prose of his dramas.
In the dramas of Aśvaghosa we do not get forms in Loc. Sing. Bhāsa has forms only in e as in muhe (Bc. 10, 11), hatthe (Bc. 10, 11), gehe (Bc. 13, 12), and only in Ardha-Māgadhi he uses the form majjhammi (Bc. 59, 11). The inscriptions of Asoka show forms in-si or-mhi but never in mmi.
In Mudrārāksasa there are forms like sampuunnamandalammi (1, 18), cānakkammi (1, 20), in which mmi is used, while there are others like cande (1, 18), akarune (1, 20), narāhive (1, *20), virahe (6, 2), and kūle (7, 3) where e is found. These verses are no doubt in Sauraseni and it appears that forms in mmi were used in that language. In the verses of Mrechakatika we find vajjhammi (10, 10) and kūvammi (10, 24), but in the prose passages e is uniformly used. In the Nātyaśāstra there are forms angayammi (10), sisirammi (12), jananammi (12), vanammi (12), nimmalayammi (26), kuñjammi (29), gayanammi (30), bahulammi (57), kusumammi (77), sarammi (107), where mmi is found, but the number of forms in which -e is used far exceeds these.
From the evidence of the Natyaśāstra Mr. Ghosh49 concludes that forms in ummi are admissible in Saurasenī. He argues that the authority of Mārkandeya and Rāma-Tarkavāgīša is doubtful as they come so late in the history of Prākrit grammar. Hemacandra does not impose such a restriction and being a writer of Gujarat his testimony for Sauraseni is of greater value. The testimony of Rājasekhara who had a considerable knowledge of the Prākrits as is seen from his Kävyamīmāńsā, uses them in his dramas. Lastly it is found in Sauraseni Apabhramśa.
This line of argument, however, overlooks one vital question, and that is of historical development. As we have seen, early writers like Aśvaghosa

Page #172

Sauraseni Prākrit
163
and Bhāsa do not use forms in mmi. By the time the verses of the Nātyaśāstra were composed we find that the forms in -mmi have made their way therein. In the early dramas they have taken their place in the verses only, while at the time of Rājasekhara they were used in prose as well. Their presence in Apabhraíśa needs no explanation. So it is clear that originally Saurasenī did not admit forms in -e, but slowly came to acquire them. It is due to the influence of Māhārāstrī, a fact made extremely probable by its entrance first in the verses, which, very early, came to be regarded as written in Māhārāstrī under the pressure of the theory of writers on poetics.
Ablative Singular :- According to Hemacandra the Abl. Sing. of a-stems is formed by the use of two inflections -ado and -ādu. (4, 276). Both Mār. and Rt. give the form in -ādo as correct, while they have nothing corresponding to Hemacandra's -ādu. In Bhāsa and other dramatists we find forms only in ado, but never in adu. Even in Jain Sauraseni this additional form is not to be met with except in a few cases where it is doubtful whether it is a short o or a genuine u. So it is very difficult to know whence Hemacandra took this termination. It can be tentatively suggested that it is the weakening of the normal termination due to the influence of Apabhramsa, in the Jain Saurasenī literature.
Nominative and Accusative Plural :- 'In case of neuter nouns the nom. and Acc. pl. forms are formed with the addition of -ni, according to both Vararuci and Mārkandeya. Hemacandra, however admits forms in i as well. In Bhāsa we find visaantarāņi, kesāņi, imāņi, avaidakusumāni, sephāliāgucchaāni, gunāni, paārāni, etc., where -ni is found while those in i are not to be met with. Ašvaghosa also has no form in i. Even Jain Sauraseni has ni as the usual termination. Ni is, undoubtedly the older form and i is an importation into Saurasenī from the Māhārāstrī Prākrit.
Besides these rules, both Mārkandeya and Rāma Tarkavāgīša give many rules which try to restrict the multiplicity of forms found in Māhārāstrī. I - and u-stems have -no -vo in Nom. and acc. pl., and -no-ssa in Gen. Sing. alone is to be used in Ins., Ab., Gen. and Loc. Sing, forms of feminine nouns. The guiding principle in all these rules appears to be that forms which depart too much from Sanskrit are not to be used in Saurasenī.
Pronouns :- As regards the pronominal declension Hemacandra has not a word to say. The eastern school of the Prākrit grammarians, however, gives a host of rules which do not admit many forms in Saurasenī. For the first and

Page #173

164
Amrita
second persons the admissible forms are Nom. Sing. tumam (Aśv. tuvas, Bhāsa both tuvam and tumam), aham; Plu. tumhe, vaaṁ (Asv. vayam, Bhāsa amhe and vaam); Acc. Sing. tumam (Pischel adds de), mam; Plu. tumhe amhe; Ins. Sing. tae (Bhāsa tue), mae; Plu. tumhehi, amhehi; Ab. Sing. tumādo (Pischel adds tumādu, Sākuntala has tatto), matto, mamādo; Plu. tumhehinto; Gen. Sing. te, de, tumha, tuha, tujjha, (Bhāsa and Aśv. have, tava), me, mama, maha; Plu. tumhānam, amhānań, (Bhāsa adds amhāam); Loc. tae, (Pischel adds tai, tui), mae, mai; plu. tumhesu, amhesu. From this it is apparent that the base tujha is not admitted in Saurasenī except in Ins. Plu. and Gen. Sing. forms. The forms of the personal pronouns in Hemacandra are extremely copious which is due to the fact that he appears to neglect the distinction of the various dialects.
Conjugation :- In the conjugation of verbs the eastern school prescribes the use of Parasmaipada only for this Prākrit (Var. 12, 27, Mār. 9, 97; Rt. 2, 1, 22), while Hemacandra allows both the Padas and gives -de as the termination for the Atmanepada (4, 274). Atmanepada is generally wanting in all the Prākrits. In Jain Saurasenī however, it is found to a considerable extent as is shown by forms like jayade, vattade, niyattade, vedayade parinamade, all from Samayasāra. The explanation lies in the fact that this Prākrit was greatly influenced by Sanskrit, which is further proved by the occasional forms of dual which are nowhere to be found in case of others. Hemacandra appears to draw his termination -de from this literature.
In the first person plural present -mha only is allowed by the eastern grammarians. But forms in .mo are to be found in all the dramatists from Aśvaghosa onwards, and there appears no reason to deny its validity.
Future :- The future termination is, according to Hemacandra ssi (4, 275), while both Vararuci (12, 21) and Mārkandeya (9, 104) give ssa. Aśvaghosa has gamiss-, and Bhāsa shows forms like karissadi, gamissadi, uvadhāraissadi, kahaissāmi, etc., but nowhere ssi. All the later dramatists show ssa only. Therefore, one is at a loss to know whence Hemacandra derived his termination unless it is a scribal mistake. The additional rule that for the first person forms in ssaṁ only are admissible in violated by the dramatists.
Passive :- Of all the grammarians Rāma Tarkavāgāśa alone says that the passive termination in Saurasenī should be -ia only (Rt. 2, 1, 23). Pischel takes it to be correct, while Grierson 50 is not willing to admit it, because it is only based on the authority of what Pischel takes to be the best MSS. As

Page #174

Śauraseni Prākrit
165
early as Bhāsa we find forms like uppajjai, vivajjai and it is very difficult to evaluate the authority of the MSS.
The Gerunds : For the formation of the gerunds Hemacandra gives two terminations, ia and -düna. The eastern grammarians allow only-ia. In Ašvaghosa we have the single form kariya. In Bhāsa forms in ia alone occur even in the two roots kr and gam, where according to both the schools dua is to be used. In the Nātyaśāstra, however, there are forms in üna along with those in ia. This is, no doubt, to be explained as the influence of the Māhārāstrī Prākrit. The termination dūņa of Hemacandra appears to be taken over from Jain Saurasenī. For the two roots kr- and gam- Hemacandra gives kadua and gadua, while other grammarians have kadua and gadua which is attested to by the practice of the dramatists. The rule of cerebralisation is scarcely to be found in this Prākrit and so the forms given by Hemacandra are of doubtful accuracy.
The Dhātvädeśas :- Lastly the eastern school gives a long list of verbal substitutes. According to them, bhū becomes bho, paśy = peccha, brū = vucca, kath = kadha, ghrä =, jiggha, bhā = bhāa, mrj = phumsa, ksud = khumda, ghurn = ghumma, stū = thuna, bhā = bhiya, srj = ghasa, chup = chigga, chuva, cary = cavva, grah = genha, sak = sakkun, and others. Most of them are given by Hemacandra but these rules want to restrict his long list of substitutes to this only. We have no material to decide how far these rules are true and reflect the real nature of this Prākrit. (cp. Mār. 9, 108-147; Rt. 2, 1, 25-31).
CONCLUSION.
From the evidence collected above we may deduce certain conclusions about this Prākrit. Of the features of Sauraseni a few like the change of t to d, passive in ia only, the first person plural in mha are to be rejected as incorrect and contrary to general usage, while others like the retention of original d and dh, the change of ks to kkh, ļ to i, ai. and au to e and o, Loc. Sing. in e, Ab. in ādo, Acc. Plu. in ni, future in ss, the gerund in ia and others are to be accepted as correct. We further see that the diverse treatment given to the conjuncts ny, ny, and jñ, the change of t to d, the intervocalic stops and aspirates, the introduction of the gerund in dūna and others are to be explained by the supposition of a gradual development in this language.

Page #175

166
Amrita
Another fact that is made clear is that the usual doubt expressed about the rules of the eastern grammarians is without any foundation. Many of their rules are confirmed by the usage of the old dramatists, and are correct to a great extent. The difference between the two schools is, further, to be explained by the fact that while the eastern grammarians are careful in keeping apart the differences between the various sub-dialects of a Prākrit language, the western grammarians are content in giving the broad features of the main Prākrit languages. Hemacandra appears to include the Jain Sauraseni literature while discussing Saurasenī, and this will explain to a considerable extent his peculiar rules. This is just parallel to his treatment of Māhārāstrī where he draws from Jain Māhārāstrī.
Annotations : 1. Konow has adopted this view in editing the text of Rājasekhara's Karpūramañjarī
in Harvard Oriental Series, No. 4. 2. Hillebrandt. Mudrārāksasa, p. III. 3. The various readings noted by Konow are so varied that one can easily make
out the text of the verses in Saurasenī. Cp. the Edition in Kävyamālā No. 4.: 4. Grierson. A. Mookerjee. Orientalia. Part II, p. 117ff; Vaidya, Prākrit Grammar of
Hemacandra. Preface p. 5. 5. Keith, History of Sanskrit Literature, p. 434. 6. His Prākritasarvasva is edited in Granthapradarśanī No. 3, Vizagapatam, 1927. 7. Ed. by Cowell, London, 1868. 8. Lüders. Brüchstücke Buddhistischen Dramen.
9. Printz. Bhāsa's Prākrit, Pisharoti, I. H. Q., Vol. I, 105. . 10. Ed. by Godbole, B. S. S. 1896. 11. Ed. by K. H. Dhruva, Poona, 1928. 12. Jacobi has given a grammatical analysis of these verses in his Bhavisattakaha
Introduction pp. *84-*89. They are also edited anew by Ghosh in I. H. Q., Vol.
VIII, No. 4, pp. 1-52. 13. Bhavisattakaha, Abhandlung, Exkurs uber die jüngern literarischen Prākrit
sprachen, Sect. 9, p. *86. 14. Nātyaśāstra, XXXIII, 408.
Bhāṣā tu Sürasenäyäm (?) dhruvayām samprayojayet //.

Page #176

Sauraseni Prākrit
167
15. For the language of these works see W. Denecke. Observations on the Digambara
texts in Festgabe Jacobi. Bonn, 1926; A. N. Upadhye. The Prākrit Dialect of Pravacanasāra or Jain Saurasenī. The Journal of the University of Bombay, Vol.
II, Prt. VI, May, 1934. 16. Linguistic Survey of India, Introductory. 17. Ling. Sur., Vol. VIII, p. 7. 18. Ling. Sur., Vol. VII, p. 5. 19. Ed. of Mudrārāksasa, p. 3. 20: Woolner. Introduction to Prākrit, p. 205 and the bibliography there. 21. Geiger. Pali Litteratur und Sprache, S. 38, p. 56. 22. Journal Asiatique Ser., XX, 495ff. 23. Geiger, op. cit., p.4. 24. Bru. budd. Dra., pp. 33-64. 25. Printz., pp. 3-7; Keith, Sanskrit Drama, pp. 120-122. 26. Printz., B. P., p. 15. 27. Woolner op. cit., pp. 74-75. 28. Cp. Bhavissattakaha, p. 84, n.3. 29. Cp. Goldschmidt Rāvanavaho. Introduction, p. XVII. 30. Jacobi Notes on Jain Literature. Modern Review, 1914; Introduction to his
edition in B. I., No. 169, 1926. 31. Published in Atmānanda Jain Granthamālā, No. 80. 32. Dr. P. L. Vaidya, A. Amg. Grammar, p. 19. 33. Vipākasūtra, Ed. by Vaidya, Poona, 1933. 34. Niryāvaliyao, Ed. by Vaidya, Poona, 1932. 35. Antagadadasāo and Anuttarovaväiadasão, Ed. by Vaidya, 1932. 36. Haragovinda Das Seth. Introduction to his Pāiasaddamahannavo, p. 28. 37. Grammatik der Prākritsprachen, p. 20. 38. Kävyālaňkāra V. 39. Elementarbuch der Saurasenī, Hannover, 1924, p.10. 40. Kavyamālā, No. 70. 41. Gra. der Pr., p. 191. 42. Printz. op. cit., p. 20.

Page #177

168
Amrita
43. Hem. 4, 293; 305. 44. It also shows nhu. Upadhye, op. cit., p. 162. 45. Bhandarkar. Collected Works, Vol. IV, p. 445. 46. Jahagirdar. An Introduction to the comparative Philology of Indo-Aryan
Languages, p. 113. 47. Gra. Pra. Spr., p. 253. 48. H. O. S. No. 4, p. 203. 49. I. H. Q., Vol. VIII, No. 4, pp. 9-12. 50. Ling. Sur., Vol. VII, p. 7.
ODO
Śauraseni Prākrit,
JBU. III. 6. 1935

Page #178

The Daśavaikālika-Niryukti
The Dasavaikālika-niryukti is one of the important works of the Niryukti literature, next only to the more famous Avaśyaka-niryukti. Even though the Āvaśyaka is more important from the point of view of Jain literature, theology and its ecclesiastical history, the Daśavaikālika gives us a better insight into the secular and profane subjects which cover a very wide field of interest. The Niryukti literature is certainly a peculiarity of Jain writings and has not as yet received due attention it requires to know its real value and significance. The major portion of these works is occupied with minute and detailed discussions about the various theological points of Jainism, for their chief aim is to supply and supplement these facts to the books, on which they are taken to be the comments. But besides these, we find many references to various philosophical schools, ethical doctrines, rules of worldly behaviour, logical discussions, information about arts and crafts, topics from sciences like economics and erotics, and various other subjects. All these are of immense importance, if we take into consideration the early date attributed to these books. Moreover the Niryuktis allude to a vast amount of folklore which is of some interest.
These works have some characteristics which we do not find anywhere else in Indian literature. Tradition attributes to them the nature of a commentary on the various texts of the Ardha-Māgadhi canon, and therefore each one is definitely joined with one of such books. The names of these works again are derived from such books. The Daśavaikālika Niryukti derives its name from the Daśavaikālika Sūtra, the second of the Mülasūtras of the Jain canon, and is supposed to be a commentary on it. Their nature as commentary is very doubtful and if at all these are taken as commentaries, they form a class by themselves, differing from all other commentaries in many important points. They do not explain the texts nor help us in interpreting them. The earliest available commentaries include them along

Page #179

170
Amrita
with the texts, and unlike most of the commentaries they are written in verses. They touch upon points which have no traces in the texts while they pass over much of the text, without a word to say on it. Moreover their method and procedure are altogether different and novel.
The name Nijjutti is an obscure Prākrit word. The commentators render it into Sanskrit as Niryukti and try to interpret it etymologically. Thus Malayagiri explains it to mean "the explanation of the Sūtras” which means nothing more than a commentary : sūtre prathamameva sambaddhānam satām arthānām vyākhyārūpā yuktir yojanam niryukta-yuktir iti prāpte yuktapadasya lopāt niryuktih/. In the Āvaśyaka itself we have the following line to explain the word :- "Nijjuttā je atthā jam baddhā tena hoi nijjutti” (88). This differs very little from the explanation given by Malayagiri, unless we take it to mean that various things are packed up in it and so it means Nijjutti. Besides this there is another cognate word Nijjūdha which occurs in the Daśavaikālika-n. in Nijjūdham kira sejjambhavena and Nijjūhagam vamde (12,13). Dr. Weber, being dissatisfied with these suggestions of the scholiasts, proposed to regard Nijjutti as a corruption of Nirutti, which would become in Sanskrit Niryukti, a well-known word in Indian literature meaning etymology (cp. Yāska's Nirukta). But such an emendation is clearly unacceptable. For one thing, the transition from Nijjutti to Nirutti is unaccountable in any satisfactory way, while we have evidence to show that the writers of these works kept the two words clearly apart. In Daśavaikālika-nijjutti, 10, 399 we find the word 'Nirutta' along with other Dvāras that can be applied to the word ‘Bhikkhu' to explain it. From this it is obvious that Nirutta or etymology forms only one part of the subject matter of Nijjutti. In fact, Nirutti is relegated to the background, while Niksepa is given prominence.
Another suggestion about the meaning of this term can be derived from its parallel Nijjüdha or Nijjühaga, which the commentators always interpret to mean "to take, to cull out from” usually referring to the Pūrvas as the source. Now, Dr. Charpentier has pointed out that the present Niryukti works are not the first exegetical works, but are produced from the more voluminous prose commentaries that preceded them and to which they formed as if the mnemonic verses, summarising their contents and helping the memory of the monks who expounded the texts following those big commentaries. Now, by this analogy we can say that the Nijjuttis were so called because they were culled out from the original commentaries which are now lost to us. This will explain the nature of these works to a greater extent

Page #180

The Daśavaikālika-Niryukti
171
than the previous one of regarding them as treatises on etymology.
But this explanation also is not very satisfactory. We have no means to ascertain the existence and nature of these hypothetical commentaries and how far they are epitomised in the present Niryuktis. Even if much traditional information is presupposed by these works it is doubtful to suppose that it was put down in the definite form of a commentary from which these works might be derived. Moreover much doubt exists about the meaning of both the words as explained by the commentators.
To interpret the name satisfactorily we must try to understand the nature of these works clearly. It can be very easily seen that these Niryuktis, though ill-arranged and always digressing to extraneous materials, have one thing in common, we mean, the elucidation of a few typical words of the texts. Thus, in Daśavaikālika, it can be seen that besides all other things the Niryukti explains nearly all the headings of the ten chapters of the Sūtra, and some more words from the body of the text in addition. It picks up the various words in the name of the chapter and begins to deal with each one of them in detail. Further it uses a peculiar method in explaining and interpreting them. It applies the various Anuyogadvāras or categories of interpretation, and points out the nature of these objects in the light of these view-points. Usually the four well-known Anuyogadvāras of Nāma, Sthāpanā, Dravya and Bhāva are found applied, even though others and of greater number are also laid under contribution (cp. 1; 8, 9, 34; 2; 157, 158; 3; 184 &). So the main function of the Niryuktis appears to be the interpretation of the various terms in the Sūtras by the application of this doctrine of the Anuyogadvāras. We can scarcely doubt that this doctrine is a peculiarity of Jainism, and must be of ancient standing, as it figures in the canon and the Tattvārthadhigama Sūtra (1. 5-8). In the light of this fact we can interpret the name Niryukti to mean 'the application of these Dvāras', a fact fully borne out by the nature of these works. It will also explain incidentally why such a literature is found in Jainism only.
Like all other Nijjuttis the Daśavaikālika-nijjutti is also attributed to Bhadrabāhu by the Jain tradition. This is found based upon the statement in the Āvasyaka-niryukti in which the author gives the ten Niryuktis which he wrote on the different works of the canon. But besides these ten, we have two other works having the title Niryukti, namely the Pinda-niryukti and Ogha-niryukti which are often included in the Mūla Sūtras themselves. This is due to a misunderstanding on the part of the redactors of the canon. We

Page #181

172
Amrita
have a reference in the Daśavaikālika-n. (304) in which the writer says that a more detailed statement about the rules of begging food can be found in the work called Pinda-nijjutti. The manner of the reference makes it clear that the work is an off-shoot of the Daśavaikālika-n., a work supplementing its 5th chapter. Similar is the case with the Ogha-niryukti which is an off-shoot of the Avaśyaka-niryukti and is a supplement to it. Because both these works were taken from the Mūla Sūtra group, these Niryuktis were erroneously grouped along with them, and were counted as one of them. This wil evade the possibility of counting them as more than ten.
There are also a few indications which would point out to a common authorship of these works. But whether this writer is Bhadrabāhu himself who lived in the 3rd century B.C. or somebody else is very doubtful. Dr. Charpentier has pointed out that the Uttaradhyayana-niryukti refers to Sthūlabhadra and others who, according to the Jain tradition, come after Bhadrabāhu himself, and therefore his authorship of these works is improbable. But there are some difficulties in following this line of investigation, and accepting the conclusions derived from it. It neglects the vital question about the composition of these works, which are, it can be proved with sufficient material, composite works of a very complex nature. Even tradition itself admits this fact, and the oldest commentators often try to differentiate between the original work and these later additions and elucidations. And unless we point out from which source the reference comes, its value for chronological purpose remains doubtful and vague.
Dr. Vidyābhūsana, on the other hand, has tried to make use of the tradition of the Digambaras in determining the authorship of these works. According to it there were two Bhadrabāhus, and the younger was a great writer. He would place him in the reign of Candragupta II of the Gupta dynasty, and mainly relies upon the tradition that he was a contemporary of Varāhamihira. On this view the Nijjuttis will have to be placed in the 4th century A. D., a time which would explain all references to late writers like Sthūlabhadra and others.
Even if we refuse to accept the theory of Dr. Vidyābhūsana as being founded on very insufficient grounds both in ascribing these works to the younger Bhadrabāhu against the unanimous opinion of the tradition to the contrary, and in placing him in the 4th century, we find the date ascribed to these works is nearer the truth than the early traditional date of the third century B. C. An indication to such a date is found in the Daśavaikālika

Page #182

The Daśavaikālika-Niryukti
173
niryukti (1, 81) in which the text refers to one Govinda as an illustration of a great disputant. Now from the Pattāvali at the beginning of the Nandīsūtra, we know that this Govinda was one of the pupils of Nāgārjuna and so probably lived about 350 A. D. This will also explain to a considerable extent the developed nature of the various sciences referred to in the Niryukti.
The Daśavaikālika-n in its present form consists of 447 Gäthās. It is divided into ten chapters and two Cūlikās to suit the divisions of the Dasąvaikālika Sūtra. According to the indications found in the commentary of Haribhadra, the oldest writer on the Niryukti, the present work is formed chiefly of two groups, one of which is called by him as the Niryukti Gāthās forming the real and original text of the Niryukti proper, and the second group of some 63 Gāthās which he calls as the Bhāsyvakrt-Gathās, and which are interspersed throughout the work, but mostly to be found in the Niryukti of the 4th chapter. Haribhadra further points out that the two Gāthās (I. 120, 163) are from a different hand which is further corroborated by the fact that the second verse of the two shows a repetition of the similes found in the verse preceding it and which belongs to the Gāthās of the Niryukti proper.
A perusal of the Niryukti and the Bhāsya-Gāthās will reveal the fact that the Bhāsya-Gāthâs are mainly intended as a supplement to the original Niryukti. Sometimes they point out the philosophical school to which the Niryukti makes a reference, while at other times it interprets the difficult verses in it (cp. 262 and foll.). But on the whole it supplies information not found in the Niryukti or develops the subject matter barely referred to in the original, to a greater extent with an intention of making the Niryukti fuller and easy. Tradition has not preserved us the name of the writers of these Bhāsya-Gāthās, nor is there anything in the Gāthās themselves to show the author of these verses. They are usually attributed to ancient writers like Siddhasena-ganin, Dharmadásaganin and Jinadāsamahattara. But nothing definite can be stated about it.
From the present Niryukti and particularly its introductory part we can collect a few facts about the nature of the original Sūtra as known to the writer of the Niryukti. From w. 6, 7, we know that the work was called both as Dasakāliya and Dasaveyāliya, which was understood to mean "the ten chapters preached at the time of evening” (cp. The Daśavaikālika-Sūtra : a Study, Patwardhan, 1933, p. 9). The present work was the production of Sejjambhava, who was enlightened at the sight of an image of Jina, intended for his young son Managa (14, 15). It was studied by him in six months at

Page #183

174
Amrita
the end of which he died. When the whole story was disclosed, the Samgha placed the work as authoritative and included it in the canon (444-446). The Niryukti also records a tradition which says that the two Cūlikās at the end were delivered by Sīmadhara to a lay woman for her benefit. (447). The source of this work was the Pūrvas particularly the Atmapravāda from which the Dharmaprajñapti was taken, the Karmapravāda from which the Pindesaņā was extracted, the Satyapravāda from which the Vākya-suddhi was culled out, while all others were taken from the Pratyākhyāna Pūrva. The work contained the ten chapters bearing the present names and in addition the two Cūlikās. He also gives us the purpose of each chapter which is a good indication of its contents. The first teaches the exaltation of Dharma, the second the firm belief in it, the third the rules of conduct in a shorter compass, the fourth abstension from killing living beings, the fifth the rules of begging food, the sixth the rules of conduct in detaiis, the seventh the purity of the words, the eighth carefulness, the ninth discipline, and the tenth the nature of the good monk. The first Cūlikā is devoted to the stability of a monk who is shaken in his faith, while the second deals with his secluded residence. All these facts are closely borne out by our present text. Again the very close and verbal interpretation of the first chapter in a syllogistic form goes a long way to show that the text was materially the same at the time of writing the Niryukti. In the 6th chapter however the text gives 'the title as Mahalliyāyārakahā while the Niryukti gives the title Dhammatthakahā which it follows in explaining the chapter. That name of the chapter, however, is referred to by the Niryukti elsewhere (22) as being its Adhikāra or purpose and so it must have been an alternative title to the chapter.,
To follow the contents of this work we have an indication of its scope in v. 399 of the tenth chapter. Therein it is stated that usually the topics discussed in a Niryukti are: i. Niksepa or application, 2. Nirukta or etymology, 3. Ekārtha or synonyms, 4. Linga or characteristics and 5. Pañcāvayava or logical discussion about the various objects chosen for the purpose of comment. This list can further be supplemented from w. 4, 5, which add such other topics as the author of the work, the cause of its writing, the people worthy of hearing it and the meaning of the Sūtras.
The discussion of Niksepa always results into a subtle distinction between the various meanings of the term when viewed from the stand-points of Nāma, Sthapanā, Dravya and Bhava. Usually the first two applications are only referred to while the last is given the greatest emphasis as leading to a

Page #184

The Daśavaikālika-Niryukti
175
discussion into philosophy or ethics. The word Dharma is treated in this manner in the first chapter. The Dravya Dharma is the modification of a substance and comprises on the one hand the Dharmāstikāya and the Pracāradharma or the objects of the sense on the other. But the Bhävadharma contain the Laukikadharmas or rules about the various restrictions of behaviour, the ordinary laws of the land as well as the special rules of the smaller communities, the Kuprāvacanika Dharma or the heretical doctrines and lastly the Lokuttara Dharma which again comprises the Sruta Dharma or the study of scriptures and Cāritra Dharma or the rules of conduct. The subtlety to which this method may lead us is well illustrated by the example of the word Vihamgama. The Dravya Vihamgama is an object which in its next birth will become a bird. The Bhāva Vihamgama is twofold Gunasiddha or Samjñāsiddha. The Gunasiddha is the world which remains in the sky which is called Viham. It is again of two kinds because the Gati may be due either to Bhāva or Karma. When it is due to Bhāva the Astikāyas become the Vihamgamas. The Karmagati is again of two kinds when it is found either in the sky called the Vihāyogati or anywhere else when it is called Calanagati. According to the first variety as also from the Samjñāsiddhi the birds are the Vihamgamas. According to the Karmagati both the souls and matter become the Vihamgamas (122-127). Other cases of such applications are to be found in Kāma 967-169), Pada (172 foll.), Pranidhi (359) and many others.
Th
The subject of etymology is of course of greater importance and interest. One thing that we should mark about the author's attitude towards this problem is that he is all the while trying to explain the Prākrit words without the help of Sanskrit. His etymologies are mainly based on Prākrit forms. However absurd the procedure may seem to us who are acquainted with the elements of comparative philology, the author has a natural disgust in explaining the words of his scriptures with reference to a language which he essentially abhorred as being the speech of the heterodox religions. But the inborn tendency of etymologising in him was very hard to die and therefore he states so many curious views and opinions. He explains the words. Ajjhayana as derived from Ajjhappassānayanam (29); Ajjhīna as Aksīna (31); Āya as Lābha (32); Jhavaņā as Ksapanā (33); Vihamgama as "viham gacchai" (123); Samana as Samamanai (159) or Samamano (160) or Sumano (161); Kāma from Ukkāma (170); Bhikkhu from "khuham bhindanto" (410); Jai as Jayamano (410); Bhāvanto as "bhavam khavento"; Bhikkhu as Bhikkhamāno or "anam khavej” (411), and many others of similar nature. In many cases we see that his interpretation is clearly wrong and much of it is based upon the mere similarity of sound

Page #185

176
Amrita
or the possibility of a secondary sense. But this discussion reveals one fact and that is the energy with which the Prakrit writers tried to copy the grammatical and etymological systems of the Sanskrit writers.
Another allied topic for which we find some traces in the present work is the science of lexicography. Herein we find synonyms collected for words like Adhyayana (4 in number v. 27), Druma (11, v. 35), Puspa (7, v. 36), Pravrajita (20, v. 164-165), Vakya (11. v. 336), and Bhiksu (28, 412-414). We will not be much in the wrong if we see in these verses giving the synonyms of the various words, sometimes a formidable list, the first and the earliest attempts of the activity which later resulted in the production of the Kosas found both in Sanskrit and Präkrit. The predelection of the Jains and the Buddhists for the use of synonyms is well known and it is but natural that they should take. up such attempts as the present to form bigger collections. It appears that the first impetus to the writing of the Kosas must have been from the Jain writers. There is certainly no lexicon which can be put as early as these works.
The question of Linga or characteristics always leads to a great deal of philosophical discussion. There are two very important illustrations of this the one about the Lingas of the soul and the other about the characteristics of a monk. In the first the soul is viewed from the point of view of niksepa or application, prarupaṇā or description, lakṣaṇa or the characteristics, astitva or existence, anyatva or difference from the body,, amurtatva or incorporiality, nityatva or immutability, kārakatva or agency, dehavyäpitva or co-extensiveness with the body, gunitva or qualitativeness, urdhvagatitva or having upward motion, nirmäyata or uncreatedness, säphalyata or capacity of enjoying the fruits of his actions and parināma or size (227-228). Of these again the Linga comprises ädäna, the fact that the soul is the holder of the body, paribhoga, that it enjoys the body, yoga, that it uses the senses, upayoga, that it has the quality of consciousness, kaşaya, its passions, leśyä, ānāpāna or respiration, indriya or senses, bandha or bondage, udaya or rising of the karmas, nirjarā or the shedding of the karmas, citta or sentiency, cetana or consciousness, samjñä or remembrance, vijñānam or knowledge, dhāraṇā or retentiveness, buddhi or grasping power, tha or discrimination, mati or sense perception, vitarka or probable knowledge (236-247). The characteristics of the monk are given in V. 416.
About Pañcavayava or the logical syllogism the present Niryukti has much to say on it. The whole of the first chapter is devoted to it while the last one also furnishes us with an illustration. Besides the usual five

Page #186

The Daśavaikālika-Niryukti
177
membered syllogism as found in the Nyāya-vaisesika school which our present work accepts with their terminology, we have two more illustrations of a syllogism having ten members. The first of them is to form and consists in adding the qualification visuddhi or purification which consists in or pointing out the statement as true by contrasting it with other perverted views of similar nature. This has led the writer into many digressions of no great value. But the second variety of the ten membered syllogism is much more valuable. It consists of the ten members as given in V. 143, which are Pratijñā or proposition, Pratijñā-vibhakti or limitation of the proposition, Hetu or reason, Hetu-vibhakti or the limitation of the reason, Vipaksa or counter proposition, Vipaksa-pratisedha or opposition to the counter proposition, Drstānta or example, Akanksā or questioning its validity, Akanksā-pratisedha or meeting of the question, and Nigamana or conclusion. Dr. Vidyābhūsana has expressed the opinion that this may possibly be the syllogism referred to by Vātsāyana at the beginning of the Nyāyabhasya (1-1-32). According to it that syllogism consisted of the following members Pratijñā or proposition, Hetu or reason, Udaharana or the example, Upanaya or application, Nigamana or conclusion, Jijñāsā or inquiry, Samsaya or doubt, Sakyaprāpti or the capacity of the example, Prayojana or purpose, samsayavyudāsa or dispelling the doubt. This will show that the two syllogisms are not identical and differ in many vital points. So it is doubtful whether the reference of Vātsyāyana is to the Jain form of the syllogism as illustrated by the Niryukti. One thing, however, becomes clear and it is the fact that there appears to have been a variety of syllogistic forms having ten members that preceded the present five-membered one which had a slow progress in acquiring its present position.
Another logical discussion in the Niryukti is that of Udaharana in which a very elaborate classification of the illustrations is given. The whole discussion is replete with references to stories to illustrate them. These stories are given in full only in the commentary of Haribhadra. The Udaharana is either Carita or historical or Kalpita or imaginary. It is again divided into four divisions : Aharana or an example intended in full, Aharana-deśa or an illustration intended in part, Aharana-dosa or one defective in some parts and Aharanopanyāsa or an anecdote. The first is again divided into four divisions called Apāya, Upāya, Sthāpanākarma and Pratyutpannavināśa, the two of which are fourfold when viewed from Dravya, Ksetra, Käla and Bhāva. The second is divided into Anuśāsti, Upālambha, Prochā and Niśrāvana. The third into Adharmayukta, Pratiloma, Ātmopanyasa and Durupanita. The last again is divided into Tadvastūpanyāsa, Tadanyavastupanyāsa, Pratinibha and Hetu. This

Page #187

178
Amrita
Hetu again is of four kinds, namely, Sthāpaka, Yāpaka, Vyamsaka and Lusaka. Each one of these 25 divisions is interpreted with reference to the story', its ethical application and its philosophical application, which has rendered the whole discussion very complicated and obscure. A critical study of it will, however, reveal the following facts.
1. The Udāharanas are in no way to be taken to mean the Drstāntas of the logical syllogisms. The commentator is clearly far-fetched and twisting in interpreting them as such. They only mean illustrations and parables to explain the point, a thing very common in the Ardha-Māgadhi canon.
2. Their chief application is ethical, the philosophical ones being very far-fetched and unconnected.
3. The divisions are not strictly logical but are based on the important point in the story.
4. The Hetus are certainly the logical reasons and the four-fold division appears to be an old one, but its exact meaning is far from certain.
The work refers to many philosophical schools. A reference is made to a vāuliya who is represented as advocating the view that there exists nothing in the world and therefore it follows that there is no soul. Haribhadra says that this is a reference to the school of the Nāstikas, which is not very accurate. It is a clear reference to the Nihilist school of the Buddhists (6970). A reference is also made to a school which admits a soul but refuses any activity on its part (74). This appears to be the Sāňkhya view. In w. 77-79 a clear reference to the Nāstikas is made, where the opponent admits only Pratyaksa as authoritative and on the strength of it denies the existence of the soul. The belief underlying the Vedic sacrifice that the offering of the oblation in fire produces rain is also hinted at (104). Many references to the Buddhists are also found. He refers to the worship of the Buddhas as being current (150). Haribhadra takes the word Sāmāyika in v. 256 to mean the Buddhas and this is confirmed by the quotation "Samaye ahamāsi gao” coming close by. The famous school of Ajita Kesakambalin, the Tajjīvatassariravāda, is also mentioned (264). The Buddhist doctrine of momentariness is stated in proving the permanence of the soul (270). V. 379 makes a mention of Kapila and his school who deny the soul the capacity of performing actions.
With reference to other secular sciences we have some information about the science of Poetics. The secondary use of a word is referred to while

Page #188

The Daśavaikālika-Niryukti
179
discussing the application of the word Dhamma to the Buddist and other religions (95). The author gives the well-known illustration of the word Simha and the stock phrase "Candamuhi dāriga” (100). The technical Upamā is found used in v. 131. Another important discussion in Poetics is involved in the discussion of Pada. There Gadya is defined as sweet, full of reason, connected with a foot having a fullstop and unlimited as to the end (177). A Padya is defined as of three kinds : Sama, Ardhasama and Vişama-a division wellknown to the writers on metrics (178). A Gīta is divided into five varieties when determined by Tantri, Tāla, Varna, Graha and Laya (179). A Cūrņa is explained as rich in meaning, deep on account of reasons, particles, prepositions, having many feet and being unlimited (180). An elaborate discussion of Kathā is also to be found. It is of four kinds. The Arthakathā deals with sciences, arts, acquisition of wealth and includes a discussion of Säma, Danda, Bheda and Upapradāna (195). The Kamakathā deals with all the aspects of love. The Dharmakathā is subdivided into four varieties : Aksepani, Viksepani, Samvejani and Nirvedani. The Miśrakathā of course includes all these subjects together (194-220). Again Kathā is divided into Akathā or a story preached by a lay man who is ignorant, Kathā or a story told by a careful monk and dealing with a right course of conduct, and Vikathā or a story preached by a monk who has fallen from right conduct (215-217).
We have some information about erotics as well. It is divided into Samprāpta or Sambhoga of 15 kinds and Asamprapta or Vipralambha of 10 kinds (325-328). Similarly the author divides wealth into six kinds : Dhanya or corns of 24 kinds (318-319), Ratna or precious materials of 24 kinds (320321), Sthāvara or immovables of 3 kinds (322), Dvipada or bipeds of 2 kinds, Catuspada or quadrapeds of 10 kinds and kupya or base metals of various kinds.
A few references to arts and other subjects are also to be met with (cp. 173, 433)?
Annotations : 1. H Q., DECEMBER, 1935. 2. Ind. Stud., XVII, 57, n. 2. 3. Uttaradhyayana-Sutra, Introduction p. 51. 4. Ibid., II. I foll., Weber, Ind. Stud., XVII, 57. 5. History of Indian Logic, II.

Page #189

180
Amrita
6. The stories are summarised by Leumann in the introduction of his edition. 7. A good many verses are found to be common to this work and the Digambara
work Mülācāra which, it is to be noted, requires careful consideration. The following can be given as the chief resemblances : D. 46=M : v. 219-220; D. 47=M. vi. 149; D. 48=M. v. 163; D. 158=M. X. 110; D. 188=M. vi. 4; D. 190=M. vi. 72; D. 191=M. v. 100 and others.
The Daśavaikälika-Niryukti
IHQ. XI. 4. 1935

Page #190

The Sūtrakrtānga-Niryukti
The Niryukti works do not form a distinct group by themselves and can be clearly separated from other works which are based on them or from later additions made to them. Sometimes it is very difficult to differentiate between the verses of the Niryukti proper and the verses introduced into them at a later stage and going by the name of Mula-Bhasya or simply Bhāsya. In very few cases tradition has tried to keep them apart which helps us in separating them from each other. In other cases they are mixed up in such a manner that it is now impossible for us to try to distinguish the various strata in the text of the Niryukti in its present form. What we possess in these cases are the enlarged and supplemented versions of the original Niryuktis, and they are called by the names of the Bhāsyas or Brhad-Bhāsyas, the names themselves being a sufficient indication to understand the modified nature of these works. But with the help of the few facts which tradition has handed down to us, it is usual to divide the collection of the Niryukti works into three groups, the division being based on the consideration of the purity of the text and the stages through which it may have passed before it attained its present form. There are Niryuktis the text of which is preserved to us in a form which we can reasonably take to represent their original one free from later additions. To this group belong the two Niryuktis on the first two Angas of the canon. Sīlānka, the oldest commentator on them, has given us no hint to think that there are verses in the text of those Niryuktis which are a later addition to them, as is done by Haribhadra while commenting on the Daśavaikālika-Niryukti where we find that there are many additional verses pointed out by him as belonging to the Müla-Bhāsya. The second group consists of those Niryuktis where verses of the so-called Mula-Bhāsya are added to the original Niryukti either to explain it or to supplement it. To this group belong the Niryuktis on the Āvaśyaka and Dāśavaikālika. In the third group come the Niryuktis which are now called by the names of Bhäsyas and Brhad-Bhäsyas like those on Niśītha and others where it is not now possible

Page #191

182
Amrita
to separate the original Niryukti and the later commentary on it. The importance of the Sūtrakṛtānga Niryukti, therefore, lies in its being the most pure and unmixed representative of the real nature of the Niryukti before it has suffered any changes.
The Sūtrakṛtānga Niryukti' is one of the smallest of the group and consists of 205 Gāthās. As usual the main function of the Niryukti is to explain the name of the text and the titles of the various chapters adding where necessary a summary of its contents and sometimes giving some introductory information for the understanding of the text. Compared with other Niryuktis like the Avaśyaka and the Daśavaikälika Niryuktis, we find in the present one less of the extraneous matter and of the long digressions which we meet with in those works. But the chief function of a Niryukti, the explanation of the terms in the text by the application of the various Anuyogadvāras, is to be found to its fullest extent. This is another indication to say that this explanation forms, in reality, the main function of a Niryukti, and others like the heaping up of synonyms, the etymological explanation of the terms in the text and the logical discussion and matters of similar nature are of subordinate importance if not completely foreign to it.
The present Niryukti gives us no help in determining its precise date. It refers to a vast number of philosophical school and systems which were current at the time. But no definite chronological conclusion can be based on them. Each one of these schools has a long and eventful history of its own and none of them can be dated later than the first century A. D. There are also a few persons referred to. So we have a reference to Abhaya, Śrenika, Kūlavāla (57), Jāmāli (125), Pārsva (205) and others, but none of them can be put later than the time of Mahävira himself, most of them being his contemporaries. So the reasoning which led some scholars to put down the date of some Niryuktis like that on the Uttaradhyayana-Sutra is in no way applicable to the present work, as it refers to no fact which will compel us to assign it to a later date than the one given by tradition.
More important is the question about the relation of the present Niryukti to others. In v. 99 it appears that the Niryukti refers to the Daśavaikālika Niryukti while in two other places Silänka would like us to believe that the lack of the treatment of some words by the writer of the Niryukti is due to the fact that he had already treated them in his earlier work, the Niryukti on the first Anga, the Acãra. That the present Niryukti is later than the Daśavaikālika-Niryukti is further corroborated by the fact that the information in verses 99-102 is nothing but a summary of the one

Page #192

The Sūtrakstānga-Niryukti
183
contained in verses 39-43 of the other. If all this is accepted, we will have to say that the present Niryukti was written later than these two Niryuktis.
Leumann has raised an important question about the division of the usual four Anuyogadvāras. It is usual to find them divided into the following divisions : Nāma or the naming of a thing by that particular word without seeing its propriety, Sthāpanā or a pictorial representation of it, Dravya or the material of which the thing is formed, and Bhāva or the real thing itself in a condition appropriate for its meaning. Of these Dravya is divided into Ägamatah and No-āgamatah. The second of these again falls into three divisions, Jñāpakaśarira, Bhavyaśarīra and Tadvyatirikta. The last of this division is further split up into three subdivisions called Ekabhāva, Baddhāyuska and Abhimukhanāmagotra. The last subdivision is found in two places in the Sūtrakstānga Niryukti where the author deals with the word Pundarīka (146 ff.) and Ardra (186 ff.). Now there is a tradition preserved in the Cūrnis on Āvaśyaka and Kalpa from which we come to know that Suhastin or Nāgahastin admitted the last division, Samudra admitted the middle one, and Mangu admitted the first division. If the tradition is accepted, and there is no reason to deny validity to it, and if Mangu is the first person to create the first division of Ekabhāva we are compelled to put the present Nirukti after him and so about the first century A. D. at the earliest. The same conclusion can be arrived at by taking into consideration the allusions to mythological and semi-historical stories about the patriarchs of the Jain community found in the other Nirtuktis especially the Āvaśyaka Niryukti. But the present line of argument is not conclusive. Unless and until we are able to show that Mangu was the first man to create this division, we are not justified in putting the Niryuktis later than him. What the tradition says is that while the other two writers were not willing to admit the first division Mangu did admit it. The difference of opinion, it will be seen very easily, is due to the very slight distinction that exists between these various divisions as it is based purely on the consideration of the relative time taken by the Dravya to transform itself into the Bhāva condition and not on any difference of a fundamental nature. It may just have been the case that the division was a very old one, and that these writers did not admit the three separate divisions as being without sufficient justification to regard them as distinct from each other. We have a parallel case in the fact that Siddhasena refused to admit the Naigama Naya which, we have sufficient evidence to prove, existed before him and for a long time.
The later limit for these works can be approximately settled with the

Page #193

184
Amrita
help of a few considerations. We find that the Avaśyaka Niryukti is often quoted by the canonical works like the Nandi-Sutra, the Anuyogadvāra and the Samaväyänga which attained to their present form as early as the fifth century. A. D., if not earlier. That the arrangement of the canonical works as settled in the council of Valabhai includes two Niryuktis as books in the group called the Müla-Sutras, as also the fact that the ten Niryuktis have for their basis the older arrangement of the canon into works called Angas and Angabahiras lead us to suppose that they must be considerably older than the second council. The latest reference to a Jain patriarch is to be found in the Daśavaikälika Niryukti, where it refers to Govind Vacata' who lived in the third century A. D. So we can put the collection of these Niryuktis between 300 and 500 A. D. a period which will explain all the references found in the various Niryuktis. But it is much more probable that the reference to Govinda is a later addition, in which case we can put the collection a little earlier.
The Sütrakṛtānga Niryukti, as can be seen from its rapid perusal, mainly occupies itself with the text and tries to summarise and comment on a few verses. As regards the authorship of that work we are told at the beginning that it was the Ganadharas, the direct disciples of Lord Mahävira who put it in so many words having heard the substance of it from the mouth of the Jina (18-20). The second chapter, it is further pointed out, was delivered by Rṣabha to his ninety-eight sons on the mount .Aṣṭāpada, to dissuade them from the battle with their brother Bharata (39). The praise of Mahavira is said to be a report of Sudharma to his disciple Jambu, while the conversation between Mahävira and Gotama constitutes the seventh chapter (58, 90). The last chapter of the book purports to be a discussion between Gotama and Udaga, a disciple of Pärśva (205). All this is no doubt the traditional information about the production of the work and is mainly collected from the indications found in the book itself. None of these writers can be taken to be the actual authors of the work in its present form which is certainly much more later in date.
The explanation of the name of the work is of some interest. The word Sutra means primarily a thread, and in its secondary application it means, according to our author, knowledge, as it is suggestive of the outward reality. This appears to be a confusion due to the similarity between the Präkrt forms of the two words Sutra and Sucaka which are in reality two distinct words of different meanings. The Niryukti explains the use of the word Sutta in connection with these works on the supposition that they are indicative of the real knowledge as preached by the Jina. Now such a Sutta

Page #194

The Sūtrakřtānga-Niryukti
185
is of four kinds, the Samjñā-sūtra in which the technical terms and their explanations are offered, Sangraha-sūtra which states a fundamental fact having a very wide application, Vrtta those written in metres, and Jātinibaddha which śīlānka explains to include four subdivisions of Kathaniya. a narration, Gadya or prose, Padya or metrical composition and Geya or song (3, P. 2). The other term in the name Kada leads into a much more complicated and recondite discussion. Krta necessarily implies two other terms, Karana or the means of doing a thing and Kāraka or the agent of the action who, in reality, is no other but the soul (4). The Dravya Karana is either due to Prayoga or some external impulse or Viśrasā Karana when it is due to its natural activity. The first is either Müla which includes the five kinds of bodies and the sense organs, or Uttara which is a further modification of the first like the eye, ear and others. The Viśrasā Karana is illustrated by the automatic formations like lightning, frost and others. Viewed in another light it includes such varieties as the Sanghata Karana or productions due to a collocation of the causes like cloth, Pariśāta Karana or a thing like a conch shell, or things requiring both, as a cart, and things without both, as the formation of a beam (7). The kşetra Karana means both the sky where all things happen, and also the cultivation of fields for corn. The Kāla Karana includes time in general and the ten stages a man passes through (10-12). Like the Dravya Karana, the Bhāva Karana which is also divided into many divisions in which is found the śrutajñāna Karana which is the meaning that pertains to the present context and means the production of the canonical works. If further connotes that the mind of the Ganadharas, the traditional authors of the works is free from all the karma influences particularly those that impede the scriptural knowledge (16-20).
We get a considerable amount of information about the text. The object of the first chapter is the description of both orthodox and heterodox schools and includes those who admit only the five elements, those who say that there is only one soul, those who identify soul and body, those who contend that soul is inactive, those who admit soul as the sixth element in addition to the usual five, those who say that the karma may remain without giving its fruit, the fatalist, and others of little importance (24, 30-32). Many of these schools, it will be very easily seen, were the well known sophistical schools headed by teachers like Ajita Kesakambalin. Pūrana Kassapa, Gosāla and others. Of these the materialists are refuted by the author of the Niryukti by pointing out that the quality of sentiency cannot be the outcome of the collection of the five elements which have qualities other than that sentiency.

Page #195

186
Amrita
Nor is it possible that one sense should know what the other sense has ! experienced (33). The school which holds that the soul is inactive is refuted by pointing out that if it is so, the soul cannot enjoy what he has not done nor can he suffer what he has committed; this will lead us to deny the five gatīs and facts like the memory of the past life. Neither can it be argued that the soul is inactive because it is not able to do all things, for no one can say that a tree is not a tree or a cow is not a cow only because she gives no milk or little milk (34, 35).
The second chapter is called Veyāliya, it means a particular metre in which the chapter is composed as also the fact of tearing. The Vidāraka or the agent of cutting is an axe when the thing to be cut is wood or some other material object, but it is the soul itself when the object of cutting is karma. It deals with enlightenment, the transitory nature of all worldly things, the abandonment of pride, and it is further pointed out that the monks should always try to remove the karmas already accumulated (36-41).
The object of the third chapter is to acquaint the monks with the Upasargas or the obstacles to their ascetic life. It may be due to the place or time or because of the rise of the karmas the last of which one should resist with all force. They are either painful as described in the first section or pleasant as described in the second section, while the third section deals with the purification of the mind and the effect of the heterodox system on the mind of the monk. The condition of those who succumbed to the influence of the words of the other religious teachers is described in the fourth section. The Niryukti illustrates the fact that a wrong deed once committed will bear its fruit by pointing out that if a man committing murder shows himself indifferent, a man eating poison and concealing himself and a man committing theft and then standing averse are all declared as criminals (45-53).
The fourth chapter deals with the dangers which arise out of attachment towards women. Both man and woman are viewed from various points of view and in contrast with each other. The first Uddesa points out that conversation with women leads to the breach of one's conduct, and in the second is described the pitiable condition of those who have fallen a victim to the allurement of women. Even men like Abhaya, Pajjoya, Kulavāla and others suffered and so every one must take care of women. Heroism does not consist in being powerful and daring in one's outward conduct but in showing steadfast belief in religion. It is equally true that women also should take care of men (54-61).
The whole of the fifth chapter is concerned with a detailed description of the hells where men of wicked deeds are made to suffer. In the

Page #196

The Sūtrakstānga-Niryukti
187
unsystematic description of the text the Niryukti tries to introduce system in some form by giving us the various functions of the tormentors in hell which are fifteen in number. It includes men who trouble the hellish beings, and the instruments of tormenting them (62-82).
The sixth chapter is a praise of Lord Mahāvīra in very glowing terms. The peculiar method of the Niryukti can be seen at its best in the treatment it gives to the name Mahāvīra. The word Mahat is viewed in the usual fourfold way, so also the other member Vira. Praise from the point of view of Dravya is the description of the ornaments on the body of the man to be praised, while from the point of view of Bhāva it consists in eulogizing his merits (83-85).
The seventh chapter deals with the abandonment of the state of being without conduct and having bad conduct, and the acquisition of the state of good conduct and being averse to the worldly things (86).
Sila, viewed from the Dravya point, is the use of objects like cloths, food and others, and viewed in its Bhāva state, it is either taking full or partial vows. The schools which are the advocates of the bad course of conduct are those who worship deities like Candī, those who regard water to be a purifying agency, and those who perform sacrifices (86-90).
Virya or energy is in its Dravya aspect, things like birds, animals and inanimate objects which include food accoutrements and weapons. But in reality Virya consists in the power of the soul in controlling the senses and showing qualities like activity, forbearance, fortitude, considerate nature as also performing penances. It is either that of a wise man or of a bad man or of an ordinary man. It is different with the monks and the laymen (91-97).
The ninth chapter deals with Dharma, and the writer points out that in this book it is not proper to describe the Dharma of the heretical schools (99-102). Samādhi is the subject of the tenth chapter. It is also called Ayana or thing worthy of acceptance. Samādhi means attachment to the good objects of the five senses, and in its Bhāva application it includes faith, knowledge, penance and conduct (103-106). Marga is the way. It is either auspicious or otherwise. It either leads to liberation or to hell. The bad one consists of killing living beings and showing attachment to the pleasures of this world. On the other hand, the right path leading to liberation consists in penance and self-control (107-115).
The 12th chapter gives in short the well known four schools of philosophy current at the time of the origin of Jainism. They include the

Page #197

188
Amrita
school of Kriyāvādins who hold that the soul is subject to suffering the fruits of its own acts, and has 180 divisions. The Akriyāvādins hold that there is nothing like soul, neither it is responsible for its own acts. It has 84 varieties. The Ajñānikas say that knowledge is not necessary for liberation and are divided into 67 views. The last school of the Vainayikas hold that discipline only is necessary for salvation and has 32 divisions. All these views are described here and the right view of the Kriyāvādins is upheld (116-121).
The 13th chapter deals with the proper interpretation of the texts. The story of Jāmāli, the first of the founders of the various schisms, well illustrates and danger that arises, from a wrong interpretation of the texts. A faithful preservation of the traditionally handed down scriptures is absolutely necessary (122-126).
The 14th chapter deals with the nature of both the teachers and the pupils, as also the instruction in the text, their meaning and both (127-131). The name Ādāniya is explained as to refer to the fact that the chapter is so composed as to make the first word of the line identical with the last word of the previous line. It also means the canon consisting of the 12 Angas which is worthy of being received (132-136).
The last chapter of the first book is called Gatha, and it is explained in many ways. It may mean either the use of that metre which, however, is not the case with the present chapter but which shows traces of an earlier version in metrical form, or that it gives in a short compass the substance of the chapters that went before it, or that it is sung in a charming tone. The chapter deals with the qualities of a monk (137-141).
The chapters of the second book are called the great chapters. It may mean that they are longer than the chapters of the first book or more important than them. The first explanation appears to be more plausible. The first chapter is called Lotus, and by it is meant all good things like the valuable objects, men of good conduct, gods of great power and everything that is marked by greatness. Here it typifies the monks of excellence (142-157). The object of the simile is to show that liberation is obtained by following the path laid down by the Jinas. Out of all the Gatīs man is capable of obtaining liberation, and there also the king is the guide of all the people and so his conversion is the best way to lead the people properly. To cross the difficult Samsāra all the magical and wrong ways of worship are of no use, and the only way to get over it is to follow the religion preached by the Jinas (158-164).
The second chapter deals with the various activities lead to a bad

Page #198

The Sūtrakstānga-Niryukti
189
existence and abstention from which is the best way to clear oneself of all sins (165-168).
The third chapter gives us information about the food of living beings. It is either food in general which all beings, yet undeveloped, take in, or Loma-āhāra receiving nourishment through ones hair, and lastly the usual method of taking food in lumps. At the start of his life the soul lives without food for three moments at the most and then begins to receive food (169178). The next deals with confession.
The fifth is called Ayārasuya or Anagārasuya, and says that one should follow the right conduct and abstain from the wrong one (181-183).
The sixth chapter is the result of the conversation between Ardraka and others. The Niryukti points out that though the whole of the canon is a complete work yet there are occasions which require the statement of a dogma in clearer terms. The author has given a synoptic view of the story which runs as follows:
In the country of Magadha there was a town Vasantapura and a householder Sāmāyika. He took to monkhood at the advice of his teacher Dharmaghosa along with his wife and began his wandering career. But one day he saw his own wife in the form of a nun and became attached to her. Knowing his intention she abandoned food and as a result of her religious penance she became in her next birth a god. He also, coming to know her act, died with a fast and became Ardraka in the town of Ardra as a son of one also is called Ardraka. His wife was born in the town of Vasantapura as the daughter of a merchant.
. Once, king Ardraka sent presents to king Srenika and prince Ardraka also despatched presents to prince Abhaya. Abhaya, in return, sent back to him an image of the Jina which was the cause of his enlightenment. But his father, to keep him in the worldly life, kept 500 other princes round him as guards. One day, he went out on his horse and against the advice of a deity, he took the role of a monk. In the town of Vasantapura he saw a girl, his former wife, who called him her husband. At that time the deity showers gold which the father takes up only for the girl. Many suitors come for the hand of his daughter but the girl refuses all of them on the plea that a girl can be given in marriage only once. To find Ārdraka out, the father keeps a almshouse and the girl detects him there. They are married and lead a life of pleasures. They get a son, when Ardraka shows signs of resuming his wandering career. But the boy in play encircles his father with a thread 12

Page #199

190
Amrita
times which is taken by them to be an indication that the father should live in the house for 12 years more. At the end of that period he departs.
In his way he meets the 500 princes who had become robbers and whom he converts to Jainism. On their way to see Lord Mahāvīra, Ārdraka holds a conversation with Gosāla, the Buddhists, Brahmins, Ekadandins and Hastitāpasas, which the text of the chapter embodies (184-200).
The last chapter purports to be a conversation between Gotama and Udaga, the disciple of Pārsva, about the formula of confession. It was held in the monastery of Nālandā near Rājagrha (201-205).
From this review of the text as taken by the author of the Niryukti it will be seen that it is not necessarily a summary of the text that is attempted by it. Besides that, it gives us some other information like the story of.. Ārdraka or the schools of philosophy in the first chapter which is a necessary information to understand the text. While summarising also it tries to introduce some system in the text as in chapters three and five. Of other interesting references we have some remarks on metres Vaitālika (38) and Gāthā (139), grammatical discussion of words like Mahat (83), Su and Ku (88) and Alam (202), synonyms for way (115) and logical discussion about Hetu or reason (35, 50-53). The canon is refered to in 136, 188, 189, some weapons are given in 93 and 98, the astronomical Karanas, seven of which are steady and four unsteady are referred to in vs. 11-13, and some other secular information is found incidentally.
Annotations : 1. The Niryukti along with the text and Sīlānka's commentary is published in the
Āgamodaya Samiti, Surat 1917. It is also found as an appendix to Dr. Vaidya's
edition of the text, Poona 1928. 2. Cp. Dr. Charpentier, Introduction to his edition of the Uttaradhyayanasūtra,
Upsala 1914, pp. 49-50. 3. Ubersicht uber die Āvasyaka-Literatur, Hamburg, 1914, p. 23. 4. Leumann quotes, "ãesa jahā ajja Mangū tiviham sankham icchai egabhaviyam
baddhāuyam abhimuhanāmagoyam ajja Samudda duviham baddhāuyam
abhimuhanāmagoyam ca, ajja Suhatthi egam abhimuhanāmagoyam icchai." 5. Daśavaikālika Niryukti, v. 81. 6. Cp. śīlānka's commentary, pp. 386-388.
The Sūtrakrtānga-Niryukti
IHỌ. XII. 2 1936

Page #200

Narrative Literature in Jain Māhārāștrī
The narrative Literature in Jain Māhārāstrī is wide in extent and varied in contents and form. It includes the lives of the mythological personages in Jainism, short stories and fables, legendary and semi-historical accounts of Jain monks of eminence and powerful kings converted to Jain religion, and popular. romance. But in all these species of literature Jainism cannot claim originality both in conception and execution. The epics dealing with their prophets are clearly modelled on the Brahmanic epics and Purānas; and the oldest available work of this nature, the Paumacariya of Vimalasūri, not only draws on the Rāmāyana both for its story and literary form, but expressly mentions its author with a tinge of contempt! We may find an older example of this dependence, if we suppose that the lost work of Bhadrabāhu, called Vasudevacariya, contained the same material as is found in later works bearing the same title, a fact made probable by the conservative nature of Jain traditions. As suggested by Muni Jinavijayaji?, with a fair amount of probability, the model for the Prabandhas, in which Jain writers take a special interest, is to be found in the Brahmanic works dealing with the lives of such ancient kings as Vatsarāja and Naravāhanadatta, even though all existing Brahmanic works of this nature are very late. In the case of popular romance, we have the testimony of the grammarian Patañjali”, that nany romances, presumably in Sanskrit and by Brahmanic writers, existed in the second century B. C., while the Brhatkathā of Gunādhya was probably a Brahmanic work. The case of popular stories and fables, however, admits of a different solution. The stories in Brahmanic works like the Pañcatantra, the Buddhist Jätakas and Jain stories, all presuppose a common heritage from ancient Indian folklore; while mutual borrowings, though existent, are late and easily detectable.
The question“ of the priority of Prākrit over Sanskrit in the writing of this secular literature stands on a different footing. It is seriously contended

Page #201

192
Amrita
that even the Brahmanic epics were originally written in some form of Prākrit, from which they were rendered into Sanskrit at the time of the Gupta revival. It is, further, pointed out that the Sanskrit language, conveniently called 'epic Sanskrit shows unmistakable traces of Prākrit influence, which are regarded, on this theory, as the survivals of its original language. But this theory is based upon the mistaken view of regarding Sanskrit as a purely artificial language, created by the energies of the Brahmanic priests, and restricted to sacred use only. No Prākrit epic can claim priority over the epics in their original form, and the question of epic Sanskrit admits of a better and simpler solution. The miserable effect of translating Prākrit originals into Sanskrit, as found in the Buddhist works of the beginning of the Christian era stands in glaring contrast with the language of the epics which shows clear traces of a living tongue. The priority of Prākrit in writing the popular, romance is supported by the illustration of the Brhatkathā; but we have earlier evidence of Patanjali of such romantic tales being current about his time, and as long as we are not able to decide in which language they were written, we cannot arrive at a definite conclusion.
It was but natural that, when the borrowal was from one religion to another, the religious element should occupy a very important place, by way of contrast. So, the lives of the Jain prophets and other religious persons in their mythology attracted greatest attention from the Jain writers. These religious persons are said to be 63 in number, and are classified in a definite scheme. First of all come the Tirthamkaras, the prophets of Jainism, who are 24 for this era. Their chief function in this life is to remove the darkness of ignorance and to preach the Jain religion for the benefit of people. Like the Buddhas, they are not regarded as ordinary men, but supernatural beings, whose lives are marked by definite steps in which the intervention of gods is found necessary. This conception about their nature has made their lives much like each other, and we find many writers treating them in a definite and well-establshed manner.
Besides these, there are the 12 Cakravartins who are universal kings, and most of them are patrons of Jainism; their life is marked by the great conquest of the world, at the end of which comes a course of asceticism leading to final liberation. Then come the 9 Baladevas and Vasudevas, who are born as brothers, and of whom the latter has to fight a victorious battle with his greatest enemy called Prativasudeva. Both the fighters go to hell after their worldly life, however glorious it may be, as it should be under the law of ahimsā, a breach of which is invariably punished with life in hell. The

Page #202

Narrative Literature in Jain Mähärastri
193
Baladevas, however, are silent and meek in nature, and being strong adherents of Jainism, go to heaven.
A striking characteristic of these Salākā-Purusas as they are called, is that many of them are found in the Brahmanic mythology. We, here, meet with the familiar names of Rāma and Krsna, the kings like Sagara and Bhagiratha and demons like Rāvans, Madhu and others. The stories relating to them are materially the same, and differ in a few unimportant points from those found in works like Mahābhārata and Rāmāyaṇa, or even other Purāņas. That the Jain versions are late can be seen from the changes which are made under the influence of the growing spirit of rationalism, and the purging off of the element of mystery and myth. Again, in the adoptation of these stories to a particular religion, we find that their atmosphere has changed to a considerable extent. Every person in the story is shown to have, at the end, come to realise the real nature of this world, and is made to take up asceticism under an omniscient Jain teacher.
Besides these lives of mythical persons, we have stories and lives of a semi-historical nature about the great Jain pontiffs who followed Mahāvīra, as the head of the Jain community, and great kings converted to Jainism. These works get the peculiar name of Prabandha a type of literature very much favoured by Jain scholars. Though dealing with historical theme, their value for history is small, and the spirit of religious preaching dominates them throughout.'
The romantic tale usually choose for its subject matter the lives of Vidyadharas, who are the demigods in Jain mythology. This gives it a very wide scope for the play of supernatural elements which are freely mixed up with things of this world. The love stories of their kings, who are also represented as worshipping and honouring Jain Tirthamkaras, occupy nearly all the Jain romances.
This narrative literature must have been wider than what we now possess. There are indications that many early works are lost to us, and we have only their names preserved in later works. The oldest of such works appears to be the Vasudevacariya of the venerable sage Bhadrabāhu, who was, in all probability, a contemporary of Candragupta Maurya. This work is referred to by Devacandra, the teacher of the famous scholar Hemacandra, in the introductory portion of his śāntināthacarita. In the Niśīthacūrni we find the mention of Naravāhanadatta-Kathā under the category of Laukika, and Magadhasenā and Tarangavati as illustrations of Lokottara Kathās. Another

Page #203

194
Amrita
story called Sulocană is mentioned by Udyotana in his Kuvalayamālā; while Siddhasena-ganin, in his commentary on the Tattvārtha-Sūtras mentions an Ākhyāyikā called Bandhumati. All these ancient works are now lost, and we cannot form any opinion about him.
The oldest available Jain epic of the nature of a Purāna is undoubtedly the Paümacariya of Vimalasūri. Very little is known of this voluminous writer except that he was the pupil of Vijaya, himself a pupil of Rāhū. Neither of these names are to be found in old Pattāvalis. The author also mentions that he belonged to the Nāilavamsa, which was also the vassa of his teacher. Now this vamśa is said to be no other than the Nāgendrakula founded by one of the four pupils of Vajrasena who lived in the first century A. D.
The date of Vimala is a disputed question. At the end of the work, itself we find the year 530 A. V. mentioned as the time of writing the work. This means that the writer wrote his work in A. D. 4, or 64 A. D. if we put the nirvana of Mahāvīra 60 years later following Jacobi. The same scholar has pointed out that the work cannot be so early. He has detected in it traces of the influence of Greek astronomy, which according to him, was introduced into India in the 4th c. A. D. The language of the work also shows some Apabhramśa features like the gerund in evi', and the use of the particle ‘kavana'. These facts lead him to assign the work to Circa 300 A. D. K. H. Dhruva has also pointed out that the use of such modern metres as gāhini and saraha, points to the same conclusion.
The relation of Vimala to Ravisena, whose Padmapurana agrees closely with our text is sufficient to prove that he is earlier than him and therefore cannot be later than 7th c. A. D. as some have thought with the wrong supposition that he copies Ravişeņa. That he belonged to the first century is not probable in view of the fact that the Nāgendravamsa was founded at the end of the century, and possibly a few generations might have elapsed between the founder of the vamsa and our author. The Greek influence also, can on no account be carried back earlier than the second century. Vimala, therefore, lived in the second or the third century A. D.
His Paümacariya is a vast epic divided into 118 cantos covering more than 9000 verses. The subject matter of the work is nothing but the Jain version of the epic story of Rāma, who is called in Jain mythology by the names of Padma. Herein, we meet all the characters of the Rāmāyana of Vālmīki, and the author refers to it at the beginning. In the main, the story is the same, with changes like the killing of Rāvana by Laksmana who plays
:

Page #204

Narrative Literature in Jain Māhārāștri
195
the part of the hero; the nature of the Rāksasas and Vānaras, who are represented as various tribes of the Vidyādharas, deriving their totemic names from the pictures upon their banners, the dropping of the golden-deer incident and many minor changes. The work also treats of the later life of Rāma, the birth of his children Lava and Kusa, and the final liberation of Rāma. The whole story is set in the back-ground of the life of Munisūvrata the 19th Tirthamkara, and this necessitates the writer to introduce so many religious sermons in the book. The work also contains a good many episodes of romantic and legendary character.
The style of the work is marked by fluency and force. The chief interest of the work being narrative, the author cares more for the telling of the story than mere poetic elaborations and discriptions of towns and other things. But in some cases where such descriptions occur, we get a glimpse of the writer's ability of composing elegant verses of real poetic flash. His vocabulary is rich and many Deśi words are to be met with. The introduction of these words served the purpose of making the epic easy to understand for the common people, which was necessary for a work of religious and propagandistic nature. This popular nature of the work is also the cause of much grammatical anomaly and the influence of the Apabhraíśa language.
Another old writer is Padalipta, whose Tarangavai is now lost to us. That the writer is old is shown by the fact that his work is mentioned in the Anuyogadvāra-Sūtra of the Ardha-Māgadhi canon, which cannot be later than the 5th c. A. D. but considerably earlier. In the prabhāvaka-carita, we find one whole Prabandha' devoted to him. From it we gather that he was born in the town of Kosala to a merchant Phulla and his wife Pratimā. He was named Nāgendra because he was born to his mother as the result of drinking the water touched by the feet of Nāgahastin. The child was given to him and was initiated to Jainism very early. He was renamed as Pādalipta when he became a monk. He is represented as visiting towns like Mathurā, Pātaliputra, countries like Lāta, Surāstra and Satrumjaya, all of them in northern India. The fact that he came to Mānyakheta when Krsna of the Rāstrakūta family was ruling there, is unhistorical. That he was a contemporary of King Sātavāhana of Pratisthāna, and composed his romance for him, as also the fact that the town Pälithānā derived its name from him, is equally untenable. Another tradition makes him a contemporary of King Vikrama of Ujjayini.
Even though this fine romance is now lost, fortunately for us, we have a short epitome of it, called Tarangalolā, made by Nemicandra, a pupil of

Page #205

196
Amrita
Virabhadra, of unknown date. From the introductory remarks of this writer, we come to understand that the original work was a fine piece of poetic composition, and equally difficult on account of the vast amount of Deśī words used by the writer. And we may readily believe that the loss of the work is to be attributed to the same cause.
Early in the 6th c. must have been written the great prose work Vasudevahindi&. It is a composite product of two authors, Samghadāsa and Dharmasenaganin. That the work is old is proved by the fact that so ancient an author as Jinabhadrakṣamāśramaņa refers to it in his Visesaņavati, and that later writers like Haribhadra and Malayagiri refer to it and use it in writing their commentaries on the Āvaśyaka-Niryukti. According to Pattāvalis, Jinabhadra is assigned to the 7th c. A. D. and so our author must be a little earlier.
...
The work is called a Hindi, a peculiar form of composition meaning the transmigratory 'wanderings of the hero. The present work is divided into 100 Lambakas, and gives a good deal of general information about the origin of the story and other extraneous matter. The chief aim of the work is to give the life of Vasudeva and Krsna, who are made contemporary of Aristanemi. But the first part of the work contains another practically independent work the Dhammillahindi which is said to represent the good effects of meritorious deeds in this very life, and thus stands in contrast with the other in which they take effect in the next life. As usual, this story is also an adaptation of the Brahmanic epic dealing with the fortunes of the Kuru race; and it also appears to make use of Harivamśa. The striking feature of this remoulding is the insignificant place to which the Pāndavas and Kauravas are relegated, and the prominence in which the Harivamsa emerges, a feature found in its formation in the Brahmanic epic itself.
How far our writers have made use of the older work of Bhadrabāhu, we cannot know. But the model of our work appears to be the famous Brhatkathā of Gunādhya, which is now found in late Sanskrit renderings of dubious accuracy. This is indicated by the use of Lambakas, to denote the divisions and the use of such terms as Mukha, Pratimukha, śarīra etc. to designate the various parts of the story, which terms later vanish from Indian literature. The style of the work is simple, but much uneven; in that sometimes the author tries to show his poetic skill in describing towns and countries, with dubious success. Whenever he comes to the narration of the story he becomes fairly simple and effective, and his narrative moves with

Page #206

Narrative Literature in Jain Māhārāștri
197
rapid strides.
We now come to the great romance of the famous scholar Haribhadra called Samarāditya-Kathö'. Our scholar, we know, was a very famous writer in the middle ages, and a vast amount of legendary information is current about him. The most famous of these stories is the one describing his meeting with the Jain nun Yākini, who puzzled this proud scholar with a verse giving the scheme of the Salākā-purusas. This caused Haribhadra to become her pupil and take up Jainism. This, it is said, is the significance of the phrase 'Yakini-Mahattarāsūnu,' which Haribhadra uses at the end of his works. Another story relates how two of his cousins and pupils Hamsa and Paramahamsa were killed by the Buddhists, and led him to take the resolution of killing in turn 14000 Buddhists. But at the teacher's remonstration, he repented of his bad intention, and as an expiatory right for the same determined to compose so many works. This, again, it is pointed out, is the significance of his virahānka which he consistently uses at the end of his works. But such stories cannot be taken as sober history. They appear to be later inventions to explain the use of these curious phrases by the writer.
Haribhadra's' date can be fixed with considerable certainty. The fact that Siddharsi the author of the allegorical story Upamitibhava-prapañcakathā, written in a A. D. 906, refers to him as his guru, is not to be explained as meaning his actual teacher, but only showing that Siddharși was influenced by the great scholar to a very great extent and so was regarded by him as his guru. This is certain, because we have an earlier writer Udyotana who wrote in A. D. 779, referring to him and his work in an unmistakable manner. Haribhadra himself refers to the Buddhist logician Dharmakīrti and so must be later than A. D. 650. We may conclude, therefore, that Haribhadra lived in the 8th century A. D.
His Samarādityakatha is a big romance in mixed prose and verse, divided into 9 Bhāvas to which it attached a bhūmikā at the beginning, and gives the lives of two persons inimical to each other, in their 9 births. In each of these series of lives there results a quarrel between the two in which the one with the bad disposition of mind, kills the other. The cause of this enmity is traced to a really tragic situation in which one is insulted and maltreated by the other unknowingly. Three times Gunasena invites Agniśarman for meals, and is not able to take notice of him on account of some worldly hurry and anxiety. This leads the insulted ascetic to make up his mind to take revenge of the king, which act is technically called a nidāna. And the chief

Page #207

198
Amrita
purpose of the story is to illustrate the bad results of such an impious wish, which one should always try to avoid. The present work is called by the writer as a Dharmakathā, and teaching of religious principles as its avowed theme.
Haribhadra is no doubt, a great scholar and well versed in Sanskrit literature. This has made him to write his work in an elaborate and ornate style, just on the model of Sanskrit Mahākāvyas. We find him using a rich and Sanskritic vocabulary, and he follows the conventions of the writers on poetics to a considerable extent in giving descriptions of towns, lakes and gardens. But he can also become simple and effective, as when in the second chapter he narrates the story of the Tāpasa and his three visits to the King. His use of poetic figures is happy and free from the defect of too much artificiality. But his prose is marred by the use of long compounds and series of puns, rendered more recondite on account of the greater flexibility of the Prākrit language. As pointed out by Jacobi, his language shows peculiarities of Śauraseni, particularly in prose.
Close to Haribhadra, we must place Uddyotanasūri, the author of the great romance Kuvalayamālākathā, of considerable extent and written in A. D. 779. We are fortunate in having a long prasasti of 29 verses, at the end of the work, in which the author gives much useful information. In the northern part of India, there was a town Pavvaiyā near the river Candrabhāgā, and was the capital of the Yavana King Toramāna. The teacher of this King was one Harigupta of the Gupta family, and lived there. One of his pupils was Devagupta, a royal descendant of the Gupta dynasty, who, in turn, had a pupil called Sivacandra, bearing the title mahattara. In his wanderings, this Sivacandra made his residence at Bhillamāla, otherwise known as Śrīmāla, the ancient capital of Gujarat. His pupil was Yakşadatta of great repute, and had a band of pupils who are represented as making the whole of Gujarat a convert to Jainism. One of these pupils was Vateśvara who built a temple the Jina in the town of Ākāśavapra. He had a pupil Tattvācārya who was the teacher of our author. Uddyotana received the knowledge of the scriptures from Virabhadra, while he learnt logic and other sciences from the famous scholar Haribhadra.
We also know something about the secular parentage of Uddyotana. His father was Samprati or Vadesara, and the name of his grand-father was Uddyotana, who lived in the town of Mahadvāra. Our author wrote the work in the town of Jabālipura, while living in the temple of the Jina, built by his

Page #208

Narrative Literature in Jain Mähäräştri
199
teacher Virabhadra. The ruling King at the time was Vatsarāja of the family of Pratihāras.
This Prasasti gives rise to many problems of great historical importance, the solutions of which are of equal interest both to Jainism and Brahmanism. Toramāna is no doubt the well-known Hun conqueror, whose capital is here located in Punjab. But the case of the two Guptas remains doubtful. It is of great importance to know whether they belonged to the royal Gupta dynasty, which appears to be the meaning of the phrases used by our author, but without sufficient evidence to accept it as settled.
The Kuvalayamālā is also important on account of its use of so many Prakrits. It makes use of Paiśāci and Apabhramsa, besides the normal Jain Māhārāstrī, and gives much interesting information about earlier authors. It makes mention of Pādalipta, Sātavāhana, Satparņaka, Guņādhya, Vyāsa, Vālmīki, Bāna, Vimala, Ravişena, Jadila, Devagupta, Prabhañjana, and Haribhadra with their various works.
A'work dealing comprehensively with all the mythological persons and possibly forming the model for the still greater work of Hemacandra in Sanskrit, is the Mahāpurusacarita of Silācārya. The identity of this writer with śīlāňka, the commentator of the first two Angas of the Jain canon, is doubtful, even though their dates fall in the same period. Our author was a pupil of Mänadeva of the Nirvrtikulīnagaccha, and wrote the present work in A. D. 868. The work deals with 54 Salākā-purusas, the 9 Prativāsudevas being included in the lives of their opponents, and not counted separately. In A. D. 917 was written a romance called the Bhuvanasundari-Kathā of more than 8000 ślokas, by Vijayasimhasūri, a pupil of Samudrasuri of the Nāilavamsa which is identical with the Nāgendrakula.
In the 10th century or about that time is to be placed the small narrative in prose called the Kālakācāryakathānaka 10, of an unknown writer. It gives the story of the overthrow of Gardabhilla the King of Ujjayinī, who is said to have carried away with force into his harem Sarasvatī, the younger sister of the sage Kālaka. To take revenge on him, Kālaka goes to the Saka Satraps called the Sāhis, and with their help conquers the King of Avantī. The story may have got some historical truth in it. It shows that as early as the first century A. D. Jainism had its stronghold in the country of Mālavā, and appears to be on its course to Gujarat. The style of the work is elaborate and the writer makes fullest use of the few chances he gets of describing seasons and towns, in which he tries to show his poetic skill.

Page #209

200
Amrita
In the beginning of the 11th c. A. D. we have a beautiful romance in Surasundaricariya" of Dhaneśvaramuni. At the end of the work, he informs us that he was a pupil of the famous writers Jineśvara and Buddhisāgara, who belonged to the Candrakulīnagaccha. The work was written in A. D. 1038 in the town of Candrāvati.
This romance is divided into 16 Paricchedas and contains 4000 Gāthās It gives in the usual method of a romantic author, the love story of a Vidyādhara chief and his marriage with his beloved, after a long period of separation, marked by the alternate use of hope and despair. The device of a story being set into another is, here, used with considerable skill and success; while the whole structure of so many story-currents running parallel, comes to a close with appreciable effect. The writer says at the beginning that the story was written at the request of his elder sister Kalyāna-Mati, and was written in an easy style devoid of all poetic embellishments to make it comprehensible to ordinary people, a fact borne out fully by his production.
Earlier than A. D. 1052 is the Pañcamīkahā of Maheśvarasūri a pupil of Sajjanopādhyāya. The work is written to illustrate by means of stories, the importance of the vow of Srutapañcami. Two of these stories, those of Jayasena and Bhavisyadatta are given at greater length, than the other eight stories. The work has 2000 Gāthās on the whole.
· A little later lived the writer Candraprabhamahattara, whose small poem about the life of Vijayacandrakevalin in 1063 Gāthās, contains stories to illustrate the good effect of the worship of the Jinas. They are eight in number and illustrate the eight different ways of performing this worship. The writer belonged to the Vijayagaccha, and was a pupil of Abhayadevasūri. The work was written in A. D. 1070, at the request of his pupil Viradeva in the town of Devāvada, while the author was living in the temple of Rşabha. The style of the work is easy and without pretensions to high poetic flights
A big work, dealing with the last prophet of Jainism, is the Mahāvīracarita of Gunacandra who belonged to the Candrakulinagaccha. The author was a pupil of Sumativācaka, who, in turn, was a pupil of Prasannacandra at whose request the present work was composed. The date of the work is given as A. D. 1082.
The work is divided into eight Prastāvas and uses prose and verse indiscriminately. The author deals mainly with the legendary life of Mahāvīra, and has little historical sense. Half the work is devoted to the former lives of

Page #210

Narrative Literature in Jain Māhārästri
201
Mahāvīra, which begin with that of Nayasāra, a village officer. Then follows a long series of lives in which the most important are those of Marici, Triprsta, and Priyamitra. At the end comes the life of Mahāvīra himself. The work ends with the attainment of omniscience by the lord and his sermons to his first disciples, who become his 11 Ganadharas.
The language of the present work is much more grammatical and chaste than that of older works like Paümacariya and others. There is no use of words without inflections nor of cases wrongly used. This is mainly due to the influence of classical Sanskrit works, that always exercised their influence over Prākrit throughout its career, and later became more and more dominant so as to render it servile and imitative. The style of Gunacandra is obviously influenced by writers like Kālidāsa and Bāņa, from whose classics he derives many of his ideas and expressions. The result of his Sanskrit studies is the use of long and intricate compounds, figures of words, only based upon paranomasia, and rare and poetic words only to be met with in Sanskrit. It also led to the decrease of purely Deśī words which were substituted by tatsama and tadbhava words as being more scholarly.
But along with this change in language the poetic value of the work has also changed. The monotonous flow of the ealier writers now gives place to a varied and careful use of different metres, even though Gāthā continues to predominate. Beautiful descriptions of kings and towns are to be met with on every page of the work, and the whole work stands out as a charming Kävya standing comparison with other works in Sanskrit of its nature. The modern date of the work has made it possible for the writer to introduce a few Apabhramśa strophes to exhibit his knowledge of that language. In still later works they grow more and more extensive. Mahāvīracarita has some 50 Apabhramśa verses, which appear to show the Gurjara dialect of the Nāgara Apabhramsa as pointed out by the Gen. sing. in -ha.
The famous commentator Abhayadeva had a pupil Vardhamāna who wrote two Prākrit works. One of it is the romance Manoramācarita having 15000 Gathās, written in A. D. 1083; and another is his Ādināthacarita in 11000 verses, divided into 5 Prastāvas, giving the life of Rşabha. It was written at Stambhatīrtha in A. D. 1103. It has many Apabhramśa verses scattered throughout the work. In the same year was written the Santināthacarita of Devacandra, the teacher of the famous writer Hemacandra. He belonged to the Purņatallagaccha, and was a pupil of Gunasenasūri. His work covers 12000 ślokas, and at the beginning of it, he refers to earlier

Page #211

202
Amrita
authors like Indrabhuti, Bhadrabāhu and his Vasudevacarita, Haribhadra's Samarādityakathā, Dākṣiṇyacinhasūri's Kuvalayamālā, and Siddharşi's Upamitibhavaprapañcakathā. It has many verses in Apabhraíśa dialect.
A year later, in A. D. 1104, śāntisūri a pupil of Nemicandra belonging to the Candrakula and the Brhat gaccha composed the Prthvicandracarita, at the request of his pupil Municandra. It has 7500 verses. In A. D. 1108 was written the Pārsvanāthacarita of Devabhadra a pupil of Prasannacandra, himself a student of the commentator Abhayadevasūri. The work was composed at Bhrgukaccha, and deals with the lives of the 23rd Tirthařkara.
In the 12th century, we find Hemacandra the great scholar of Gujarat writing his Kumārapālacarita??. He is one of the many-sided and fruitful writers, and he is as great a scholar as a good poet. To him, more than to anybody else Gujarat owes its greatness in literature. He was not only a writer of Jain works, but has also written text books on grammar, lexicography, poetics, metrics, and was well-known by his title of Kalikālasarvajña, which he eminently deserved. He was born in the year 1089 A. D. at Dhundukā a small town in Gujarat, as the son of a merchant. His parents were pious Jain lay-followers, and in his early childhood he was converted to that faith. As a Jain monk he spent the greater part of his life in the capital of Gujarat. His patron was first the Cālukya King Jayasimha Siddharāja ( A. D. 1094-A. D. 1143 ). His successor King Kumārapāla was converted to Jainism by him. After this Gujarat became the centre of Jain religion, and many of the acts of this King were directed by his teacher.
Considering the close relationship between King Kumārapāla and his famous teacher Hemacandra, we do not find much historical information in his work, as we should expect. In fact the motive that guided Hemacandra in writing his work was neither historical nor poetic, but purely grammatical. To illustrate his grammar called Siddha-Hemacandra after the names of both the writer and his patron, he composed a Dvyāśrayakāvya of which the Kumārapālacarita forms a part, and it is intended to illustrating his Prākrit grammar. As such the work shows no merit except that of supplying illustrations for the various rules in his grammar. The last two cantos are composed in various Prākrit dialects like Saurasenī, Māgadhi, Paisāci and Apabhramsa.
An elder contemporary of this writer was another Hemacandra belonging to the Maladhārigaccha, and a pupil of Abhayadevasūri. He is wellknown for his great work in the form of a commentary on

Page #212

Narrative Literature in Jain Māhārāṣṭrī
203
Viseṣāvasyakabhäşya of Jinabhadra. He wrote the life of Neminatha the 23rd Tirthamkara in the Präkrit epic called Neminathacarita in 5000 Gāthās.
One of the pupils of this Hemacandra was Śrīcandra. In his life as a householder, he appears to be a state officer in the reign of King Siddharāja. His big epic, the Munisuvratasvāmicarita, which includes the story of Rāma also, was written in A. D. 1135, at the request of one Dhavala of the Poravāda family, in the town of Asavallipura. At the end, we have his Prasasti giving much useful information about his contemporaries.
A co-student of this Śrīcandra was Lakṣmaṇagani who is the author of the big work Supäsanähacariya, of some 8000 Gāthās. From the end of the work, we know that, he wrote it in A. D. 1142. It describes the life of the 7th Tirthankara, giving his earlier lives, wherein he acquires the good merit to be born as a Jina. A major portion of the work is devoted to the preachings of Jina, after he has obtained omniscience; but even this part is made interesting by the introduction of so many stories to illustrate the various vows of Jainism, a fact showing clearly the religious spirit dominating all the writings of Jain. monks. The style of the work is flowing and possesses a charm of its own when he is describing some supernatural objects. The writer's skill of adapting sound to the sense can be seen in his description of the miserly merchant; though he is unable to come out of the dry narratives of a host of bare names, he exhibits much poetic skill when he comes to describe the Abhiseka of the prophet on mount Meru.
About this time lived another Śrīcandra who belonged to the Candragaccha, and was a pupil of Devendrasūri. He composed his Sanatkumāracarita of 8127 Ślokas in A. D. 1157. In the beginning he refers to Haribhadra, Siddhamahākavi, Abhayadeva, Dhanapāla, Devacandra, Säntisūri, Devabhadra, and Hemacandra of Maladharigaccha. The story of Sanatkumāra, the sovereign King, is full of romantic and supernatural elements. He was carried away by a horse and deserted into a dense forest, where he met with a number of difficulties from which he comes out successful, and obtains in the manner of a true romance, so many daughters of Vidyadharas as his wives.
Another contemporary writer is Haribhadra belonging to the Vatagaccha, and a pupil of Śrīcandra. His patron was minister Pṛthvipāla, of the two Solanki Kings Siddharāja and Kumārapāla. He is well-known for his Ap. work Neminähacariu written in A. D. 1159, in the same year when Kumarapala was converted to Jainism by Hemacandra. He also wrote a big

Page #213

204
Amrita
Prākrit epic called Mallināthacarita in 3 Prastāvas, giving the life of the 19th Jina who is regarded by the “vetāmbaras as a woman, which was due to his deceitful behaviour in his former life. Another work of this author is the Candraprabhacarita in 8000 verses.
At the end of this century, just after the death of King Kumārapāla and his teacher Hemacandra, we find Somaprabha writing two of his big Prākrit works about A. D. 1195. One of it is the Kumārapālapratibodha"3, originally called by the writer Jinadhammappadivoha; a big work giving so many stories as illustrations of the various principles of Jainism. It is divided into 5 Prastāvas. It begins with the description of Gujarat and its capital Anahillapātana, and enumerates in quick succession the various Kings of the Cālukya family founded by Mūlarāja. We then come to Kumārapāla who shows some inclinations towards the problems of religion and its different solutions. His minister informs him of Hemacandra and gives a short account of his birth and monkhood. Then follows the meeting between the King and this illustrious scholar, and the whole work is represented as forming a conversation between them. At the end of each story, the King is shown to accept the particular vow of Jainism, and gives order to be observed throughout his Kingdom.
As Jinavijayaji, the learned editor of this work, has observed, in view of the fact that the present work is written only 11 years after the death of King Kumārapāla, we find little historical information about the times in which the author lived, and the then conditions of Gujarat. The aim of the author is only to give stories and old religious sermons to impress the principles of Jainism on the mind of his readers. The two great personages only form a convenient and attractive background to this aim. The style of the work is fairly simple and smooth. It is written in mixed prose and verse, and many Apabhramśa and Sanskrit stories are found, particularly in the 5th Prastāva. Another Prākrit work of this writer is the Sumatināthacarita, giving the life of the 5th Tirthamkara. It occupies more than 9000 verses, and like his other work contains stories and legends to preach Jain principles.
Interesting is the romance, called Malayasundarikathā, in which an unknown Jain writer has worked many popular and fairy tales. All the usual motives of a fairy tale are to be found in this romance. It describes the love story of prince Mahābala and Malayasundarī, their separation and reunion. It is earlier than the 14th c. A. D. when a Sanskrit epitome of the work was made by Dharmacandra.

Page #214

Narrative Literature in Jain Māhārāștri
205
Still later, we have another work dealing with the life of the Jain Tirthařkara śāntinātha by one Munibhadra, written in A. D. 1353. A few years later, in A. D. 1371, is written a romantic story of a King Śrīpāla, by one Ratnasekhara, to illustrate the effect of worshipping the Navapadas.
In the 15th c. we have the Rayaṇaseharakahā of Jinahassa, a pupil of Jayacandra. It was written in prose and verse, and covers 8000 ślokas. Still later we have the short story of Kummāputta in 180 Gāthās, written by Jinamāņikya or his pupil. It gives his life as an illustration of the wonderful effects of bhāvasuddhi, which helps even a householder to obtain omniscience.
From this brief review of these works, we come to know that they are useful to us in many ways. Besides the help they give us to understand the Prākrit languages and the Deśī vocabulary, their value as representing various types of poetic composition is not inconsiderable. Even though they are not of much historical value, the scraps of information they supply about the middle and dark ages of Indian history, are not only welcome but necessary to give us a connected picture of the many-sided activities of the same, in the present condition of our knowledge. This information is more valuable because many of these works give definite dates and places of their composition, a thing rarely to be met with in Brahmanic and Buddhist works of the same age. But their chief contribution to our knowledge lies in giving much-needed information about the social conditions of India in the middle ages. The society, they represent, is, though formally one of Jain faith, is fairly typical of all the people in those times, and can be taken to be such, without any serious mistake14.
Annotations :
1. Paümacariya ii.
2. Jaina Sahitya Samsodhaka Vol. 3. No. 2. p. 170.
3. Mahābhāsya 4. 3. 87.
4. Keith, Classical Sanskrit Literature ch. i.
5. Ed. by Jacobi, Bhāvanagar. 1914. 6. Winternitz G. I. L. Vol. iii p. 570. 7. Kävyamālā Ed. p. 47. 8. Ed. Bhāvanagar, 1930. 9. Ed. Jacobi B. I. 1926. [Annals, B. O. R. I.]

Page #215

206
Amrita
10. Ed. Jacobi Z. D. M. G. Vol. 34. 1880.
11. Ed. Rajavijayaji, Benares 1916.
12. Ed. by Pandit in B. S. S.
13. Ed. in G. O. S. 1920. [Annals, B. O. R. I.)
14. The writer wants to acknowledge here his great obligations to his revered
teacher, Prof. A. N. Upadhye, M. A., Rajaram College, Kolhapur.
OOO
Narrative Literature in Jain Māhārāștrī
ABORI. XIV. 1935
22

Page #216

Didactic Works in Prākrit
The idea that poetry is only a vehicle of giving pleasure to the readers was never current in India and it has played no important part in colouring the writings of Indian poets. Like Plato and Horace in the west the guiding principle of both the poets and their critics was to instruct the readers in nearly all the fields of human activity. Pleasure was no doubt thought to be one of the important effects of poetry but instruction in subjects both religious and secular was thought of equal if not of greater value by all and the poets carried this view into practice with zeal in writing their works. Conditions like these should naturally lead to the production of a vast didactic literature which is actually the case in India. They were still stronger in case of Jainism as it was more or less an ethical religion which valued moral advice more than anything else.
Jain writers have written works of a didactic nature with a view to preach the fundamental principles of their religion and also to impart to their lay followers such instructions and moral advice as they thought fit in the interest of their faith both for its maintenance and progress. This advice included, in case of the lay followers, the bestowal of the alms to the wandering monks, the building of the temples dedicated to the Jinas, the erection of the alm-houses and other kindred institutions. In case of the monks the advice took a more theological and philosophical turn and included the usual rules of good conduct and such instructions as would help them in freeing themselves from the bonds of the worldly life.
It is not always possible to distinguish the works of purely didactic nature from others of similar contents, either from the theological and philosophical literature on the one hand or from the story literature on the other. Principles of ethics and worldly advice are not distinct enough to admit of a clear separation from other related topics and this is particularly the case with the Jain literature where the moral element takes precedence over every

Page #217

208
Amrita
other thing. The habit of giving illustrations to ethical rules had often led the author to migrate into the alluring province of story-telling while the usual habit of drawing moral conclusions from the current stories and anecdotes helped to bridge whatever small difference there remained. Even though the writers mostly refrained from giving full stories in their works or remained satisfied by alluding to the chief characters or the main incidents of the story, their commentators were not able to resist the temptation of giving them in full and sometimes at great length. This was not always without justification as they were sometimes absolutely necessary for the understanding of the text. In such circumstances what should lead one to call a work didactic should be the intention of the author to offer words of advice more than to give out a treatise on ethical matter or some collection of stories, which is usually found expressed in choosing an appropriate title to his work..
The presence of didactic verses in the Ardha-Magadhi canon shows us that this spirit of moralising was present in Jainism from the very beginning. Along with the rules of conduct intended for the monks, on whom a stricter code of behaviour was imposed, we find there verses giving advice which can be applicable to monks as well as lay men with the sole intention of helping them in behaving well in the present world., In works like the Uttaradhyayana, Daśavaikālika and others we find such verses in abundance. In the Jñätädharmakatha and other story-books of the canon the habit of moralising from popular stories is found considerably developed. In the first group of works we find the general morality of the prudent people of this world which Jainism shared along with other Indian religions while in the later group even though Jainism draws many of the stories from, the common stock of folklore current at the time, the fact of moralising from them is certainly its own peculiarity and the moral culled out from them is distinctly Jainistic and incorporated into it from outside. This didactic element is found to increase in the Niryukti works which are the earliest books besides the canon. There we find good many stories incorporated anew and this also fort the special reason of advising the readers. Gnomic verses are abundant and the germs of the anthologies can be traced in the fact that the treatment of various subjects is put together by chance coincidence and numerical. consideration. In the Niryukti on Daśavaikälika we find a chapter dealing with topics like the acquisiton of wealth and love and its fulfilment. In the Avasyaka-Niryukti we find a saying like 'a small fire, a young king and at young serpent are not to be slighted as being young' which can find its way in any anthology.

Page #218

Didactic Works in Prākrit
209
The oldest didactic work in Prākrit appears to be the Upadeśamālā of Dharmadāsaganin. Tradition makes him a direct disciple of Lord Mahāvīra which would make him a writer of the sixth century B.C., a thing highly improbable both in itself and from the nature of his work. Tradition further supplies us with some information of his personal life before and after his becoming a monk. He is said to have been a king called Vijayasena of the town Vijayapura. He married two ladies Vijaya and Ajayā, the former of whom gave birth to a son who was treacherously taken away by the second wife and cast away in a forest to die. As luck would have it, the boy was found by a shepherd who brought him up as his own son and named him Ranasiṁha. When the king came to know the whole story of the exposure of his son he became disgusted with the affairs of this world and turned out a monk. Then he took up his second name Dharmadāsa. His son Ranasimha suffered a good deal in his life as a boy at the hands of his ministers who drove him off from his kingdom. He wandered for a long time during which he showed great personal valour and was successful in getting back his kingdom, while winning many royal princesses as wives. His father now tried to enlighten his son of the vanity of this world and was successful with the help of the present work composed for that purpose.
: It is not possible to determine the veracity of the tradition preserved by his commentators. But the fact is clear that the writer is not a contemporary of Lord Mahāvīra himself. Both the language of his work and the references to later patriarchs and others make it evident that he lived much later than is supposed by the tradition. He is further found to make a good use of the various books of the Ardha-Māgadhi canon while supplying illustrations in his work. We are thus forced to conclude from the internal evidence of his work that the work is written by some later writer Dharmadāsa by name and he was probably confused with some disciple of Mahāvīra.
The work deals with the duties of monks and pious householder given without much consideration for system. After the customary salutation to the first and the last Tithamkaras of Jainism. Rsabha and Mahāvīra, the author advises his readers to wander about in this world without attachment as did the great sages of old, to cultivate good qualities like forbearance, compassion, discipline and many others which distinguished the religious teachers and to overcome all obstacles as they did. Like the devout Goyama the first disciple of Mahāvīra, we should always try to hear them preaching and act upon it. The value of the teacher is as great as of the moon among the stars. Even the best of the gods have to suffer death much more the other persons. Both

Page #219

210
Amrita
the flower and the fruit of the tree of passion is undesirable, the one is the bad thought and the other the wicked act. A pupil who has no respect for his teacher can gain nothing by remaining with a teacher. The object of one's liking is always good to one for, the tigress regards her young one to be meek and compassionate. One who accepts the words of his teacher will be benefited with it as with a dose of medicine. Even the deadly enemy is not able to do as much harm as attachment and hatred are able to do. Passions burn down the penance of a soul as does the forest-conflagration the forest.
Along with such moral advice the author has given us a lot of theological information. The condition of the soul is the real cause of the binding of the Karmas. In reality body and soul are quite distinct and for the sake of religion one should abandon even the body. There are the five Asravas which lead the soul through the worldly existence. Many topics that are usual in the philosophy of the Jain faith are enumerated at great length. The striking feature of the work however is the illustrations of mythological and historical persons to explain the moral rules. In this the author has drawn greatly on the canonical books in giving instances of Hariesa, Bharata, Bahubali, Sanankumāra and others and probably on the Niryukti literature in referring to the later patriarchs like Prabhava, Sthulabhadra, up to Vajra. Historical references appear to be, king Sanvahana of Benaras and his thousand wives, Kūniya's murder of his father. Seniya, and the fact that Canakya the teacher of Chandraguptá killed king Parvataka. In some part of the work the author appears to muse and advise his own self to meditate and to be careful.
The work really ends with verse 537 which gives in a cunning manner the name of the author. This is however followed by a small praise of the book itself and the importance of studying it which is certainly added by some one else afterwards. Nearly the whole of the book is written in Gäthäs the usual metre of the non-canonical works. There are however, a few verses in Anustubh and Rathoddhată many of which are quotations from earlier works like the Uttaradhyayana and the Niryuktis. On the whole the work shows no great poetic merit. In few verses there are similes like that of an actor for the wandering soul which are trite and often used. The list of allusions is however long and gives credit to the author for his knowledge of the canon. The language is simple but lacks the quality of elegance and smoothness. That the work was popular from comparatively very early times" in clear from the number of commentaries on it. As old as the tenth century is the commentary of Siddharsi the famous writer of the Upamitibhavapraprapancakatha in Sanskrit. In A. D. 1161 Ratnaprabha wrote another

Page #220

Didactic Works in Prākrit
211
commentary called Doghattītīkā at Bhrgupura and was revised by Bhadreśvarasūri. Another one is written by Rāmavijayaganin who gives at great length the stories explaining the allusions in the text.
A more bulky work is the famous Upadeśapada of Haribhadra who calls himself the son of Yākinī and who uses the catch-word Viraha at the end of his works, both the peculiarities being met with at the end of the present work. The work extends over 1039 gāthās written continuously without any divisions into chapters. Having saluted the Jina he expresses his wish to offer a few words of advice culled out from the scriptures preached by the Jinas for the enlightenment of men weak in intelligence. He starts with the fact that human life is difficult to get and illustrates it with examples which are taken from the scriptures. Similarly he deals with the right faith, the relation between the teacher and the pupil, the four types of intellect and their examples, the usual vows of non-killing and others and even enters into the technical field of Karmas and their workings. He concludes his work with the advice of serving the knowers of the doctrine, hearing their advice, giving of gifts, abstaining from tormenting others, not to entertain thought of pleasure, constant thought about the nature of the world, to honour worthy people, not to insult any one in the world to observe the rules of people, not to blame others; to have attachment towards virtue, not to associate with wicked persons, and always to avoid passions and carelessness.
But the major portion of the work is devoted to the summary statement of the stories to illustrate the various precepts, and in this Haribhadra follows the curious method of the Niryuktis in which only the most important facts of the narration are given in a mechanical way. The hints are sometimes so meagre that they remain obscure unless the reader is acquainted with the story which they profess to epitomise and naturally the commentator Municandra has given them at great length in Prākrit verses except for one long story of Ratnaśikha in prose. The source of all these tales given by Haribhadra is the story books of the canon and of later works the Āvaśyaka Niryukti, on which he has drawn not a little.
Unlike other works of Haribhadra the present work does not make a delightful reading both on account of its method of giving the stories in the most uninteresting manner and the consequent roughness of the language in which sometimes a mere string of uninflected words follow one another. Many of these stories are further mere suppositions without life in them as in case of the Drstāntas to point out the difficulty of obtaining human life.

Page #221

212
Amrita
The only merit of the work appears to be the putting together of many stories from diverse sources and giving them in a form easy for memorising.
A later work of a more refined nature is the Puspamālā of Hemacandra. He belonged to the Maladārigaccha and was a pupil of Abhayadeva. Of him Rājasekhara tells us that he was originally a royal minister called Pradyumna and was converted to Jainism by Abhayadeva along with his four sons. His co-student Śricandra praises him in the following words. 'Hearing his greatness in preaching the religion, king Siddharāja approached him in the temple of Jina to hear him preaching. He caused Jayasimha to exert greatly to increase the importance of Jainism in Gujarat and in building temples crowned with golden sikharas. Hemacandra visited the holy places of Girnar and satruñjaya and died with the usual method of fasting for seven days. From this it is evident that the literary activity of Hemacandra falls in the reign of the famous king Jayasimha who ruled at Anahilapātana from A. D. 1094 to 1143 A. D. This writer is otherwise famous for his voluminous commentary on the Višesāvasyakabhāsya.
His Puspamālā consists of 550 verses in the gathā metre and deals with some twenty topics of usual interest to the Jain monks and laity. The author appears to have named his work as Upadeśamālā which he compares with a garland of best flowers and hence the popular name to distinguish it from other works bearing the same title. It begins with a salutation to the Jina and then deals with Ahimsā which, according to the author is the best part of the religion of the Jinas (4). Its sequel the Jivadayā is then praised and it is pointed out that those who neglect it wander through the forest of life (8, 12). Then follows the praise of knowledge, conduct and right faith, in which the author indulges in giving many technical details and refers to ancient sages and kings (103). Topics like the begging of alms and the praise of a perfect monk (220) and the transitory nature of this world (444) are also included.
The work distinguishes itself from other works of the like nature by its ornate style and some poetic ability on the part of the poet. While describing the perfect monk he presents us with a series of similes to bring into prominence the various qualities of the monk, but many of them are trite and drawn from the canon. Some of his expressions are happy and concise while use of puns to base the figures of speech is also found (32). At the end the writer expresses the pious wish that his work should receive a cordial welcome for all the time to come.
Chiefly famous for the didactic works is the king-poet Asadha. He was

Page #222

Didactic Works in Prakrit
213
of royal descent. For his life we are chiefly to depend upon the tradition preserved by his commentator Balacandra who comes only two generations after him. He tells us that Asaḍha was the son of king Kaṭukarāja and his queen Analadevi, who were ruling over the town of Bhinamāla, the ancient capital of Gujarat and situated on its borders. He had a brother Jāsaḍha. Asadha had two wives Pṛthvidevi and Jaitalladevi the latter of whom gave birth to two sons the elder Räjada and the younger Jaitrasimha. From his first wife Asadha had a son Arisimha by name. Rajada the eldest of his three sons, died at a very early age and this incident turned the bereaved father into a poet for composing his two didactic poems. Abhayadeva a pupil of Bhadresvarasūri, who in turn was a pupil of Devendra, was able to console the bereaved father of his grief at the death of his eldest son, which was brought about by the advice which his royal disciple put down in the form of his works. There appears nothing improbable in the tradition and taking into consideration the date of his commentator which is A. D. 1221 who preserves it we can accept it as a true account in the main. But the nature of the books is such as to possess nothing distinctive about them and of the abnormal situation in which they were composed. They lack the warmth of emotion which one naturally expects out of such conditions. The fact that Åsadha was the son of Katukaraja and a pupil of Abhayadeva is corroborated by the concluding verses of both of his works.
Our author is given the honorific title of Kaviśobhāśṛngāra and is said to have lit the fame of Kālidāsa by composing his comment on his famous Meghadūta. He lived at the end of the twelfth century and the beginning of the thirteenth as his Vivekamañjarī was written in A. D. 1191.
The Vivekamañjarī consists of about 140 verses in gäthä and sloka metres. It begins with a salutation to the last Jina and purports to b instructions to the souls who do not know the real nature of the life becaus of the absence of the lamp of discrimination and therefore this want is to be filled by the present Vivekamañjarī. The author then gives the scheme of his work which is to include the usual Jain rites like the Causarana, cultivation of the good qualities the censure of wicked deeds and the reflection on misery which he regards the seeds of the purity of the mind (6). This scheme he carries consistently throughout the work. He points out the real Mangala which consists in offering salutation to the five dignitaries of the faith and enumerates the 24 Tirthankaras (11-12). In addition he refers to other Tirthankaras both past and present and requests them to remove his sin. Then follows the praise of the disciples of Mahāvīra like Goyama (22), and other

Page #223

214
Amrita
mythological persons like Bharata (27-28), Bāhubali, Sanankumāra, Gayasukumāla, Thūlabhadda (33), and many others both historical and legendary. The mothers of the Tīrthankaras and other great women famous in the Jain mythology are enumerated and the whole of the early part of the work is a mere table of names. Then follows a description of the religion of the householders and the monks with its usual vows and other observances. Only at the end we find the Jiva directly addressed and some advice given which would suit the occasion of the work. The tract concludes with the statement that the substance of the preaching of the Jinas is found in the Namokkāra (129).
The style is simple but most of it reads very dry as the major portion of the work is devoted to the enumeration of the names. Only a few show some kind of grace (55) and in one (35) the author tries to make use of the device of alliteration. Much of the matter and even the wording is the common property of the Jain community and there is nothing distinctive about the work. His other work the Upadeśakandali consists of 125 verses and is probably of similar contents. On the whole Asadha appears to be no poet but a mere compiler of ordinary calibre.
Another writer about this time or a little earlier is one Paramānandasūri who is said to be a pupil of Devabhadra. At any rate he is earlier than A. D. 1253 the date of his available Ms. His Prākrit work is called the Hitopadeśāmrta and extends over 525 gāthās. The author himself has written a commentary on it.
At about this time lived Yaśodevasūri who was most probably a pupil of Uddyotanasūri who was a co-student of Municandra. He wrote his Parithasanthiyana or Suyanaryana in A. D. 1232. A writer of the name of Yaśodevasūri but of doubtful date has written a Prākrit didactic work called Dharmopadeśamālā of some hundred gāthās. It is not certain whether he is identical with the previous writer. His work begins with a salutation to the Śrutadevi and gives many illustrations from the annals of the Jain church history.
At the beginning of the fourteenth century lived one Jayakīrti who calls himself a pupil of Jayasimha. Nothing is known about him except that his work is commented upon by one Somatilaka who wrote the Tikā in A. D. 1337 and named it Tarangini. Jayakīrti names his work Silopadeśamālā which is a small tract of 144 verses. It is a glowing praise of Sila or good conduct and its importance in getting the final goal of religion. sila is the source of prosperity both in this as in the next world and with its help alone the sages

Page #224

Didactic Works in Präkrit
215
of ancient times obtained liberation. The breach of good conduct however is equally disastrous and many have suffered on account of it. In course of his advice he quotes a verse from the Daśavaikälika as being the Agama. As usual the work is replete with allusions to the ancient sages and mythical persons. This fact has given the commentator an opportunity to load his comment with long stories about them which alone gives some value to it.
Silopadeśamālā is better in many respects than the works of Asadha both in style and poetic ability of the writer even though it reaches no high mark in both. The author has expressed himself in clear and not uninteresting terms and it makes a smooth reading.
Upadeśacintāmaņi is another Prākrit work of this nature and is written by Jayasekharasūri. He was a pupil of Mahendraprabha and a co-student of Muniśekhara and Merutunga. He wrote along with it a commentary on the work in A. D. 1379. The book is divided into four chapters and contains moral advice. A little later lived Munisundara who wrote the Upadeśaratnākara. He belonged to the Tapāgaccha and was given the title
lisarasvati. Born in A. D. 1379 he became a monk in 1386, a Vācaka in 1409, Sūri in 1420 and died in A. D. 1446. His book consists of three chapters called Tatas each of which is divided into four Ambas with a further sub-division into Tarangas. The author has written a commentary on it, but it is available upto the end of the second chapter.
In conclusion we can note the less known works, the Dhādasī gāthās 36 in number of an unknown author but earlier than Śrutasāgara who quotes a verse from it, and therefore earlier than the 16th century; a Vairāgyaśataka of 104 verses teaching renunciation by describing the horrible state of this life; the Indriyajayaśataka with 102 verses, the Sambodhasaptatīkā of one Jayasekhara; the Hitacarana of Sakalacandra written in A. D. 1573; the Dharmopadeśa of Laksmīvallabha written in A. D. 1688 and consisting of 107 verses and many smaller Kulakas of little importance.
This brief review of the didactic works is sufficient to show that these works are not so much important for their poetic quality in which all of them deplorably lack, nor for the short and pithy statements full of wisdom which make the Sanskrit works of this type so much valued, but for the sincerity of the writer's mind who believes firmly in the creed he is preaching and has devoted his life for its cause and earnestly wishes the welfare of his readers.
Didactic Works in Prākrit
JUB Vol. VI, Part III 1937

Page #225

Hymns in Prākrit
Jain writers have written a considerable amount of hymns or Stotras of no small merit. As in other departments of literary writings, they have tried to vie with Brahmanic poets in composing hymns of devotion, and have vied with considerable success. Ever since the days of Rgveda, the production of hymns addressed to various gods was continued by Indian poets, and this activity has resulted into a vast amount of hymnal literature of no mean value. In its long and eventful career, however, many changes were effected both in the subject matter of these Stotras and in the spirit of devotion underlying them. To the gods like Varuna, Sūrya, Indra and Vişnu, new gods like Krsna, Rāma, and Durgā, were added, either taken over from the gods of the original inhabitants of India, or from local deities of various places, claiming greater attention. The Jain writers also offered panegyrics to the prophets of their own religion, who soon acquired a position of gods to whom prayers and worship can be offered, and still later, their immediate disciples like Gotama and others acquired their own share of praise. .
But along with this widening of the subject matter of these hymns, a similar change came over the spirit underlying them. The naive simplicity and the matter-of-fact dealings of the Vedic poets gave place to a more and more spiritual and devotional sentiment, in which the poet raised the object of his devotion to a higher plane and lowered himself down to a great extent, a fact aided by the new theory of Bhakti or devotion. Similarly while praising the gods of one's religion the poet passed with imperceptible steps to the principles of his creed, and so the more philosophical and logical hymns were produced. And when once the fact that a short hymn can be used to express in a succinct form a particular dogma or a few outstanding ethical principles in a charming and attractive manner was realised, the hymns began to draw their subject matter from all the varied departments of knowledge from great philosophical problems to a list of mythological personages.

Page #226

Didactic Works in Prākrit
217
The best part of this hymnal literature of Jains is no doubt to be found in Sanskrit, but a good deal of it is also to be met with in Prākrit languages. Even though Sanskrit on account of its greater popularity and circulation enforced itself on the heterodox religions, and soon became the sole vehicle of expression, Prākrit was still continued for writing hymns and other works by the Jain writers for a long time, and this activity has produced many Stotras worth attention.
The oldest of these Prākrit Stotras is undoubtedly the UvasaggaharaStotra attributed to the high priest Bhadrabāhu. Bhadrabāhu, we know, was a great pontiff of the undivided Jain community, and lived at the time of Candragupta the Maurya. One of the traditions says that he migrated to the south to avoid the great famine that overtook the country of Magadha at his time, and died in the country of Mysore, while another of them makes him retire to the country of Nepal, at the time of his death. The attribution of the present Stotra to this sage is, however, not very accurate. But there is also nothing to state against it except the fact that the language of the hymn is much modern and certainly later than that of the Niryuktis. This short Stotra contains barely five stanzas in Gāthā meter, praising and eulogising the 23rd Tirthankara Pārsva, mainly pointing out his characteristics of helping others and rescuing his devotees from distress and difficulties. This fact probably led to the magical use to which this hymn is put by later writers, and a commentary on it tries to interpret it in the spirit of a spell to be used against many a calamity of various kinds like snake bite and others. The name of the hymn is also obviously derived from the beginning words, a device used very much by the Jain writers in naming their hymns.
Of a similar nature appears to be the Bhayahara Stotra of the celebrated poet Mānatunga whose other Sanskrit hymn the Bhaktāmara-Stotra is much more popular and certainly the better of the two. Both the sects of the Jain community claim him as belonging to their creed, and his Sanskrit hymn is commented upon by the writers of both the sects. According to a very widespread tradition, he was a contemporary of the two famous poets Bāna and Mayūra, and composed his hymn to show the great supernatural power of his deity. But it is doubtful whether he is to be put so early. Jacobi's conjecture that Siddhasena wrote his Kalyāṇamandira-Stotra on the model of this hymn has nothing to recommend it, while it is just possible that Mānatunga comes later than this celebrated logician, and may have lived about the 8th or 9th century. His Bhayahara is also addressed to Pārśva who appears to be the most favourite of the Tīrthańkaras to receive praise.

Page #227

218
Amrita
Another of his hymn is called the Paramesthi-Stavanam and contains only 35 Gāthās.
Nandisena appears to be also a very old writer. Winternitz would place him as early as the 8th or 9th century. At any rate he is earlier than Jinaprabha (12th century), who calls him a very old and celebrated poet and considerably earlier than him. Tradition would make him a direct disciple of Lord Mahāvīra himself, a fact scarcely to be believed. His poem called Ajitaśānti-Stava is a hymn of praise addressed to the two Tirthankaras Ajita and sānti, and contains in its present form 40 verses, but which originally contained according to the opinion of his commentator, only 37 stanzas. Tradition further tells us that, while the poet was on his way to the pilgrimage of the mount Śatruñjaya where both these Tīrthankaras spent their rainy season, he conceived the idea of composing the present hymn addressed to these prophets.
We are more fortunate in knowing something of Dhanapāla the author of the Prākrit hymn Rsabhapancāśikā. Originally a Brahmin by birth, he was converted to the Jain faith by his brother Sobhana who is famous for his Sanskrit hymn in a very ornamental and scholarly style. Dhanapāla's father was Sarvadeva. Dhanapāla appears to have been patronised by the two kings of Dhārā Siyaka and Vākpati, even though Merutunga relates in his Prabandhacintāmani another tradition which makes him a contemporary of Bhojadeva, the famous king of Dhārā and adds something about his family relations, about his quarrel with his brother and the final reconciliation. From the conclusion of his Prākrit lexicon Pāialacchināmamālā we know that he lived and wrote that work for his sister Sundari about A. D. 972-73. Afterwards he was converted to Jainism and then composed his hymn to show his respect for his newly adopted faith.
The best of Dhanapāla's hymns is undoubtedly the 50 stanzas in honour of Rsabha the first prophet of Jainism. It is composed in Gāthās and describes in its first part (Gäthās 1-20) some important incidents in the life of Rsabha, like his birth in the house of Nābhi, his ceremonial bath on the mount Aştāpada, his promulgation of the various arts and crafts, his bountiful gifts, his Dīksā ceremony, his subduing the non-Aryan tribes, his obtaining the omniscience, and things like the Samavasarana which is in fact a general theoretical outline of the life of most of the Tirthankaras. In the second part of the last 30 verses the author has greater freedom of thought in describing the greatness of his religious teacher and shows great poetic skill in depicting

Page #228

219
Hymns in Prakrit
it. He points out that Rṣabha conquered Madana, which is a clear indication of his greatness over Brahmanic gods like Hara and Hari, and his help to his followers to get final emancipation. In verse 37 we find a fine poetic idea. based upon the two meanings of the word Payattha which means both the principles of Jainism and ordinary objects. He also exhibits his poetic ability in many other ways and uses many a poetic figure. The language of the hymn. is also no doubt Jain Mähäräṣṭri but here and there a few Desi words like Lattha and a few Apabhramśa forms like Paim are also found. Moreover, the author's command over Sanskrit can be easily seen even in his Prakrit.
His other Präkrit hymns like the Vira-stava in 30 Gäthas and another of the same name are not of great importance. The second is a curious attempt of using both Sanskrit and Prakrit in one and the same verse, in which the first line is in Sanskrit and the second in Präkrit. It contains some ten Gāthās, describing the heroic deeds of Mahāvīra, like his removing the Meru mountain by the touch of his foot, defeating the gods who had come to test his powers, his great scholarship and the ridiculous position of his teacher and many others. But the hymn has no other merit nor any poetic quality that will charm our minds.
Abhayadeva who is famous for his Sanskrit commentaries on the 9 of the Angas of the Ardha-Magadhi canon, has also written a few Stotras. He was a pupil of the equally famous scholar Jineśvara, and his literary activity falls in the middle of the 11th century A. D. He was the son of Maheśvara and Dhanadevi both of them being residents of Dhārā. He is reported to have converted Sankaradāsa a Brahmin teacher of the Paramāra kings of Dhärä, in A. D. 1054. Besides his famous Apabhramśa hymn the Jayatihuyana he has also written a Mahavira-Stava in 22 Anustubhs, giving the various qualities of Lord Mahavira, in a purely dogmatic enumeration of his excellences which makes it dry and devoid of all poetic beauty and charm.
Next we come to the prolific and scholarly writer Jinavallabha. His teacher was Vardhamana, whom he succeeded on the chair of the pontiff, and died in A. D. 1110. He lived at Citrakūta and was helped by his friend. Devabhadra in his work. Among the great number of his works he has composed a number of hymns addressed to the various Tirthankaras. His Pārsvanatha-Stava is composed in 22 Gāthās. Therein he praises the ability and readiness of Parsva to render service to all people in troubles, and in giving the desired objects of his devotees. The hymn shows a fine use of the power of alliteration and beautiful poetic ideas, while the language is elegant

Page #229

220
Amrita
.
and simple. (Vs. 8, 17). Another of his hymns is the Panca-Kalyānajina-Stotra also called Caturvimsatijina-Stava in 26 Gāthās, and is much inferior to his previous song. It describes the various characteristics of the 24 Jinas in the usual enumerative fashion. His Vira-stava is composed in 44 Mālini verses and is tolerably good. In his Ullāsikkama-stava he praises the two Jinas Ajita and śānti.
One of the famous pupils of Hemacandra is Rāmacandra who succeeded him and met a cruel death at the hands of the successor of king Kumārapāla. Rāmacandra is more famous for his Sanskrit dramas and dramatic theory, while many of his hymns are in Sanskrit. The only hymn of his in mixed Sanskrit and Prākrit is the Adideva-Stava a small piece of 8 Gāthās and not a great production.
Dharmaghosa is chiefly famous for his many hymns. He belonged to the Tapāgaccha; Devendra was his teacher and Vidyānanda one of his costudents. He often calls himself by his other name Dharmakīrti, often occurring at the end of his poems. He became Upādhyāya in A. D. 1263, Sūri in A. D. 1271, and died in A. D. 1300, enjoying a long life.
His chief work is no doubt the Rsimandala-Stotra, which is not exactly a hymn but rather approaches a chronicle or a Pattāvali. It purports to give the praise of the various sages who obtained liberation or various 'heavens, as the result of accepting asceticism and the life of a monk at the hands of the various Tīrthankaras. The scope of the work is very wide. The author gives the various sages that were converted to Jainism by Rsabha and others, the contemporary monks of Ajita, Sāgara, Malli and the vast number of the pupils of Mahāvīra. Much of his information he draws from the canonical books like Nāyādhammakahão, Viāhapannatti and others, while the successors of Mahāvīra upto Devardhigani and Munivrsabha he draws from the Kalpasūtra and the Nandi-Sūtra-Pattāvali. In the verses of praise of these various sages the author gives suggestions about their personal lives. He deals at great length with the lives of Sthūlabhadra and Vajrasvāmi the two patriarchs of great popularity. The language of the hymn is easy and the author describes the sages in as simple a manner as possible. But the vast amount of mythological allusions makes it a little difficult to follow them unless the reader is already familiar with them.
Other hymns of this author are of no great value. There is a short collection of Stotras called Jinastotrāni giving the various births of Rsabha, Candraprabha, śāntinātha, Munisuvrata, Neminātha, Pārśva and Mahāvīra, the

Page #230

221
Hymns in Prakrit
Ajitasäntistava in 17 Gāthās and a host of others. The Pañcavi-mśatigunastavanam describes the 25 excellences of the preachings of the Tirthankara, while a panegyric is addressed to the present period of time in Duṣamākālasamstavanam.
More well-known as a scholar is Jinaprabha who calls himself a pupil of Jinasimhasūri, who lived at the beginning of the 14th century. His Rsabhadevajñā-stava in 11 Gāthās is a hymn which describes the importance of following the rules laid down by Rsabha and its praise. More artistic and full of elaborate puns is his Pärśvanäthalaghu-Stava in which Pārsva is compared with the 9 different planets by the use of paranomistic adjectives, each verse thus yielding two senses. His Gautama-Stotra bestows praise on the first Ganadhara while the Pañcaparamesthi-Stava eulogises the five divinities so often praised in Jainism.
About Dharmanidhāna we know next to nothing. The concluding words. of his hymn would make us believe that he was possibly one of the pupils. of Udayaprabha and so lived in the 14th century. His hymn is called Caturvimsikastavana having 27 verses in Śärdülavikrīḍita metre and gives in a schematic form the lives of the 24 Tirthankaras while at the end he collects together in 3 verses their emblems. On the whole the hymn shows no great merit. At about this period lived one Devendra who calls himself a pupil of Sanghatilaka of the Rudrapalliya Gaccha and wrote two hymns, one the Adideva-Stava giving the praise of Ṛsabha in 9 Gäthās and another called Sāśvata Caitya-stava in 24 Gāthās.
Jinabhadra, the author of the Dvādaśängipadapramāṇakulaka is to be distinguished from the more famous and ancient Jinabhadra, the celebrated author of the Viseṣävasyakabhäşya. The present writer belonged to the Kharataragaccha and lived upto A. D. 1418. The hymn contains 21 Gäthās and is devoted to the description of the 12 Angas including the 14 Purvas now lost to us. The information adds nothing to our knowledge of these works and he follows the tradition with strict adherence.
In the 15th century lived Somasundara a very famous and prolific writer in Jain literature. Besides commentaries and other independent works he has written five Stotras addressed to Rṣabha, Nemi, Säntinätha, Pärśva and Mahavira. The chief characteristic of these short hymns is the variety of languages used therein, which are Sanskrit, Präkrit, Śauraseni, Mägadhi, Paiśācī, Cülika-Paiśăci and Apabhramśa, a feature found in many Jain Stotras and they show a considerable amount of metrical variety. A similar feat is

Page #231

222
Amrita
performed by his pupil Ratnat Sekhara whose hymn Caturvimśatijina-Stavana in 24 Mālinī verses and one concluding Śārdūlavikrīdita is composed in a manner so as to be considered written in either Sanskrit or Prākrit or Sauraseni, a fact which the Sanskrit rhetoricians call as Bhāsāślesa. Another of his pupils Jinakīrti wrote Namaskāra-Stavana in A. D. 1437 with a commentary on it.
OOO
Hymns in Prākrit
JBU. III. 3. 1934

Page #232

A Few Parallels in Jain and Buddhist Works
The two canons, of the Buddhists in Pāli and of the Jains in ArdhaMāgadhī, present us with a few interesting parallels worth consideration. The study of such parallels is interesting both for its own sake and for the light it throws on the problem of the relation in which these two religions stand with each other. The real explanation of the similarity found therein, whether it is a case of borrowing or one of common inheritance or even one of accidental coincidence, is to be decided in each particular case by considerations of its individual peculiarities. And as such their examination will help us in forming an idea about the exact relation in which these
religions stand, particularly in their literary traditions. . Both the religions, Buddhism and Jainism, arose in the same country
of Magadha and at about the same time. As such they partook of the same surroundings which goes a long way in determining many of their common features. But besides this general similarity of spirit and form which can be explained as due to the influence of the time-spirit we find something more to think of in the present case. The canons of both these religions show similarity not only in the general moral and disciplinary tone due mainly to the fact that they embody the same general principles of ethics which are common to both these religions which is in its turn due to the circumstance of their birth and early growth, but also in matters of composition and wording which requires something more to explain them. They raise the important question of borrowings and the authentic nature of one tradition as against the other and the question of their respective age.
Even though it is now admitted on all hands that Jainism as a religion arose a few decades earlier than Buddhism, or even a few centuries before it, if we accept the view that not only Pārsva was a historical person but that the traditional date of his birth and death is equally trustworthy, a fact not beyond reasonable doubt; the question of the formation of the two canons of

Page #233

224
Amrita
these two religions stands on a very different footing, and is in no way connected with it. It is yet very difficult to believe that the present ArdhaMāgadhi canon, which tradition itself admits to have suffered much recasting and reduction, and which has the still greater disadvantage of being repudiated by the whole of the Digambara community of the Jains, can be reasonably attributed to a period to which the Pāli canon of the Buddhist is attributed at the latest!. It is true that the Pāli tradition also shows us the Buddhist canon as going through the similar stages of redactions at various stages of its history, but their last council falls in the reign of king Asoka in the third century B. C.?. while the last council of the Jains comes in the fifth century A. D. in the days of the kings of Valabhi?. So if we are to believe in these traditions alone it is clear that the Pāli canon will have to be put much earlier than the Ardha-Māgadhi one.
This problem of the relative priority of the two canons is further rendered more difficult and complicated by the supposition of an ArdhaMāgadhĩ canon earlier than the present one, and a similar canon of the Buddhist which again according to Lüders will have to be supposed to be written in old Ardha-Māgadhī4. All such speculations have no doubt some indications in the present canons themselves and can on, that account be said to rest on facts. In the case of the Jain canon we even possess an outline of the older canon preserved to us with more or less accuracy, while it will be very unwise to put down the whole of the Pāli canon at the time of Asoka. But it is equally true that it is not possible now to separate them from their later additions with anything like certainty. For the present purpose then of comparing a few parallel passages from the two canons it is better to set them aside and to start without the supposition of earlier works not to be found to day, even though the parallels themselves are adduced to prove their existence.
To begin with, we have a number of stories common to the two canons which we take for consideration. Of all the works of the Jain canon the Uttarādhyayana is the most important as it preserves many interesting stories and parables which are also to be met with in various works of the Buddhist. Here we meet the story of the two persons called Citra and Sambhūta who were fast friends at the beginning and wandered a series of lives together but at the end suffered very different fates because of their characters. This story is also found in the Jātaka collections in the CittaSambhūta-Jätaka“. As pointed out by Dr. J. Charpentier the two chapters show similarity not only in the general outline of the story and its main incidents

Page #234

A Few Parallels in Jain and Buddhist Works
225
but even in the verses found in them which are common to both the books. The story in the Uttaradhyayana is in verses only while the one found in the Jātaka books is in mixed verse and prose as usual. This fact along with the fact that the Gathas of the Jātaka are decidedly older than the prose which is very late as can be seen both on linguistic and logical grounds would lead us to suppose that the story as preserved in the Uttaradhyayana is the older of the two. But the Jain version says nothing of the earlier lives of these two friends which are however referred to in their conversation'. The Jātaka gives us all the details about this earlier part of the story which cannot be regarded as a later modification of it or an addition to it. So also we have a few cases in which the order of the verses in the Jātaka books appears more in accord with the general trend of the story than the one found in the Jain version. This is to be explained on the supposition that the Jain version has suffered in its arrangement while the Jātaka books were more fortunate in having a commentary which numbered its verses very early and arranged them rigidly which has saved it from any further change. On the other hand the Jain version also lost the earlier part of the story which is preserved to us in the prose of the Jātakas even though it is put down much later than the writing of the chapter in the Uttarädhyayana.
Another story common to these two works is that of IsukāraR in the Uttaradhyayana and the Hatthipālajātaka in the Jätaka books. The story relates that a king and his preceptor had no son, but with the help of a treespirit the Purohita was able to obtain four sons who were all religiousminded. To test their zeal in the matters of religion and to know for certain whether they will live in the worldly life or not both the king and the Purohita approached them in the garbs of monks and found that all of them turned out monks. This led the wife of the Purohita and himself to take up to asceticism and consequently the king and the queen also do the same thing. In this case it will be seen that the story of the Jātaka books is fuller and gives many details about the birth of the four sons of the Purohita which are wanting in the chapter of the Jain work. There the story begins abruptly with the statement that all the characters in the story were born in the same town descending from their heavenly abode. Another difference between the two versions is about the number of the sons the Purohita had, they are four in the Jätaka while only two in the Uttaradhyayana chapter. This fact again is made use of in the Jätaka books to give rise to four different occasions for the renouncing of this world by the four sons and taking to ascetic life, which is occasioned by seeing the king and the Purohita in the garbs of a monk,

Page #235

226
Amrita
and the repetition of the same situation for four times. In the story as preserved in the Uttarādhyayana there appears to exist no relation between the Purohita and the king, while in the Jātaka they are represented as consulting each other and plotting together to test the sons of the Purohita as to their intention of becoming monk. In the version of the Jains it appears that when all the members of the family of the Purohita took to monkhood, the king came in the possession of their property according to the rule of the Dharmaśāstras. This occasioned a complete change in the mind of the queen who turns her mind to nunhood and also advises the king to the same effect. This fact appears to be more natural and appropriate than the one found in the Jātaka books. Considering further, facts like the story of the four sons acting exactly like each other, their long-drawn moralising, the improbable story of the spirit in the tree in the beginning which is pleased to give four sons to the Purohita alone and not a single one to the king who was in greater need of an heir, and the curious relation between the king and the purohita, it is clear that the Jain version is not only earlier but is better preserved and the more interesting of the two.
The story of a low-caste man attaining to a high position and showing the ill-founded faith of the Brahmānas in their idea of greatness in birth is found in the Hariesijjaħ10 of the Uttarādhyayana and the Mātangajātaka'l of the Buddhists. The two stories show a good deal of divergence in all matters except the central idea of the approach of the Cāndāla to the feast of the Brahmāns the wrong treatment given to him by the priest in one case and by his own son in the other, the sound beating they receive at the hands of the demi-gods who attend on the Mātanga, the approach of the woman the daughter of the king of Pāñcāla in case of the Jain version and of a merchant in case of the Buddhist version, her revealing the greatness of the Mātanga, and the recovery of the Brahmāns from the illness. And it is interesting to note that this part of the story is to be found in the verses which are common to both the versions to a great extent. The Jain version adds. little to the body of the text but the commentator12 gives us the back-gound of the whole story. He relates how the daughter of the king of the country of Pāñcāla went to a temple and saw there the Cāndāla whom she abhorred. But a spirit possessed her and to get her out of its clutches the king became ready to give her to the same low-caste man. But the sage refused to marry as its being against his monkhood. Now once he goes to the sacrificial ground of the Purohita of the king where he is refused food. And there the story begins in the Uttarādhyayana. The story in the Jātaka is much more

Page #236

A Few Parallels in Jain and Buddhist Works
227
complicated and expanded. There also the daughter of a rich merchant meets a Mātanga and feels disgusted at his sight. The man is beaten by her servants. But he goes to the house of the merchant and lies at the door until he is given the same daughter in marriage which is done at his persistence. A son is born to them, but in the mean time he becomes a sage and to confer prosperity on his wife makes the people believe that she is the wife of the great Brahmā. While the son grows old and is worshipped by all the people, the main incidents of the story happen. From this the Jātaka proceeds to give another story of the same Mātanga only because it also deals with the same theme. Otherwise it has no connection with the main story. A comparison of the two versions will make it clear that the Buddhist story is much more elaborate and of mixed motifs. The Jain version, on the other hand is much more simple and to the point. But there is one consideration which should lead us to think that the Jain version is the older of the two. On a careful reading of the Buddhist story it is seen that the attitude behind it is much more haughty and full of bitter feelings than the one which accentuated the writer of the Jain version. This can be seen in the facts like the plain deception of the Brahmāns and the administration of the food as a cure for the beating. This must have also led the writer to include the other story in the same Jātaka. And such an attitude must have arisen in later times as the effect of sectarian bias. The original motive of writing such stories appears to be to show the hollow foundation of the greatness claimed by the Brahmāns on account of birth alone. And this is clearly seen in the Jain version and in a much more humane and sympathetic form.
. Another book of stories in the Jain canon is the sixth Anga called the Jñātādharmakathā. Here also we meet with a few parallels in the Buddhist works. The illustration of the tortoisel3 in it has a clear parallel in the Samyutta-Nikāya II 14 which by the nature of the case appears to be a very ancient simile developed into an illustration and used by both the religions for the specific purpose of moralising on the control of the senses. The story15 of the two brothers going on a voyage and suffering a good deal at the hands of the deity on the Ratnadvīpa who killed so many ship-wrecked people and the winged horse who helped them in flying from that island, and the fall of one of the two brothers because of the temptations of that Yaksini has a parallel in Valāhassa-Jātaka6 where the part of the winged horse is played by the compassionate Bodhisatta.
Of greater importance is the story in the chapter of this book called the Amarakankā" which related the Jain version of the Brahmānic epic of the

Page #237

228
Amrita
Kuru family considerably changed and modified to suit the Jain religion. In tracing the early life of Draupadi we find a very curious story of a girl called Sukumārā which has a very distinct parallel in the story of Isidāsī8 in the Therī-Gāthā. The story tells of a girl who sinned in giving bad food to a monk, and as a result of which she was born in her next birth with an unpleasant touch of her body. She was married to a son of a merchant but in the very next day of their marriage he ran off to avoid her unpleasant touch. She returned back to her father's house where her father married her a second time with a monk who had come to his house to beg food. He also ran off the next day and the girl disgusted at her life took to the life of a nun. The Buddhist story of the unhappy girl is materially the same. The form in which the story is found in the Jain version is fragmentary and incomplete and is relegated to the back-ground, showing great inferiority to the other version in the Therī-Gāthā which is more complete and better told. But against the natural supposition of regarding the Buddhist version as earlier on account of its artistic superiority we have many textual indications in the Pāli version itself which go to prove a very different result. As remarked by Mrs. Rhys Davids 19 the whole spirit of the poem is non-Buddhistic and shows many traces of Jain tendency of valuing mortification and penance as more important. To add to this, we find such a technical term as Nirjarā20 used in the poem, and the name of the teacher of the unhappy girl is given as Jinadattāli not without significance. So it is more than probable that the writer of the Pāli poem had before him some Jain version of the story even though it may not be identical with the very meagre survival of it in the sixti Anga of the Ardha-Māgadhi canon. The Pāli tradition itself admits that the present poem in the Therī-Gātha is much later than the bulk of the work and was introduced into the collection by the Samgītikāras.
Two other stories in the Jñatadharmakathā have parallels in the anecdotes told by Buddhagosa in his Visuddhi-Magga. Considering the nature of his work and his usual method of relating stories from earlier literature we can fairly suppose that the present two stories are also drawn by him from the canon or possibly from the older Atthakathās in Ceylon. This is more probable as he omits all details and satisfies himself with a bare reference which shows that the stories were very famous and already known to his readers. The story22 of the merchant becoming a frog because of his falling away from the right path to which he was first introduced and his consequent liberation is told in the Visuddhi-Magga without the previous life of the merchant, while the cause of his death is different in the two versions. In the

Page #238

A Few Parallels in Jain and Buddhist Works
229
Jain version23 the frog is trampled down by the hoof of Śrenika's horse, while in the Buddhist story a cowherd kills the frog with a stick. The second parallel24 is not so marked as Buddhagosa only gives a passing remark about the central idea of the story without adding details. But the idea and the statement is so queer and out of the way that we are forced to think that there must be a story behind it. The Jain work2 relates the story in full. It tells us that a merchant was pursuing a thief who had carried away his daughter. But before he was able to catch hold of him the thief killed the girl and escaped. Now the father and his sons who were pursuing him found themselves in a thick forest without food. So to save themselves they ate the flesh of the girl and in this manner came safe to their own town. Buddhagosa makes the father to eat the flesh of the son instead of the daughter.
Besides these parallels in story and fable, we find similarities in the two canons about some important philosophical discussions. The most important among them is dialogue in the Rayapaseniyam 26 and the Payāsi Suttanta? in the Digha-Nikāya. Similarities in wording and similes and expressions leave no doubt as regards their mutual relation. Either both must have followed very closely a common source or one must have made a considerable use of the other. The Jain version as found in the second Upānga forms the central theme of the work. It turns on the point of the existence of the soul independent of the body in which it is embodied. Kesi the follower of Lord Pārsva tries to prove the soul as existing and refutes the arguments of king Paēsi who is a follower of the heretical teacher Ajita Kesakambali. The Pāli version makes the king bear the name Pāyāsi who holds conversation with Kumāra Kassapa who is also shown triumphant in refuting the arguments of the king. Some scholars are inclined to think that the Jain version is the later of the two, but without sufficient reasons28 On the contrary there are a few facts which point unmistakably to the conclusion that the Jain version is the older of the two. The vehement denial of the soul which is the main function of this story is a little inconsistent with the general spirit of Buddhism. Herein Buddhism agrees more with the views of the king than his opponent who is shown as successful in both the versions. According to the Buddhist tradition itself the present Sutta is not of equal age with the others of the same collection. It is even admitted that the real name of the king was Paēsi and not Pāyāsi which is an unmistakable sign of the authentic nature of the Jain tradition as against the Buddhist version. On the other hand the Jain version is shown to be contemporaneous with Mahāvīra as Kesi the disciple of Pārsva is shown in other works29 to hold conversation

Page #239

230
Amrita
with Gotama the chief disciple of Lord Mahāvīra. So here also we will have to admit that the Jain version of the dialogue is of better authenticity and we can go so far as to assert that it was the Jains who first tried to refute this doctrine of Ajita as being the exact contradictory position of their own theory of the existence of the soul. The Buddhist took this refutation bodily from them even though a little inconsistent with their own vehement denial of a. soul.
Another philosophical discussion common to the two literatures is that about the refutation of the philosophy of Gosāla. In the Jain works we find it stated in the Upāsagadasão 30 and the Bhagavatī31 while in the Samaññaphalasutta 32 of the Digha-Nikāya is found a summary of his views. The life story of Gosāla as found in the Bhagavati Sūtra is not to be met with in the Buddhist works. But the statement of his doctrines is common to both of them in a very similar phraseology. It is just possible that both of them were copying from the works of that sect. But the other alternative is more probable. It can very easily be seen that Gosāla was more intimately connected with Jainism than the Buddhism, even though it is very difficult to decide the exact relation in which he stood. But in view of the fact that his doctrines are taken by him from the Pūrvas and his claim to be the last prophet of Jainism it appears that he represented another line of the school of Pārsva while Mahāvīra succeeded in asserting himself as the true continuer of the orthodox line. So it is more probable that the Jain version had better chances of giving the views of the school of Gosāla in a more authentic form than its rival religion Buddhism.
Apart from these similarities extending over a long incident we have a good many verses33 in common in the works of these two religions. Particularly the Uttarādhyayana and the Sūtrakrtānga show many verses in common with the various books of the Pāli canon like the Dhammapada the Thera and Therī-Gāthās the Suttanipāta and stray verses from the Nikāyas. Dr. Winternitz34 has suggested the solution of such similar verses in the supposition that there existed before both these canons a floating mass of poetry dealing with ascetic life and ideals which was incorporated in the works of both these religions.
Annotations : 1. Keith, Buddhist Philosophy in India and Ceylon. Ch. i. pp. 15-24. 2. Cp. Rhys Davids. Buddhism its History and Literature pp. 187-195. Winternitz,

Page #240

A Few Parallels in Jain and Buddhist Works
231
His. Ind. Lit. Vol. II, pp. 4-5; Kern. Manual of Indian Buddhism, p. 103ff. 3. Charpentier. Introduction to his edition of the Uttarādhyayana, pp. 15-16. 4. Bruchstucke Buddhistischen Dramen, Einleitung. 5. Ch. XIII, pp. 115-119. 6. No. 498, Fausball, Vol. VII, pp. 390-400. 7. Ch. XIII, Vs. 5-7. 8. Ch. XIV, pp. 119-125. 9. No. 509, pp. 473-490, Vol. IV. [Annals, B. O. R. I.) 10. Ch. XII, pp. 109-115. 11. No. 497, Vol. IV, pp. 375-389. 12. Cp. the commentaries of śāntyācārya and Devendra. 13. Ch. 4th. 14. Cp. Oldenberg, Buddha, p. 313. 15. Jñātādharmakathā, Ch. 9th. 16. No. 196, Cp. Winternitz, op. cit. p. 131. 17. Jñāt., Ch. 16th. 18. Ed. by Müller, Verses 400-447, pp. 260-271. 19. In the introduction of her Psalms of the Sisters, P. T. S. 20. V. 431. 21. V. 427. 22. Ed. Mrs. Rhys Davids, P. T. S. Vol. I, p. 208. 23. Jñāt. Ch. 13th. 24. Visuddhi-Magga, Vol. I, p. 347. 25. Jñāt, Ch. 18th. 26. Ed. by Dr. Vaidya, Poona, 1934. 27. Ed. P. T. S. No. 23, Vol. II, pp. 316-358. 28. Cp. B. C. Law. A History of Pāli Literature, 1933. Vol. I, p. 109. 29. Cp. Uttar. Ch. 23rd. 30. Ed. by Dr. Vaidya, Poona, 1930. Chs. 6 and 7. 31. Uvās pp. 139-192. 32. Ed. P. T. S. Vol. I, pp. 47-86.

Page #241

232
Amrita
33. I give below the common verses from Uttarādhyayana and Sūtrakrtänga as fas as
I am able to detect. Utta. I, 17. Thorī-Gāthā. 247; II. 3. Therga-Gā. 243; II, 2425. Suttanipäta. 932; III, 17. Sn. 769; IV. I. Dhp. 182; IV, 3. Thera-Gā. 786; V. 21. Dhp. 141; VIII, 13. Sn. 927: IX, 34. Dhp. 103; 44. Dhp. 70; 48-49. MāraSany. 11, 10.6; X, 28. Dhp. 285 : XXV, 16. Sn. 268; 17. Sn. 136; 31. Dhp. 264; XXVII, 8. Thera-Gä. 976. Sütrak. I, 1, 1, 3. Sn. 394; I, 2, 1, 2. Sn. 578; I, 2, 1, 15. Samyutta-Nikāya. IX, 1; 1, 2, 2, 11. Thera-Gã. 1053; I, 2, 2, 15. Dhp. 373; 1, 2, 2, 17. Sn. 810; I. 3, 2, 21. Thera-Gä. 1154; I, 3, 4, 7. Theri-Gä. 508; I, 3, 4, 8. Dhp. 245; I, 5, 0, Dhp. 307; I, 7, 15-16. Therī-Gā. 241-244; I, 7, 25.
Dhp. 325; I. 8, 7, Dhp. 5; I, 8, 19. Sn. 400; etc. 34. Hist. Ind. Lit. Vol. II. P. 121, p. 125 etc.
OOO
A Few Parallels in Jain and Buddhist Works ABORI.XVII. 1936 35

Page #242

Parallel Passages in the Daśavaikälika and the Ācārānga
The chronology of the different books of the Ardha-Mägadhi canon is as yet unsettled except in a general way, by which some books like the Acaränga, Sütrakṛtānga are assigned to the oldest stratum while others like the metrical Painnas are said to form the latest additions. Other books are put in different places between these two extreme limits. When we come to decide more exactly the place of a particular book and its relative chronology with reference to some other work of the canon, we find that such general considerations are 'of no great use and the relative chronology of any two books of the canon or even parts of books must be decided by a close comparison of these works with each other, with a view to find out which is older and which is younger.
There are three chapters in the Daśavaikälika which have a very close parallel in three chapters of the second part of the Acaränga, from which an attempt can be made to decide the relative chronology of these two books of the canon. These parallel passages were already noted by W. Schubring and Patwardhan who have also concluded about their chronology. Schubring has expressed his opinion that the Daśavaikälika is younger of the two and naturally draws upon the passages of the Acäränga in writing its own chapters. Prof. Patwardhan has also come to the same conclusion and expresses it with greater conviction. Before, however, these conclusions can be accepted, it is necessary to go through their arguments and also to compare both the texts more closely.
The view that the chapters of the Daśavaikälika are younger and draw for their sources on the Acaränga appears to be based upon considerations of a general nature, that the latter work is older and that the Daśavaikälika is accepted by all to be a work of a compilatory character. Here it must be pointed out that both the parts of the Acäränga cannot claim the same.

Page #243

234
Amrita
antiquity and the second part from which all the three parallels are to be derived is considerably younger than the first one. Another reason for this supposition must have been the general impression that the language of prose portions of the canon shows an older form than the one used in the metrical portions and because the passages in the Acārānga are in prose and those of the Daśavaikālika are in verse, the verses must be younger. Prof. Patwardhan further thinks that this position is strengthened by the fact that Das. VIII. 49 makes a reference to Āyāra., thus showing its acquaintance with it.
All these considerations of more or less general nature are not sufficient to prove the relative chronology of these few parallel passages. The Ācārānga is one of the oldest books of the canon may be accepted in a general sense, but the text itself is found in two distinct parts of different ages, and while the first part can be said to be fairly old, the second is younger. As all the parallel passages occur in the second part only the high antiquity of the first part is not sufficient unless it is proved that the second Srutakandha is also equally old or older than the Das. The other fact that the Das. is admittedly a compilation made by Sejjambhava for the benefit of his son is not sufficient in calling them as versification of the prose of the Ācārānga. The reference to Ācāra. in the Das. verse is at best doubtful. We are thrown back upon an actual comparison of all the parallel passages to decide the chronology, and no argument of a general nature would settle the case one way or the other.
I give below the two texts in parallel columns where some kind of verbal agreement is to be found. Generally the topics of all these chapters are the same and the contents of both the books are nearly the same.
ACARANGA
1. II. 1.1.1. se jjam puna jānejjā : asanam vā pānam vā khāimam vā sāimam vā pāņehim vā panaehim vā biehim vā hariehim vā saṁsattam ummissam sīodaeņa vā osittam rayasā vā parighäsiyam tahāppagāram asanam vā ....
2. II. 1.1.2. se tam āyāe egamtam avakkamejjā ahe jhamathamdilaṁsi vā....tao samjayām eva parittavejjā /
anabhikkantabhajjiyam
pehāe... no
3. II. 1.1.3. taruniyam vä сhivādim padigahejjā.
4. II. 1.3.9. tivvadesiyam vā vāsam vāsamānam pehãe, tivvadesiyam vā

Page #244

Parallel Passages in the Daśavaikālika and the Acāränga
235
mahiyam samnivayamāņam pehāe mahāvāena rayaṁ samubbhūtam pehãe tiricchapätimă vă pāņā samthadā samnivayamānā pehāe....
5. II. 1.5.2. se tuttha parakkamamāne payalejja vā pavadejja va, se tattha payalamane vā pavadamane vā.....
6. II. 1.5.6. se jjaṁ puņa jānejjā samanaṁ vā māhanam vā gāmapindolagam vā atihiṁ vā puvva-pavitthaṁ pehāe no te uvāikkamma pavisejja vā obhāsejja vā / se tam āyāe egamtam avakkamejjā aṇāvāyam asamloe citthejja / aha puna evam jānejjā padisenie vā dinne vā tao tańmi niyatție, tao samjayām eva pavisejja vā obhāsejja vā /
7. II. 1.6.2. no gahavaikulassa dagacchaddaṇamattae citthejjā... no gāhāvaikulassa siņānassa vā vaccassa vā samloe sapadiduvāre citthejjā, no gāhāvaikulassa aloyaṁ vă thiggalaṁ vā samdhim vā dagabhavanam va....nijjhāejjā /
8. II. 1.6.4. tahappagāreņa purekammakaena hattheņa vā (matteņa vā davvie vā bhāyanena va) aphāsuyaṁ anesanijjam jāva no padigahejjä /
9. II. 1.6.5-6. no purekammaena udaulleņa tahappagārena udaulleņa hatthena vă 4 asanam vā 4 aphāsuyaṁ anesaņijjam jāva no padigāhejjā / no udaullena sasiniddhena sesam tam ceva/ evam sasarakkhe udaulle sasiņiddhe mattiyā ose, hariyāle himgulae manosilā amjane lone geruya-vanniyasedi - ya - soratthiya - pitthakukkusa - kae- ya - ukkattha - saṁsatthena /
10. II. 1.6.10. se jjaṁ puņa jānejjā asanam vā (pāņagam và khāimam vā säimam vă aganinikkhittam tahappagāram asaņam và 4 aphāsuyam jāva no padigahejja / assamjae bhikkhūpadiyāe osimcamane vā nisimcamāne vā amajjaman pamajjamāne vā āyāremāne vā uyattemāne vā aganijīve himsejja
11. II. 1.7.1. assamjae bhikkhūpadiyāe pīdhaṁ vā phalagaṁ vā nisseņiṁ vä udūhalaṁ vā āhattu ussaviyā duruhejjā se tattha duruhamāne payalejja vā pavadejja va, se tattha payalamāne pavadamāṇe hattham vā pāyam vā .... lūsejjā pānāni vā abhihanijja ...taṁ tahappagäram mālohadaṁ asanaṁ vā 4 jāva no padigähejja.
12. II. 1.7.7. aha puna evam jānejjā cirädhoyam, ambilaṁ vokkamtam parinatam viddhattham phāsuyam jāva padigahejjā /
13. II. 1.8.3. se jjam puna jānejjā sāluyam vā viraliyam vā sāsavnāliyam vā
14. II. 1.10.2. se egaio maņunnaṁ bhoyanajāyam padigāhettä paṁteņa bhoyaņeņa padicchãei : mă metam dāiyam samtam datthüņaṁ sayamāie.
15. II. 1.10.3. se egaio annataram bhoyanajāyam padigāhettä bhaddayam (bhaddayam) bhoccă vivannan virasaṁ āharati /
16. II. 1.10.4. assim
khalu padigähiyamsi appe siyā bhoyanajäe

Page #245

236
Amrita
bahuujjhiyadhammie, tahappagaram no padigahejjä /
17. II. 1.10.5. se jjam puna jānejjā bahuyatthiyam vă maṁsam maccham vā bahukamtagam, assim khalu padigahiyamsi appe siyā bhojanajāe bahuujjhiyadhammie, tahappagāram.
18. II. 1.11.1. ime pinde ime loe ime tittae ime kaduyae ime kasāe ime ambile ime mahure
19. II. 4.1.13. pannavam se bhikkhū vā 2 amtalikkhe ti vā gujjhānucarie ti vā sammucchie ti va nivaie vā paoe vadejjā vā vutthavalāhage tti/
20. II. 4.2.7. se bhikkhū vā 2 manussam vā gonam vā mahisam vā migam vā pasum vā pakkhiṁ vā sirisivam vā jalayaram vă se ttaṁ parivūdhakāyaṁ pehāe no evan vadejjā thulle ti vā pameile ti vā vatte ti vā vajjhe ti vā pāime ti vā
21. II. 4.2.9. se bhikkhu vā 2 virūvarūvão gão pehāe no evaṁ vadejjā, tam jaha : dojjhã ti vā dammā ti vā gorahā ti vā vahimā ti vā rahajogga ti vä/
22. II. 4.2.10. se bhikkhu vā 2. virūvarūvão gão pehāe evan vadejjā, tam jaha juvam gave iti vă dhenú ti vā rasavai ti va hasse ti va mahalle ti vă mahavvae ti vā samucchane ti vā /
23. II. 4.2.11. se bhikkhủ vă 2 taheva gastuṁ ujjāņam pavvayāni vānāņi ya rukkhā mahallā pehāe no evam vadejjā pāsāyajoggă ti vā toranajogga ti vā gihajoggā ti va phalihajoggā ti vā aggalajoggá ti vā nāvājoggā ti vā udagajoggā ti vā doni-pīdhacamgaveranangala- kuliya - jamta-latthi-nābhi-gamdiasana-sayana-jana-uvassayājogga ti vā/
24. II. 4.2.12. se bhikkhū va 2 tahe vā gamtum ujjāņāim pavvayāņi vaņāņi vā rukkhā mahallä pehāe evaṁ vadejjā : tam jahā : jātimartā ti vă dihavattā ti vā mahālayā ti vā payātasālā ti vā pāsādiyā ti vā /
25. II. 4.2.13. se bhikkhu vā 2 bahusambhūtā vanaphalā pehãe no evam vadejjā, tam jahā : pakkā ti vā pātakhajjā ti vā velociyā ti vā tālā ti vā pehā ti vā
26. II. 4.2.14. se bhikkhu vã 2 bahusambhutã vinaphala pehe evan vadeja, tam jahā : asanthadā ti vā bahunivvattimaphalā ti vā bahusambhūya vā bhūtarūvā ti vā evappagāram bhāsam asāvajjam jāva bhāsejjä /
27. II. 4.2.15. se bhikkhu vā 2 bahusambhūyao osahio pehāe tahā vi tão no evam vadejjā, tam jahā : pakkā ti vă niliyā ti vã chaviti vā laimā ti vā bhajjimāti va bahukhajjimā ti vā /
28. II. 4.2.16. se bhikkhū vā 2 bahusaņbhūyão osahio pehāe tahā vi tão evam vadejjā, tam jahā : rūdhā ti vā bahusambhūtā ti vā thirā ti vā usadhā ti vā gabbhiyā

Page #246

Parallel Passages in the Daśavaikälika and the Ācāränga
237
ti vā -pasūtā ti vā sasārā ti vā eyappagāram bhāsam asāvajjam jāva bhāsejjā /
29. II. 15.29. tao naṁ samane bhagavam Mahāvīre uppannaņānadamsanadhare Goyamadiņam samaņānam niggasthānam pamca mahavvayāiim sabhāvaņāiim chajjivanikāyāim äikkhai bhāsai parūvei, tam jaha : pudhavikāe jāva tasakāe //
30. II. 15. padhamam bhante mahavvayam paccakhāmi savvam pāņaivāyam, se suhumam vā bāyaram vā tasaṁ và thāvaraṁ vā neva sayaṁ pāņivāyam karejja 3 jāvajjīvāe tiviham tivihenam manasā vayasă kāyasā tassa bhamte padikkamāmi nimdāmi garahāmi appānaṁ vosirāmi.
.
DAŠAVAIKĀLIKA
V. 1.57. asanam panagam vā vi khāimam sāimam tahā / pupphesu hojja ummissam biesu hariesu vä // V. 1.59. asanam pāņagam vā vi khāimam säimam tahā / udagamsi hojja nikkhittam uttinga panagesu vā //
V. 1.72. vikkāyamāņmam pasadham raena pariphāsiyam / V. 1.85-86. hatthena taṁ gaheūnaṁ egaṁtam avakkame //
egamtam avakkamittá acittam padilehiyā jayam parithavejja parithappa padikkame //
V. 2.20. taruniyam vã chivadim ămiyam bhajiyam saim / dintiyam padiyāikkhe na me kappai tārisam // V. 1.8. na carejja vāse vāsante mahiyāe vā padantie / mahāvāe ya vāyante tiricchasampāimesu vā // 'V. 1.5. pavadante vā se tattha pakkhalante vä samjae/ V. 2.10-11. samanam māhanam vā vi kiviņas vā vaņīmagam / uvasaṁkamantam bhattatthā pānatthae vä samjae // tam aikkamittu na pavise na citthe cakkhugoyare / egaṁta avakkamittā tattha citthejja samjae // V. 2.13. padisehie vā dinne vă tao taṁmi nivattie / V. 1.25. siņānassa ya vaccassa saṁlogam parivajjae/ V. 1.15. āloyaṁ thiggalaṁ dāram samdhiṁ dagabhavanāņi ya / caranto na vinijjhāe saṁkatthānam vivajjae/ V. 1.32. purekammeņa hatthena davvīe bhāyaṇena vä/ V. 1.33-34. evaṁ -

Page #247

238
Amrita
udaulle sasiniddhe sasarakkhe mattiyâ-üse / hariyāle himgulae manosilā amjane lone // geruya vanniya sediya soraţthiya pittha kukkusakaye ya // ukkatthamasansatthe amsatthe ceva bodhavve // V. 1.61. asaņam pāņagam vā vi khāimam säimaṁ taha / aganimsi hojja nikkhittam
V. 1.63. evam ussakkiyā osakkiyā ujjāliyā pajjāliyā nivvāviyā ussimciyā nissimciyā uvvattiyā oyāriyā dae/
V. 1.67-69.
nisseņiṁ phalagam pīdham ussavittāna māruhe / mascakilaṁ ca pāsāyam samaņattāe vā dāvae // durūhamāņi pavadejjā hattham yāyaṁ ca lūsae/ pudhavijīve vi hissejjā je ya tam nissiyā jaga // tamhā mālohadam bhikkham na padigenhanti samjayā / V. 1.76-77. jam jānejjā cirādhoyam maie damsanena vā / ajīvam parinayam naccă padigāhejja samjaye // V. 2.18. sāluyam vā virāliyam kumuyam uppalanāliyam / munāliyam sāsavanāliyam. V. 2.31. siyā egaio laddhum lobhena vinigūhai / mä meyam dāiyam samtam datthūnam sayamāyae // V. 2.33. siyā egaio laddhuṁ viviham pāṇabhoyaņam / bhaddagam bhaddagam bhoccā vivannam virasaṁ āhare // V. 1.74. appe siyā bhoyanajāe bahu ujjhiyadhammie / dintiyaṁ padiyāikkhe na me kappai tārisam // V. 1.73. bahuatthiyaṁ poggalam animisam vā bahukamtayam V. 1.97. tittagam vā kaduyam vā kasāyam
ambilaṁ vā mahuram lavanaṁ vā /
VII. 52-53.
taheva meham vā naham vā mānavam na deva deva tti giram vaejjā / sammucchie unnae yā paoe vaejja vā vuttha balāhaga tti //

Page #248

Parallel Passages in the Daśavaikālika and the Acäränga
239
astalikkhatti nam būyā gujjhānucariyatti ya/ VII. 22. taheva manusam pasum pakkhiṁ vā vi sirīsivam / thüle pameile vajjhe paime tti ya no vae // VII. 24. taheva gão dojjhão dammā gorahaga tti ya/ vāhimā rahajogga tti nevaṁ bhāsejja pannavam // VII. 25. juvaṁ gave tti ņam būyā dheņuṁ rasadaya tti ya / rahasse mahallae vā vi vae saṁvahane tri ya //
VII. 26-29.
taheva gamtum ujjānam pavvayāni vanāņi ya / rukkhā mahalla pehāe nevaṁ bhāsejja pannavam //
alaṁ pāsāyakharbānam toranāņa gihāņa ya / phalihaggalanāvānam alam udagadoninam //
pidhae caṁgabere ya namgale maimam siya / jamtalatthi ya nābhī vā gamdiyā vā alam siyā //
āsaņam sayanam jāņam hojjā vā kimcuvassae / VII. 30-31. 'taheva gaṁtum ujjānam pavvayāni vanāni ya / rukkha mahalla pehāe evam bhāsejja pannavam //
jāimastă ime rukkhā dīhavattā mahālayā / payāyasālā viļimā vae darisaại tti ya // VII. 32. tahā phalaim pakkāim pāyakhajjāim no vae / veloiyāim tālāim vehimāim ti no vae / VII. 33. asamthadā ime ambā bahunivvattimā phalā / vaejja bahusambhūya bhūyarūva tti vā puno // VII. 34. tahosahữo pakkāo niliyão chavi i ya / läimā bhajjimão tti pihukhajja tti no vae // V. 35. rūdhã bahusambhūyā thirā ūsadhā vi ya/ gabbhiyão pasūyão sasārão tti älave //
IV. ima khalu să chajjīvaniyā nāmajjhayanam samaņeņam bhagavayā Mahāvīrenam kāsaveņam paveiyā...tam jahā : pudhavikāiyā...tasakāiyā.
IV. padhame bhaṁte mahavvae pāņāivāyāo veramaņam / savvam bhamte pānaivāyam paccakkhāmi, se suhumaṁ vā bāyaraṁ vā tasaṁ vā thāvaram vā/ neva sayaṁ pane aivāejjā nevannehi pāņe aivāyāvejjā pāņe aivāyante vi anne na samaņujānāmi, tiviham tivihenam maņeņa vāyāe kāenam na karemi...tassa bhamte

Page #249

240
Amrita
padikkamāmi niņdāmi garihāmi appānam vosirāmi /
An exact parallel of this very nature is to be found in the remaining portions dealing with the other four vows, which need not be repeated here, as the first comparison is sufficient for all the purposes.
Before we note down the result of this comparison it is to be pointed out that though apparently the text of the Acārānga is in prose, there are indications in the text itself which would go to show that in its original form it was in metrical form or at least contained many metrical lines. In fact what is proved by SCHUBRING for the first śrutakandha of the book is true of the second as well, though to a less extent. We can note the following lines in the part chosen for comparison : samie sahite sayā jae Ay. II. 1.1. 14; II. 1.2.7; padisenie vă dinne vā tao tammi niyattie Ay. II. 1.5.6; sasarakkhe udaulle sasiņiddhe mattiyā ose 7 hariyāle himgulae manosilā amjane lone / geruya-vanniya-sediya-soratthiya-pitthakkusakae ya / Ay. II. 1.6.6.; sāluyam vă viräliyam Ay. II. 1.8.3; mă metam dāiyam santam datthunam sayaOé äie / Ay. II. 1.10.2; bhaddayam bhaddayam bhoccă vivannam virasam āhare / Ay. II. 1.10.3.; appe siyā bhoyanajāe bahuujjhiyadhammie / Ay. II. 1.10.4; bahuyatthiyam vă maṁsam macchaṁ vā bahukantagam / Ay. II. 1..10.5; vedejja vă vutthabalāhage tti / Ay. II. 4.1.13; taheva gamtum ujjänam pavvayani vanāni vā / rukkhā mahallā pehāe / Ay. II. 4.2.11. .
This list can be greatly extended if we make slight changes usually of pping the disjunctive particle vā or ti vā which would turn many of these prose passages into perfect metrical lines. But what is more important for our present purpose is the fact that all these metrical lines also occur in the Daśavaikälika in exactly the same form, where they are preserved in their original metrical garb. From the comparison itself we can make out the following points :
1. In most of the parallel passages the order of the words to be enumerated is the same. Cp. Nos. 4, 7, 9, 14, 15, 18, 20, 21, 22, 23, 27 and nearly all the remaining passages with slight changes in them. Now it is impossible to believe that exactly the same order of the words can be kept throughout if we regard the Daśavaikālika passages to be a versification of the prose of the Ācārānga. It is more natural to suppose that the verses were re-written in prose where the order can be easily preserved.
2. It is equally striking that to read the prose passages as verses we are required to drop merely the particles like vā or ti vā which have no sense of their own, and again it appears more probable to suppose that they were added by the writer who turned the original verses into prose, than the other way.

Page #250

Parallel Passages in the Daśavaikālika and the Acāränga
241
3. We have further individual cases by examining which one can clearly see the process of turning the verses into prose. The mechanical method of inserting the disjunctive particle after every word has led the prose writer to produce passages which have no meaning. Thus we find the compound udagadoni of the verse written as udagajogga tti va doni; the sentence damma gorahagā into dammā ti vā gorahagā ti vā; niliyão chavi into niliyā ti vā chavi ti vā—all these prose passages giving no consistent meaning.
4. It is impossible to suppose that the same order to such an extent can be preserved in changing the prose into verses, particularly in such a case as No. 18 where the words in the same order have given two lines of such a refined metre as Svāgatā.
5. In the prose passages we find the order of the words which are otherwise in prose somewhat peculiar and not the normal one. Cp. vadejjā vā vutthabalāhage tti; taheva gamtum ujjānam pavvayāņi vaņāņi vā; aha puna evaņ jāņejjā cirādhoyaṁ etc.
6. In one case we find that the verse uses the form uvassae as Loc. sing. which the writer of the prose has understood as the Nom. and has added the word in the list of things in 'No. 23.
All these considerations go to show that out of the two parallel texts the one found in the Daśavaikālika is the older and is preserved in the original form while the prose of the Acāränga is younger and is a mutilation of the original verses.
DOO
Parallel Passages in the Daśavaikālika and the Ācārānga
NIA Vol. 1, No. 2, May 1938

Page #251

Kundakunda and his Philosophy
In spite of the rich mine of thought found in the history of Indian philosophy, there are very few outstanding personalities which contributed to its development. This lack of philosophers, in spite of the great number of philosophical schools and systems, is no doubt the result of the Indian mind to concentrate on the thought itself more than on the thinker, and to care more for the books than their writers. The lack of historical material and the consequent vagueness about the personal lives of a few great names that have survived, is also responsible for this to a certain extent. But the fact remains that in comparison with the histories of thought of other nations, both ancient and modern, India presents a curious spectacle in which we find philosophy mainly developed in schools and systems, and not in a rapid succession of great personalities and their individual opinions.
But even when the general condition of the Indian philosophy is of this nature, we do have a few great names that can stand comparison with any philosopher of the western nations. Such a person is Sankara whose personality has influenced an existing system of philosophy to such an extent as to give it a distinct personal colour. Such were also the two great outstanding personalities. Mahāvīra and Buddha whose personal opinions can be gathered to a certain extent, even though they are now found mingled with later developments. And it is certainly a very interesting and also an instructive study to collect and arrange material to illuminate such other personalities that lie buried in the history of Indian philosophy.
To such an end the difficulties are obvious. It has become a common place of observation to say that India gives us nothing as regards the personal and private life of her heroes, and this is a great draw-back in our study of the individual philosophers. Similarly the Indian thinkers themselves cared more for the net result of their philosophical thought and less for the method by which the result was obtained. Thus a thinker, when he finds some of his

Page #252

243
Kundakunda and his Philosophy
principles in agreement with any other system, does not hesitate to identify them with those of that system, and thinks them to be the same, even though set in an altogether different back-ground. The practical effect of such a procedure is to minimise the individual differences. But, above all, the method of trying to put whatever new ideas one has, on the ancient authors, and of interpreting and re-interpreting their works to conform to later developments is the chief cause of reducing the individual contribution to a minimum. Moreover the very spirit of Indian thinkers was to obliterate all traces of personal facts from their writings and to view the subject from at purely objective point of view.
Particularly in the case of Jaina philosophy we meet these difficulties to a greater extent than probably in others. Herein there is very little development and the system has not much evolved to newer and newer forms. As Jacobi has pointed out long ago, Jainism shows the greatest amount of orthodoxy and adheres to its primitive principles with a remarkable tenacity of its own. Further, the fact that the main philosophical principles of the system are few and simple, makes any great change impossible; while all divergences between author and author pertain to minor questions of division and sub-division. The tendency to try to trace all the facts in the established canon, also has a considerable effect in eliminating the individual opinions and contributions. But, in spite of this, a few great names stand out clearly before us to demand our attention, like kundakunda, Umäsvāti and Haribhadra. In this place, we take for consideration the first of the three great writers on Jaina philosophy.
Of the personal life of Kundakunda we know very little. The few facts which are handed down to us by late traditions and writers of religious stories are unimportant and reveal their legendary character very clearly. In a late work called Punyasravakatha we have the life of Kundakunda, in which he is said to be a cow-boy in his former life, and is described as finding a copy of the Jaina scriptures in a forest, miraculously saved from being burnt. by the forest conflagration that had raged all around it. He took the sacred books and presented them as a gift to a Sadhu who had come for food at his master's house. As the result of this good deed he was born as the son of the wealthy merchant in whose house he formely served; and soon obtained proficiency in the Jaina scriptures. Another story relates that he cured a lay woman who was possessed by a demon, with his supernatural powers. But such stories, however interesting to read and instructive in exposing the psychology of the devout writers who wrote them down, are of

Page #253

244
Amrita
little avail for historical purpose; and often they will have to be summarily rejected for obvious reasons.
Besides these stories we have some more reliable information about Kundakunda. He appears to have borne a number of names, and a late record attributes to him as many as five different names, viz., Kundakunda, Padmanandi, Gridhrapiccha, Vakragriva Elacharya. But from other sources we know that some of these names were foisted on him on account of some confusion with later writers. In a popular colophon to the Tatvārthādhigamasūtra of Umāsvāti, we find the author to bear the name of Gridhrapiccha, and one of the traditions makes him a pupil of Kundakunda. So it may be that in later times the name of the pupil might have been transferred to the famous teacher. Similarly the other two names of Elacharya and Vakragriva were those of two later Acharyas, who, being less known, were possibly merged in the personality of this famous philosopher. This will explain, to some extent, the long list of books attributed by tradition to him, which probably contains the names of the works of these later writers. But nothing definite can be put forth, as these works are found only in their names. The real name of our author appears to be Padmanandin, while Kundakunda the better known of the two, is said by tradition, to be received by him, from his native village called Kondakunda.
It can now be taken as proved that Kundakunda was born in South India, even though one late tradition makes him a resident of a town, Varapura, in the country of Malwa. The exact place of his birth is variously given as Kondakunda or Hemagrama, the latter of which is identified with Ponnur. His father is said to be Kundashresti and his mother Kundalata; the names clearly showing their imaginary origin. We know nothing of his childhood and later life. He must have taken early to religious studies and soon obtained proficiency in them. His teacher is given as Jinachandra by the Pattavali of the Nandhisangha; while Jayasena, his commentator gives. kumaranandi as his teacher's name. According to another tradition Kundakunda himself was the teacher of Umäsväti, the well-known Jaina philosopher and theologion, who wrote the first authoritative book on Jaina philosophy in Sanskrit. According to a tradition preserved by Jayasena, and following him noted by Balachandra, and Brahmädeva, that in his later life, Kundakunda appears to have been associated with a king called Siva Kumāra" Mahārāja, for whose instruction he is said to have composed some of his works.

Page #254

Kundakunda and his Philosophy
It is further ascertained that Kundakunda belonged to the Mula or Dravida Sangha of the Jaina community and was possibly a Digambara, even though his traditional date goes back to the actual split of the Jaina Community into Svetambaras and Digambaras. According to the Pattavalis, which preserve the names of the teachers who succeeded Mahavira as the head of the Jaina Community, Kundakunda was born at about 52 B. C., took the pontiff's chair in 8 B. C., and lived a long life of 95 years. He is represented as succeeding Jinachandra I who is also taken to be his teacher. But this date is called in question by Dr. Pathak, who relies on the tradition preserved by Jayasena, according to which Kundakunda was a contemporary of a king known as Sivakumara Maharaja. He is identified by him, with the early Kadamba King Sri Vijaya Sivamrigesa who ruled in the fifth century A. D. in the western part of Mysore. Prof. Chakravarti, however, tries to confirm. the traditional date by identifying this king with the Pallava prince Sri Sivaskhandhavarman who ruled at Kanchi and issued the well-known Mayidavolu inscription.
245
From the traditional list of books attributed to Kundakunda it appears that he was a voluminous writer. These works come up to more than 50 different Pahudas of which about a dozen are available to us, and there is no apparent reason why the genuineness of his authorship of these works should be questioned. Of his other works, their bare names suggest to us nothing of their contents and authorship, and it remains doubtful whether they really existed or not, and if existed who was their author. It is just possible that when Kundakunda was recognized as a great writer of ancient times, many more works were fathered on him, so as to increase his greatness.
Of the existing works the most important are no doubt, the three Pahudas called Samayasara, Pravachanasara and Panchastikayasamayasara; all dealing with the main principles of Jaina philosophy. The Samayasara contains about 445 gäthās and deals with topics such as, views of oneself, views of others, the relation of the subject and the object, bondage, liberation and knowledge all pure. The style is logical and full of dialectical arguments used to prove the thesis. The Pravachanasara is the most popular of his works, and gives in three main divisions the most important of the Jaina doctrines. The first chapter, called the Jñänädhikāra, deals with knowledge and the nature of omniscience; the second Jñeyādhikara, summarises the metaphysical system of Jainism, while the last called Chariträdhikara deals with the ethical code of the Jaina monks, and the way to liberation. The work is found in two different recensions slightly differing from each other, one being shorter than

Page #255

246
Amrita
the other. The shorter one is found commented on by Amrtchandra while jayasena gives the longer one. Less philosophical and more schematic is the third Pahuda, which professes to give in a summary for the information about the five Astikāyas so well-known in Jaina philosophy. The major portion of the work, however, is devoted to the description of the Jivāstikāya which occupies the most important place among them all. It contains 180 Gāthās. Of the lesser works, the Satpahuda is a collection of six small tracts dealing with Darśana, Sūtra, Charitra, Bodhi, Bhava and Moksa. His other works are the Rayanasāra giving information on a variety of topics, Barahanuvekkha, dealing with the twelve philosophical reflections, and the Niyamasara, which also covers nearly the whole of Jaina philosophy.
It is to be noted that all these works are written in a Prākrit language which is designated by Pischel as Jaina Sauraseni, and is known by that name to modern scholars. Philologically viewed, it appears to be a mixture of Ardha-Māgadhi and Sauraseni. It has the peculiar Sauraseni changes like the change of a to & and to 4 and Ardha-Māgadhi forms like digal and TE. It can well be called Jaina Sauraseni, if we mean by that the Sauraseni language used by the earlier Jaina writers and influenced by their canonical speech, as we understand the parallel expression Jaina Māhārāstrī applied to the language of the Post canonical Prākrit literature of the Svetambara writers. The fact that this language was used by the Jaina writers alone and that also in the South India makes it probable that it was only a literary language and not a vernacular of the country in which it was used.
In giving the philosophical views of Kundakunda, it will be convenient to state at the beginning that what is tried here is not an epitome of his great works. Such an undertaking will include nearly all the details of the Jaina philosophy, as the works of Kundakunda are a storehouse of information on nearly all the points of the Jaina philosophy. Nor will such a description differ from the general outline of Jaina philosophy as found in other works, or even from that of the Ardha-Māgadhi canon to a very great extent. What is to be traced here, is only the distinctive feature of the handling of Jainism at the hands of Kundakunda, and his personal contributions to it.
To begin with, it will be profitable to start with his method of philosophising which is of great service in knowing the personality of an author. The method of the Jaina canon and the books mainly dependent on it, is purely dogmatic. The doctrine is stated with much repetition and emphasis, and the explanation mostly consists in giving the divisions and

Page #256

Kundakunda and his Philosophy
247
subdivisions of the subject. Synonyms serve the function of definition, and further explanation is thought to be of no use. The point of discussion is stated in as many ways as possible, and the whole thing is repeated with slight verbal changes. But such is not the case with Kundakunda. He, no doubt, is bound by the doctrines of the religion to which he belongs, and to which he tenaciously adheres; yet his method is not so dogmatic nor is he so prone to details. Being well acquainted with the development of other philosophical systems of India and their logic, he shows due regard to the play of reason in proving his thesis, and he even tries to attempt refutations of the possible objections which may suggest themselves to him. His arguments are more in accord with experience and exhibit his power of logical acumen to a considerable extent.
Of the various Pramanas current in Indian Philosophy, it appears that Kundakunda lays greatest emphasis on the scriptures of his religion, as was the case with many other great philosophers belonging to other systems. According to him the knowledge of the Jaina scriptures is absolutely necessary not only to understand the principles of his religion, but also to achieve its practical goal. Accordingly these scriptures form the basis of his philosophical system. It should not however be taken to represent a complete disregard of experience which really forms and should form the basis of all reasoning. On a closer scrutiny it will be found that the scriptures are valued so much, not because of some mysterious sanctity attaching to them, even though they are regarded as sacred, but because they are words of an omniscient being, who is ex-hypothesis, an infallible authority. So, the ultimate appeal is to the experiences of such a being, which are faithfully recorded in the scriptures, and the validity of it must stand or fall according as the scriptures represent the experiences, correctly or not. Now to decide it the Jainas afford no other alternative but the pragmatic test of trying to realise it. No amount of a priori discussion will be of any use to decide it.
After understanding the position of Kundakunda as regards his authority, we naturally come to the central conception of his philosophy, his solution of the epistemological question, which determines to a great extent all other views of our author. In this connection, the discussion of knowledge at the opening of Pravachanasāra is of great importance and a valuable contribution to the Jaina theory of knowledge. The author takes a firm stand on the very basic conception of it, namely, that the soul itself is knowledge and not a thing to which knowledge can be added as some extraneous thing. In fact, knowledge, along with happiness constitutes the very essence of the

Page #257

248
Amrita
soul or jiva. Both are co-extensive, and the very idea of one without the other is self-contradictory.
The application of this principle decides many questions for him. He points out that when one knows something what happens is that, the soul turns into knowledge, and the knowledge itself is nothing but the objects. focussed about the soul. So, we can say, that, the knower, the knowledge and the objects of knowledge are one and the same thing from the point of view of synthesis. But this should not lead us to suppose that they are identical in the sense of being physically one and the same thing. The author is careful enough in pointing out the fact that the relation of the knowledge and the objects known is not one of physical contact and infusion into each other; but it is one in which both identity and difference are kept together. When one sees an object, his eye is said to see it, even though there is no physical contact between the two, and this relation applies equally well to all other cases of knowledge.
This knowledge is of two kinds, the one, Indriyajñäna, or knowledge obtained by the senses, which is limited, and proceeds by the steps of Avagraha, Iha, Apaya and Dharaṇā. From the strictly, metaphysical point of view, this cannot be called knowledge at all. Kundakunda says, the whole universe is one in the sense of being an integral unity, and as the senseperception can but know only a part of it, it can be said to know nothing. A full understanding of anything involves the knowledge of all other things, as all things in the world are inseparably connected. This sensuous knowledge is the outcome of the soul's contact with his physical body, and pertains to things which can be grasped by the senses, and is thus limited by the structure of the instruments it uses. This consideration leads our author to call it as Parokha or indirect, as against the more common usage of regarding it as Pratyaksa. But in glaring contrast with this sensuous knowledge stands the Atindriyajñana which is otherwise called Kevala. The characteristic feature of this knowledge is its power to view all things directly. Nothing is indirectly known by it. It views things without any stages and gradations, as does the ordinary knowledge. Another trait of this knowledge is that it transcends the limits of space and time and can know the past, present and future modifications of all the objects. It leads to no bondage, as it arises out of the. complete distinction of the impeding karmas, and necessarily leads to final emancipation. This alone is the true knowledge worth the name, for to know a thing perfectly is to know all things in the universe.

Page #258

Kundakunda and his Philosophy
249
From this dual character of knowledge is deduced the two Nayas or view-points, called the Vyavahāranaya and the Nischayanaya; the former being only a stepping stone to the latter. As Kundakunda describes it, the necessity of the Vyavahāra is felt in making the conclusions of the latter comprehensible to the ordinary people, as the use of a Mleccha language is necessary for addressing an aboriginal. This method of viewing things from the ordinary and the real view-point, is no doubt related with the Nayavāda of the canonical Jainism, but stands in a very different footing as regards its practical application. One such interesting application is given by Kundakunda in stating that the worship of the idols of the Jinas and the praise of their physical beauty is to be regarded as meritorious only from the Vyavahāra point of view. From the stand-point of Nischayanaya, none of them can be regarded as meritorious, because neither the idol nor the body of the Jina has anything to do with the real soul of the Jina himself to which must our praise be directed.
The ontological system of Jainism is clearly realistic. The very first dogma of its metaphysics is the assertion of the two independent and eternal principles of soul and matter which together comprise the whole world. They are substances and one does not depend on the other for its existence. To these are added the two other entities of common experience time and space which are regarded as real existing entities and not merely subjective experiences dependent on mind for their existence. In addition to all of them, are the two peculiar conceptions of Dharma and Adharma which, of course, play a very subordinate part in the system. Kundakunda treats of them all in his Panchāstikāya and points out their individual characteristics and differs, there in, very little from the traditional description. But the further dogmas of Jainism fare differently at his hands. According to them, when the soul and matter come in contact, there follows their intermixture resulting into this worldly bondage reducing the soul to its present abject condition. The position of Kundakunda is, here delicately stated and maintained with the constant use of the two view-points; and one can clearly see the influence of idealistic systems working in them, even though he maintains the realistic interpretation as far as possible. He finds difficulty in knowing how the soul can be the cause of the various Karmas that bind him and change its nature, as the two principles are like poles asunder, and have nothing in common. He repeatedly says that the soul is not the cause of the karmic changes, as he cannot effect anything outside his own nature. It is incomprehensible how the soul can make, changes in the outside world; any kind of interaction is

Page #259

250
Amrita
impossible. But Kundakunda is equally unwilling to go to the length of saying that the soul is totally inactive and does not do anything, which opinion reduces the bondage to a mere illusory phenomenon. In fact he combats with great force this position of the Sankhya philosophers. The solution of the problem arrived at by him, in the present case, is from the vyavahāra point of view the soul can be said to make changes in the karmic matter, but from the Nischaya point of view he can be said to effect changes in his own Bhāvas which eventually result into the bondage of the karmas. This is called the Bhāvabandha in which the soul assumes a state worthy of the influx of the karmic substance, and the karma which is in constant motion, enters it and forms the Karmabandha. When once this principle of the two-fold nature of a process, one psychical and the other physical, is established, it was easy for him to apply the same to other processes like Nirjarā and Saṁvara. Its application, however, to the principle of Moksa leads to the interesting result of the existence of the Kevalins who are yet men of this world and who can be compared with the Jivanmuktas of the other systems.
Kundakunda's exposition of the Jivatatva is interesting as affording some insight into his psychological system. From the very beginning, he is careful in distinguishing the really psychical element of our being from the accompanying physiological and organic phenomena like the five senses, vitality, life and respiration, which he definitely classes as material. The real soul is characterised by Upayoga which further develops into knowledge and conation, and the knowledge of its real nature is the necessary requisite of obtaining final emancipation. Besides this usual division of Upayoga into knowledge and conation, Kundakunda gives at the beginning of his Pravachanasāra, another division based upon ethical considerations, into three kinds, Subha, Aśubha, and Suddha. The Suddhopayoga is the best of them all and consists in understanding the distinction between the soul and matter, and knowing the real nature of the former. It is obtained by the complete destruction of all the karmas, and it knows the soul to be constituted by: knowledge and happiness, and contemplates on its own nature. The subha and the Aśubha Upayogas, though distinguished as good and bad from the Vyavahāranaya, are both classed as bad from the Nischayanaya, which makes no distinction between the two.
This distinction of the various kinds of Upayogas naturally leads us to his ethical system, wherein, Kundakunda is less original and adhers to the traditional code of behaviour both for the monks and laymen. The rules governing the behaviour of the monks are strict and are so formulated as to

Page #260

Kundakunda and his Philosophy
251
produce the Suddhopayoga, ultimately leading to liberation. Herein also, he insists more on the purity of the soul than on the outward actions themselves, while the earlier Jainism appears to have laid equal stress on both. The most essential condition of a true monk is to irradicate the four passions which are the main roots of all misery and of this mundane existence. The ethical code of the pious layman is simpler and less strict, as was to be expected. They are to live a life free from all gross crimes and to help their co-religionists to their best ability and help to continue the faith. In fact there's is an ideal life for a man of this world.
OOO
Kundakunda and his Philosophy, Jain Gazette. XXX.10. 1933
37

Page #261

A Passage from Haribhadra's Samarādityakathā
Among the numerous works in Jain Māhārāstrī, the Samarādityakathā of Haribhadra undoubtedly occupies a prominent place. But in the absence of a close study of this masterpiece of a great poet and scholar, many of its beauties remain hidden from us. A number of passages from his work are in need of a closer interpretation and explanation. As a striking example I discuss here a small passage from the sixth Bhava of this work, which must be explained in a manner different from the attempts made so far and which reveals a fine sense of humour and resourcefulness on the part of Haribhadra.
Dharana, the hero of the story, who is married to Laksmī, his inveterate enemy of many previous births, passes through many adventures. Once while he is sleeping in a temple outside a town called Mahāsaras, along with his wife after an escape from danger, a thief enters the place to avoid being caught by the constables pursuing him for robbing the royal treasury. Dharana, due to the fatigue of the earlier day, is fast asleep and Laksmi alone becomes aware of the presence of the thief Candarudda. With a diabolic idea in her mind, she approaches him stealthily and wants to know from him his difficulty. Candarudda explains to her that he is being followed by the constables and he is hiding in the temple in order to escape them. Lakşmi promises to help him and explains her plan to abandon her real husband Dharana to the constables as the thief and eloping with Candarudda by claiming him as her husband in the presence of the king if need arises.'
To this suggestion Candarudda makes the following answer, at the same time refusing to follow her suggestion and giving reasons for it. His words are: (Candaruddena bhaniyam) / sundari, atthi eyam, kistu ahaṁ emha vatthavvao caucaranapadibaddho / ao viyāņāi me tam agahiyanāmam savvaloo ceva ettha mahiliyaṁ ti / The context makes it clear that the general sense of the passage is to point out the difficulties which Candarudda feels in claiming Lakşmi to be his wife in open court. For one thing, he is a resident of this place and naturally people know him rather well. Secondly he is

Page #262

A Passage from
Haribhadra's Samaradityakatha
253
already married and all the people of the town know his wife too well to believe that Laksmi could be that lady. He himself suggests the use of a magic pill to escape and elopes with Laksmī, as the story progresses.
Two words in this short passage caucaranapadibaddha and agahiyanāmam... mahiliyam deserve special consideration. The Sanskrit chāyā and the annotations given so far have all failed to point out the real significance of these expressions which Haribhadra has chosen with a purpose. Caucaranapadibaddha is vaguely explained as bhāryāyukta 'married' and agahiyanāmam is taken to mean that the people of the town know her but not her name. While the sense of the first word can suit the context and gives us the required sense the meaning of the second is both logically and contextually absurd. To know his wife but not her name can in no way add to the difficulties of claiming Laksmī to be his wife for the thief. It is simply irrelevant and Haribhadra, a careful writer, is not likely to add such an expression without some significance.
The passage will gain in meaning and value if we attempt a more precise sense of these two words used here. The robber Candarudda is speaking with a strain of sarcasm and alludes to his hopelessly unhappy married life and is using language with a sting in it. Both appear to be extempore formations based on popular expressions intended to give vent to bitterness. In popular Marāthi there is a famous expression caturbhuja honē in the sense of getting married and usage has humorously extended its use to the act of getting captured. No explanation is available for this use except the vague idea that the man will have four arms, by counting the arms of his wife. This will not however explain its extended sense, which implies that the man is imagined as having four arms, which in the later case may be due to the custom of tying ropes to his arms when led captive. Apart from this problem which must be left to the study of the Marāthī idiom, it is evident that Haribhadra had before him, this or a similar expression in the sense of 'to get married' which alone would explain the twist he has given to it to allude to the unhappy marriage of the thief, possibly with a shrew. In contrast with the expression caturbhujapratibaddha, if we may guess such a Sanskrit word, Haribhadra coins a parallel word catuhcaranapratibaddha meaning 'having the ill luck of getting married.' But there is, I think, a deeper suggestion in it. Caturbhuja, in Classical Sanskrit, is the most frequent epithet of Visnu and we know that Indian custom considered the bridegroom as a form of this great god at the time of marriage, as can be seen from the phrase vişnurūpine varāya. Though the words of the ritual are late, they reveal an essentially popular idea involved in marriage. Thus caturbhujapratibaddha may well have been intended to allude to this temporary

Page #263

254
Amrita
deification of the vara by identifying him with Visnu. Now Haribhadra has turned the tables as it were and coined the word catuhcaranapratibaddha, exploiting in full the other sense of catuḥcarana a synonym of catuspāda 'a beast'. The expression thus means, if rendered fully: 'I have been led in a wedlock, thereby degrading myself to the state of a beast'. The contrast between caturbhuja and catuhcarana gets its full force, alluding to the unhappy marriage of the speaker.
Still more striking is the sense of the other expression agrhitanamā (nāmni) mahilikā. What the thief wants to say is the fact that he cannot agree to claim Laksmī as his wife because his real wife is well known and is still alive. This would become clear if we recollect the original Sanskrit expression after which this word was newly coined by Haribhadra. This is undoubtedly the word sugrhītanaman. Long ago Lévi discussed the meaning of this word along with three others, occurring in the inscription of Rudradāman. in work's of dramaturgy, in Bāņa's works and in the lexicons. He came to the conclusion that the word sugrhītanāman is used in two distinct senses (i) as a mode of address used by the disciple, a son or a younger brother to designate a person to whom he owes respect and (ii) in a funeral sense as it were, when a deceased person is respectfully referred to. This second usage is frequent in the works of Bāna (cf. evam uparate'pi sugrhitanāmni tāte) and in the inscriptions (cf. inscriptions of Rudradāman and Mangaleśa). Lévi defines the sense of the word as 'to mention the name of a person, more especially a dead person accompanied with qualifications which bring good fortune, and which, thanks to their value as omens, may have a happy influence on the posthumous destiny of the deceased or on the future destiny of the living' (IA. xxii p. 167).
Here is again an extempore creation of a word by Haribhadra meant as a counterpart to sugrhitanāman, used in the sense of one whose name no one would like to mention for fear of misfortune and an implied suggestion that the person is yet alive. The robber, as the sequel tells us, has no objection to run away with Lakşmi but he cannot claim her as his wife in public for unfortunately his wife is alive and a source of ill luck to him. And people know her too well. He is thus sorry to tell her that he is married, that his wife is still alive and people know her well. Therefore he cannot accept her suggestion.
We here see Haribhadra using language with a suppleness and a fine sense of humour which is as suggestive as it is forceful. It expresses with extreme brevity the suppressed feelings of the robber about his unlucky marriage. "
A Passage from Haribhadra's Samaradityakathā, .
Indian Linguistics, 1955

Page #264

Pārsva's Historicity Reconsidered
Most of the historians of Jainism believe that both Pärśva and Mahāvīra, the last two of the 24 prophets of Jainism, were real historical persons. In the early days of Jain studies, scholars were less inclined to give credence to the Jain traditions and most of it was regarded as suspect. Nowa-days, the pendulum has swung to the other extreme and there is a growing tendency to put greater faith in the Jain traditional accounts and to regard them as historical without a close scrutiny of their reliability. In the interest of truth, it will be necessary to re-examine the question of Pārsva's historicity in the light of all the available evidence and to reconsider the arguments which are generally urged to prove that he was a historical person and responsible for the formulation of the Jain faith..
In the old Jain writings, mostly canonical, we get a number of references to Parsva and his teaching. The information thus obtained amounts to very little, and consequently we lack a clear picture of his personality and of his views. The Jaina tradition in the Kalpasūtra furnished us with very few facts of his life. He is said to have been a son of king Aśvasena of Benares and his wife Vāmā. He lived for 30 years as a householder, then became an ascetic and after performing penance for 84 days became enlightened. He lived for full hundred years and died on Mt. Sammeta in Bengal some 250 years before Mahāvīra. He is uniformly referred to by the use of the epither purisadaniya 'beloved of all men' which may point to his genial temperament and kind nature.
In his personal life, there is nothing unusual and improbable. But this in itself is not sufficient to regard him as historical. While the duration of hist life of hundred years is based on the mythological scheme of decreasing age of each succeeding prophet and is itseld suspicious, the period of 84 days for his penance is certainly too short and is again due to the same consideration of a decreasing number of days for the austerities of each preceding prophet.

Page #265

256
Amrita
We may further note that while Mahāvīra, his successor, is brought into close relations with definite historical persons, like ruling kings and contemporary religious teachers, no such association is available in case of Pārsva, though we must also remember that he is, in no way, connected with purely mythological persons as well, as is the case with his predecessor Nemi.
We know something more of his teaching. From Vyakhyāprajñapti we understand that Pārsva believed in the form and eternity of the world as did Mahāvīra. The followers of Pärśva preached that selfcontrol (samyama) leads to the stoppage of Karman (anhaga) and penance leads to its purification. Mahāvīra agrees with this as well. Both these pairs occur in the list of items causally connected and stated in a verse found in close association with the earlier statement both in Sthānanga and Bhagavati. We may, therefore, reasonably conjecture that it formed a part of Pārsva's teaching. There we are told that waiting upon the teacher (pajjuvāsaṇā) results into the hearing of the doctrine (savana); this, in turn, gives rise to knowledge (nāṇa) and so on to discrimination (vinnāna), stoppage of Karman (ananhaga), penance (tava), purification (vodāņa), absence of any activity (akiriyā), and finally liberation (nivvāna). This view is in substantial agreement with the later teaching of Jainism though such a series of causation never gets any importance in the fully developed Jain system of dogmatics. More, well-known is Pārsva's teaching called căujjāmadhamma which is contrasted with the pañcamahavvaiyadhamma of Mahāvīra. The four vows of Parsva usually enumerated are savvão pāṇāivāyāo veramaṇam 'abstention from all kinds of killing', savvão musāvāyāo veramaṇam 'abstention from all types of falsehood', savvão adinnādāṇāo veramaṇam 'desisting from all kinds of thefts' and savvão bahiddhādāṇāo veramaṇam 'abstention from all kinds of giving out.' The exact meaning of the last vow is a matter of some doubt. The traditional explanation, as for instance given by Abhayadeva, in his comment on the passage in the Sthānānga, takes it to mean either abstention from sexual intercourse (bahirdha maithunam) and other possesions (ādānamparigrahah) or abstention from possession not necessary for religious observance. With greater probability Schubring suggests that the reference is to the giving out of the semen and the vow demands chastity. In view of the special emphasis on the confession in the usual phrase pañcamahavvaiya sappaḍikamma dhamma of Mahāvīra in contrast with that of Pārśva, it is natural to imagine that the formula of pratikramana was lacking in the preaching of Pārsva. From the story in the last chapter of Sūtrakṛtānga, we come to know that in the vows to be accepted by the followers of Parsva specific words like

Page #266

Pārsva's Historicity Reconsidered
257
tasabhrūe and thāvarabhūe were inserted, which restrictions were not admitted by Mahāvīra. Finally there was the outward difference between the two sects; Pārśva allowed the use of a white garment for the monks, while Mahāvīra preached complete nudity.
Numerous argument are advanced to prove that Pārśva was a historical person. Of these, the following deserve a careful consideration. It has been pointed out that the Jaina canon supplies us with many indications of his teaching and tells us of his followers. One of them Kesī who plays an important part in the two works Uttarādhyayana and Rājapraśniya, is mentioned in such a manner is to suggest that he was a real person. In Sūtrakrtānga Udae Pedhālaputte, another disciple of his, holds a diputation with Gautama the chief disciple of Mahāvīra, while in Bhagavati Pārsva's follower Kālāsavesiyaputte expresses his desire to exchange his law for the vows of Mahāvīra. Secondly we are told in Ācārānga that the parents of Mahāvīra belonged to the following of Pārśva and are called pāsāvaccijjā samanovāsagā. Jacobi has, moreover, brought forth a far more important evidence to prove that Pārśva was a historical person. In the Sāmaññaphalasutta of Dighanikāya, it is stated that Māhāvīra preached the cātuyyāmadhamma. A Jain monk is called cātuyyāmasaṁvarasamvutto, while, in reality, it was Pārśva and not Mahāvīra, who taught the cāujjāmadhamma, as we know from the Jaina sources. Thus a natural blinder on the part of the Buddhist writer proves that followers of Pārsva's teaching must have existed at the time of Mahāvīra, to make such a confusion even possible.
We should not, however, overlook the limitations of all these arguments which are not conclusive in proving Pārsva's historicity. The Jain tradition attributes the cāujjāmadhamma not only to Pārsva but to all the 22 Tīrthakaras except Rsabha the first and Mahāvīra the last prophet. The distinction thus appears to be the result of a priori theory. Both Sthānānga and Uttarādhyayana emphasise that the 22 prophets from Ajita to Pārsva preached the cāujjāmadhamma and explain the difference on the basis that their lives were without hardships. The presence of Pārsva's followers at the time of Mahāvīra, and the statement that his parents were followers of Pārsva are chiefly due to the ordinary considerations of logical necessity. Believing in the eternity of the religion, the writers could not but make the religious preaching before the time of Mahāvīra as the work of his predecessor Pārsva. And we find in the canon stories of earlier Tirthankaras like Nemi their contemporaries and disciples. Even the statement about Mahāvīra's parents following the creed of Pārsva is subject to doubt. There we are told that they

Page #267

258
Amrita
performed pratikramana which is thus attributed to Pārsva, while we have reason to believe that it was lacking therein. The passage in the Buddhist canon is also not conclusive. It not only commits the mistake of attributing the cātuyyāmadhamma to Mahāvīra but also incorrectly states the nature of these four vows which does not agree with Pārsva's teaching. We are thus led to believe that the writer was not well acquainted with the caujjāmadhamma of Pārsva. The statement is clearly the result of a superficial acquaintance of the Jain scriptures, which regards everything contained in it as the doctrines of Nigantha Nātaputta. Moreover, there is no evidence to show that the mistake was committed at the time of Mahāvīra or even of Buddha, as the canonical passage is obviously of later date. There is still less reason to believe that Pārsva lived some 250 years before Mahāvīra. In fact, the Jaina tradition itself tells us that Subhadatta, one of the eight disciples of Pārsva, became the head of the church after him, and was followed by Haridatta, Āryasamudra, Prabha and Keśi, the last a contemporary of both Mahāvīra and his disciple Gautama. He was thus separated from Pārśva by at most four generations and a period of 250 years is certainly and unusually long one for them.
We thus find that the evidence available so far is not sufficient to prove that Pārsva was a historical person, and some more evidence must be brought forth if we are to believe in his historicity.
OOO
Pārsva's Historicity Reconsidered
All India Oriental Conference 13th Session 1946

Page #268

Group 3 : Sanskrit Language

Page #269


Page #270

Patañjali on P. VIII. 1.1.
The evaluation of the teaching of the classical works, the language of which is no more a spoken tongue, depends upon a correct and precise interpretation of their words and arguments. This can be illustrated from a passage of the Mahābhāsya on P. VIII 1.1 Heta which introduces the topic of repetition of a word to convey some additional meaning of the words used. The necessity of using the word or in the sūtra is explained by the Vārttika I, which points out that the possibility of repeating the last sound of the word by applying the traffic rule 316 START should be avoided. This is justified by pointing out that the meanings of such formations, viz. fra in the sense of continuity (37192) and all in the sense of all-inclusiveness (1) could not be conveyed unless the whole word is repeated. Then follow the words se af R A ' rinfo fedett quefira TRIA. Avoidance or exclusion is one sense in which the pre-verb of is repeated as laid down in the sutra VIII. 1. 5. Hence the most natural interpretation of the words of wat would be to state that at least in the case of y, even the repetition of the final sound can convey the expected meaning of exclusion."
This sense is supported by the introduction of the next Vārttika which gives another reason for the use of the word om in the sūtra. The sūtras in this group use the words to be repeated in the Gen. case and therefore a reference to that case is necessary in the first sūtra so that what word is to be repeated is made clear. This alternative is suggested to avoid the contingency of repeating only the last sound of a word even in the case of a single rule in this group.
Kaiyata and Nāgesa find no means to show that a repetition of the final sound could in any way convey the intended meaning of this process of fara Hence they take the words of Patañjali to mean that an unacceptable possibility may arise in which the final sound when repeated should convey the sense of 'avoidance'. It is thus an suffers for the futur. But by the same meaning

Page #271

262
Amrita
it will constitute an şeyes for the yayfer. As no conclusion is drawn from the situation, Patañjali appears to take it in the second sense, and hence he tacitly accepts such a possibility and thus comes to a stalemate in the argument.
How was such a possibility imagined by Patañjali ? A clue for it can be found in the general discussion about repetition which follows later. Patañjali classifies repetition into three types, according as the element repeated is a पद, a वाक्य or a मात्रा. Of these the last two are rejected by him as unsuitable for this purpose : न च वाक्यद्विर्वचनेन मात्राद्विर्वचनेन वा नित्यता वीप्सा वा
red. In the light of this remark, the application of the rule 37615 paper can give us a repetition of a hill, if the final sound of a word is a vowel. This condition is satisfied by the pre-verb f and the repetition of its final 3 will give rise to the form and this can be accurately described as a repetition of a E. Hence what Patañjali could mean by his statement in question is can possibly give us the meaning of परिवर्जन. Katyayanas Varttika परेर्वजने वावचनम् on the sūtra makes the use of a simple in this context as an acceptable idiom and Patañjali supplies the example uf famosi e da:. For the sake of argument Patañjali reads i i.e. uf with a lengthened final vowel. Like other pre-verbs of disyllabic nature 371, ufa etc. uf has a lenghened form in a large number of compound words. PW has recorded as many as 24 such cases from RV onwards. We may add to it a technical word from Jain literature mine which always occurs in this form.
This view has an unexpected result on the interpretation of the Vārttika I on the same sūtra VIII. 1.5. Kātyāyana says that u can be repeated when no is involved. Patañjali's example is id qe da:. This bars the use of परिपरि occurring in a hypothetical expression like परिपरित्रिगतम्. This is strange because such a restriction is valid for all the pre-verbs admitting of repetition, like अधि or adverbs like अधः, उपरि etc. A समास-like formation is later admitted in the sūtras VIII. 1. 9. and VIII. 1. 11. for such repeated words which do not form compounds like Tanah etc. Patañjali's attempt to deny this Vārttika any validity by attributing the meaning of वर्जन to the समास and not to yf is not convincing.
What is then the significance of this Vārttika? I venture to suggest that what Kātyāyana means is that any attempt to explain it in a compound as a use of repetition and assign it the meaning of Taufa is not correct. This specific meaning of ufis available only outside a compound and no compound with repeated f is acceptable.
Patañjali on VIII. 1.1.
प्रसन्न पारिजात
2

Page #272

Pāņini - VIII. 1. 7
In the context of repetition (dvirukti) which Panini deals with in VIII. I, a problem of interpretation arises when we take a close look at the Sūtra - 39082Ezta: a VIII. 1. 7. It lays down that the pre-verb 37 and the two adverbs 34f1 and 37877 are repeated in the meaning H1414. As usual, the Loc. is used to indicate the semantic content of the formation. Now a question is raised whether it is the specific sense of the whole expression or a mere condition which qualifies the normal meaning of the words used. In the present case, the commentators and other grammarians differ from each other, and the practice of the later writers is found to be not decisive because of the influence of the theory of the grammarians, which they follow. The महाभाष्य of पतञ्जलि is not available on this Sutra and the usual explanations come from the Kāśikā. The wording of the Kāśikā itself admits of different interpretations because expressions like सामीप्ये विवक्षिते or न सामीप्यं विवक्षितम् are not sufficiently precise for our purpose. Both the pit and the 449 are definite on this point, as can be seen from their words like 39tai falou and उपर्यादीनां योऽर्थस्तस्य समीप्यद्योतनाय द्विवचनम् On the other hand, the वृत्ति on the
S cho VI. 3. 3, appears to suggest that 'nearness', or proximity is the meaning of these reduplicated or repeated words, because his counter example is 37f8afa. Hemachandra is far more definite and his own example at the 2014 20. 83 is clinching. Taking me as the intended meaning in this has the peculiar effect of making all the three repeated words. उपर्युपरि, अधोऽधः, and अध्यधि synonymous, which implies the rejection of their original meanings, which are at least two, 'over' and 'below', basically different and of opposite directions. This is exactly what the commentator on the व्याश्रयकाव्य does, as can be seen from his words तरून् अधोधो वृक्षाणामासन्नः, 3fa44f4f G:GRHH, and 3787787414, 342 3477774 where the context prevents the original meanings of these words.
The lateness of these explanations may cast doubt on their validity. But

Page #273

264
Amrita
TL
there is an accidental statement about the same in a work clearly older than the Kāśikā. While discussing the difference between Tag and oft as semantic concepts, the 1441 presents us with a passage which has some textual value. It reads like a summary of the H64107 on this section of Pāṇini. वात्स्यायन takes नित्य to mean अनुपरम, वीप्सा to mean व्याप्ति and uses the words अतिशय, परिवर्जन सामीप्य and प्रकार as does पतञ्जलि in the same senses. He, however, adds, examples which also occur in the E14T and in case of the present where ysret's comment is not available, he supplies the example 378zfe ga sai f ufifa H, which is certainly older and clearer than that of Kāśikā. Here the meaning given is सामीप्यम्.
The examples given by the later grammarians are also not conclusive and raise more problems in interpretation. Kāśikā gives no examples for 34fet, but uses the other two adverbs 34 and 3787: to make a further distinction of an into a temporal and spatial proximity. In the latter case, the distinction between 34 and 34f4f is taken to mean that the first only indicates the upward direction only, while the second gives the sense of immediately above. But examples which go against this distinction are explained away as in the case 34f f t ¿ ref the meaning is said to be only an upward direction and no proximity. When 34f is repeated without a noun following it as in उपर्युपरि पश्यन्तः सर्व एव दरिद्रति the meaning is said to imply वीप्सा and not
I . In the case of temporal proximity, another type of difficulty ar which is solved differently by the न्यास and the पदमञ्जरी. In the example उपर्युपरि G:44, 14 remarks og G:GHSnifa4, 3fachlightfa al canard, which implies two types of temporal proximity, the immediately preceding and the immediately following moment qualifying the word g:0. He thus takes G:GH as Nom. sg. and 34f is made to refer to an earlier and later time both of which are doubtful. ga, on the other hand, takes g:04 as the Acc. sg. governed by 344ft on the basis of the Mihaifa, 394a aref feruifay fergi द्वितीयानेडितान्तेषु ततोऽन्यत्रापि दृश्यते ॥ Kasika intended दुःखम् as Acc. sg. is clear from its other examples 34ef4f 4144, and 378TSET FR4.
Taking a historical point of view, we may note that the usage of repetition only begins at the later stage of Vedic language, and that also in the sense of any which will have to be defined more precisely. The repetition is also confined to the preverb 377&t, which usage is abundantly illustrated in the sita and TEJ Es at the end of the Vedic period. The repetition of 37874 and 34f appears only to begin at this stage and then becomes prevalent in the classical period, and 37eze becomes obsolete except in the grammatical literature. कात्यायनश्रौतसूत्र uses उपर्युपरि and अधोध: only once

Page #274

Pāṇini - VIII. 1.7
265
(8.9.10; 9.2.16) while dache has one case of 34ef4f (17.7). There are scores of examples of 378zmfy in the older site and TET LES and usually mean 'above', or 'over' something though the commentators indiscriminately explain it as either उपरि or समीपे both, in phrases like सामीप्ये, उपरिभावे and उपरिष्ठात् समीपे all under the influence of the grammatical teaching as understood by them.
A scrutinization of the examples in the state of 371464441A RA, 47G/ST, ARTE, 41a, fexuel doft and ster gives us the following result. In almost all the cases, they express the meaning above and the other shades of meanings are either due to the context of other words or the specific situation in which they occur. Mostly the Acc. of the words are used but the Loc. is also found, which fact is true of the unrepeated word 31 as well. When the objects are two or more, the repetitive sense is also found, otherwise some emphasis is implied. All these nuances of meaning are also found in all kinds of repetition and cannot be singled out for these three words alone. What Pānini then means by 11 is the quality of being in close proximity in the activity in the direction of above (with 3hfe and 34ft) and below (with 37€:) and can be rendered by such English expression as 'higher and higher and lower and lower',.but without the component of comparison which is present in the English expression. With a plural object, the meaning will be one above the other without a break’ a kind of continuity which is indicated by the word in between any two stages of the act expressed, whatever the actual physical fact. In the words of the air, it is the HId ry which is indicated in the repetition of these three words more than 31CERT, 37877901, 3414rty on the part of the speaker and the hearer. cf. JENGO on 2.1.67.
000
ni VII. 1.7,
Pāṇini VIII. 1.7,
BDCPRI 1991
BDCPRI 1991
2

Page #275

Idiom-formation and Pāṇini
Idioms are always a problem for linguistic analysis, both on account of grammatical irregularities and semantic changes of an unpredictable nature being involved. This has led to the traditional mode of treating them as stereotyped and in no need of any kind of explanation. But both the facts of formal irregularity and meaning development are of constant occurrence in a language and need some kind of historical explanation based upon usually accepted changes.
In the grammar of Pāṇini, many, if not all, formal irregularities are treated as individual facts and all such formations are given without analysis, and are traditionally called nipatita. Even then from the days of Patañjali, the question is raised 'for what purpose a particular form is so treated and what irregularities are involved in it. In some cases, alternative suggestions are made when more than one irregularity is involved, and sometimes alternative analysis of the given word or form are suggested. But usually no ,attempt is made to explain the unusual sense of the idiom by referring to semantic changes involved in their formation, because, semantic changes are not taken note of in Pāṇinian grammar except in a very general way of making precise the meaning of each basic unit postulated for the analysis. Thus the roots and suffixes with substitutes are given meanings and the elements called agamas are treated as having no meanings of their own. In this situation, an attempt has to be made to trace the origin and development of meanings of idiomatic formations with whatever evidence is available for this purpose. This involves some amount of speculation and this must be combined with evidence coming from different directions, particularly the precise and original meaning of the basic elements of the idiomatic expression.
Pāṇini gives in his grammar meanings of the completed formations only when it has some bearing on the expression itself. Otherwise he assumes

Page #276

Idiom-formation and Pāņini
267
that the meaning of the expressions are known to the speaker, as the language analysed is his native tongue. This applies as much to simple words as to idiomatic expressions. This distinction is vital. While normal formations like krt and taddhita have their meanings explained by the meanings of the corresponding suffixes, idiomatic expressions can call for such an explanation only when they contribute a complex formations of two or more elements like the upasarga, the gati and the karmapravacaniya, where the two elements have their own specific meanings, continue and produce a semantic change in the final expression. One such group of cases is provided by Pāṇini's grammar in his treatment of gatis which are diverse elements combined with some verbal meanings as seen in his definition fher. These are collected together in 1.4.60-80. All these formations are considered as idiomatic in nature, and one obscure idiom is investigated detail to see what process the idiom formation involves.
Sūtra 1.4.66 Utane Grecitand' tells us that Pāņini had in his language two idiomatic expressions with same general meaning, formed with the first elements as out and joined with some verbal nominal form. This much is certain because he states the same meaning for both the idioms by using the sg. of the compound expression Trend. What particular verbal form was used is not made clear and there is no Vārttika and Mahābhāsya on the Sūtra available. Later writers from Kāśikākāra and others have surmised that the root could be to kill, because the given meaning uses a nominal form of that root in the word ufreid. Once the root is fixed, it is natural to take its gerund as the second member Eral, because the purpose of giving these two words, the designation na can only be to form a compound expression with a nominal form of a root (Eger), which should lead to a change of forms from bral to FR (cf. p.VII.1.37). All examples given for this Sūtra are mostly the same UTERY and ECT, but we should not necessarily exclude any other verbal form. Sūtra 1.4.69 suggests that Panini would have stated the root itself for this purpose, if it were the only root used, as in the case of ac in relation with 3409, and then avoided to give the meaning of the idiom as being unnecessary. The only reasonable conclusion we can draw in this context would be to think of a verbal form having the meaning 'to kill' for example a root like मार् (tr.), may be taken and a form like मनो-मारं can be constructed, but no such case is attested.

Page #277

268
Amrita
If we scan the meanings given in these gati Sūtras of Pānini, we find them as follows : 34 :, 40, 374RUe, maglucard, 3194681, 3777eit, 34RT, 34247 and den, all of which are such as to indicate the general meaning of the idioms formed and not the specific meanings of individual words used in them. In this respect they differ from other meanings given in Pāņini's Sūtras, when they refer to a specific word, to restrict its meaning for a particular grammatical operation 3. 1. 46 fiş 2. 4. 46 uit T CTA. Most of the wife words are given the Hill of quite under this condition of restricting their meanings cf. 1. 4. 83-98. This presumes that the specific meaning either originally belonged to the particle or was a figurative use of that word, or may have no specific sense to denote (cf. 1. 4. 93). Rarely in case of compounds of two nouns, the specified sense necessary for compounding them is given which is the result of the compound formation and can be called the meaning of the 4414 itself, as in the case of 2. 2. 26. faş14PTRIC where the resulting meaning can be easily referred to the primary meanings of the two words which are compounded. This is not the case of the GTAS implied in P. 1. 4. 60-80 which can be considered as having a meaning that can not be derived from the meanings of the fact word, or the word with which it is compounded. The roots which are mostly used in the process of idiom-formation are given as कृ, भू, हन्, धा, नी, गम् and वद् which are susceptible to numerous collocations to develop idiomatic expressions. Neglect of such general considerations have led the commentators to suggest all kinds of meanings to be assigned to the words or the idiom as a whole. Thus Kāśikā has assigned the meaning of 'desire 39146141 to the word all, and the sense of satiation or satisfaction of the desire to the word ofera. Both the expressions are given the same sense, so that कणेहत्य and मनोहत्य with पिबति means ataq fuafa lag t fraa: 1 star fent ref: The primary meaning of the word of is given as 'grain' and is said to be a distinct homophonous word of out in the idiom, as no semantic change can be easily imagined. The
14 goes a step further and says that means exclusive desire (अभिलाषातिशय) and thus मनस् also means the same thing, because it is put along with the first (HTML) The commentary a14-RAT raises the question 'how can we get the meaning श्रद्धाप्रतीघात, when neither the word कणे nor मनस has the meaning of cel ?' It answers by saying that for this very purpose the word कणे is explained as अभिलाषातिशय. It is strange that the meaning of अभिलाषा - should not be connected with first and then extended to cut by association, unless these commentators consider that 371141 has nothing to do with in the language. Finally TTT in his EROGATOR comes to the

Page #278

Idiom-formation and Pāṇini
269
conclusion that अतिशयाभिलाषानिवृत्ति is the meaning of कणे.
The later grammarians have tried to make the sūtra easier to understand by paraphrasing the somewhat obscure word slecitand by words like to (FCERA) and TH (ES).
The finding out of counter-examples for the absence of a 4414 in this idiom has equally puzzled the commentators. Kāśikā's examples on Brail a: etc. are not explained in what sense they are used. Only alMHI makes a feeble attempt to explain these examples by stating 74: gugalcea: 001: afged Eral TG: grief: H y gfe falez Tarief: thereby suggesting that the two idioms have not exactly the same sense. In the first case, if aface is not a purely scholastic paraphrase of the loc. of the word out, it could mean 'for the sake of a grain, killing is done because to take it to mean striking on a single grain, is, to say the least, verging on absurdity. His sense of the second idiom is much nearer the truth, and means, probably, that 'suppressing the inclination of the mind to drink poison, he went' but with two vital modifications : i) it is the intended meaning of the idiom and not its counter-example, and ii) what is intended is 'someone forces oneself to drink poison by suppressing the natural tendency of the mind against it'. Modern translations of Basu, Böthlingk, Renou, Katre and others offer no help whatsoever.
The best and possibly the only way to deal with these idioms would be to ascertain the original or older meanings of the words in the Sūtra and fix their normal usage. Then it will be possible to ascertain what figurative meanings can develop from them to give rise to these idioms. These words are श्रद्धा, प्रतीघात, कण, मनस् and the verb हन् or its synonyms. Of these the only meaning that can be assigned to put is a 'grain of corn' and 47 has the ordinary sense in Skt. i. e. mind as an internal organ of thought. The root 67 means to kill and then figuratively to suppress and foreld means a counterblow, hence to counteract something, to set aside a natural inclination. The word space has been studied in detail and has a long history from the time of RV to the present day. Here some distinction should be made for the three stages of its existence. Its original IE is sense-based on the comparison of its elements at (a) ET with Latin credo and the reconstructed base kred-dhe. Basing oneself on the European idiom to put one's heart into something the meaning of B MI or desire is taken as due to a natural semantic change. It is only in late Sanskrit, illustrated by examples from Mahābhārata, Rāmāyana, Kathāsaritsāgara and the lexicons that the sense of desire is attested for store. No sure examples for this meaning are found in the older

Page #279

270
Amrita
literature, and this meaning cannot be assigned to Pānini himself. Other than P. 4. 2. 66, the use of this word and its derivatives at P. 3. 2. 158, 5. 2. 101, 4. 3. 12, 5. 2. 85, do not show this meaning at all. As the equation between G and kred' is objected to on phonetic grounds, it is not safe to make use of it for this purpose. In fact, E. Benveniste has pretty nearly proved that the word for 'heart' in its earliest sense was more a physical organ than a mental state, and hence ideas like desire, will etc., were not associated with it. This is true of the whole range of Sanskrit literature and hrd or hrdaya is not associated with the mental states like will, desire, hope etc. Hence, Köhler's attempt to prove the sense of desire for act as being as old as its other senses, faith and devotion, cannot be easily accepted. As Benveniste remarks, we see that there is no need for the generosity, desire to give, which Köhler believed was the semantic constituent of (this) word'.. He believes that the original and basic meaning of the word was 'trust'.
The words derived from stased like asta, Go, Stalca, SH, (FYRICS) श्रद्धान, श्रद्धामनस, श्रद्धामनस्या, श्रद्धामय, श्रद्धालु, श्रद्धावान्, श्रद्धिन, अद्धिव, श्रद्धेय, and श्रद्धेयत्व, do.. not show any component of desire' or 'liking in them. The only compound which can be interpreted to mean, 'to give according to ones desire is rarely found in Sanskrit, while on the contrary, to give or get a 'boon (CP) according to one's desire is very common and usual. The change of meaning from trust, believe, faith etc. to desire is far more natural than a change in the reverse direction. The same is true of the word यथाश्रद्धम्.
With tl meaning trust, faith, conviction, the word ufagra get its usual sense of suppression or counteraction, something against those or inspite of Tecl belief. The use of the verb 5 used with verbs like 474 could mean ‘killing one's mind', suppressing its natural inclination. In this situation, one can also imagine that what is done is against one's desire which must have led a host of commentators to assign this meaning to steel. That a similar idiom existed in early days can be seen from a few examples in which 67 is replaced by the root hrefa which has become more usual in the Middle Indo-Aryan stage. One can consider the following two examples in this respect.
RHICHYPE II 157 3404 Teu Alfas tu tuffa het fe; II 163 fafetours मणु मरइ; II 171 सो जाणिवि मणु मारि; दोहापाहुड १४ मोहु विलिज्जइ मणु मरइ; ९५ अप्पा परहं ण मेलविउ मणु मोहिवि सहस त्ति.
The following idioms in some Modern Indian languages can also be considered. Marathi HRU- T GC Jau (Date); HISU-TGE2 Suofas

Page #280

Idiom-formation and Pāṇini
271
जाणे; Gujarati मन मारQ to restrain the mind; Hindi मन मारना etc.
The cumulative evidence given above leads to the conclusion that one of the two idioms stated by Pāṇini in 1. 4. 66 means 'to do something against one's trust, faith,' where the root is used in a slightly figurative sense 'to suppress.'
II
Once the meaning of the idiom Fler is ascertained, it will be now possible to deal with the twin idiom Ery. We have now the advantage of knowing what the meaning condition of this idiom is. It must be the same as for the other idiom. But we have the further problem of knowing what the form and meaning of the first element of this idiom is. Whatever the theoretical and recondite explanations of the grammatical tradition, it is the loc. sing of the word and it forms a fhre with without the loss of the case termination of the Loc., a kind of 370kr. Pānini admits 3461477 of the various cases, which are dealt with in 6. 1. 1-24 and enjoins the अलुक् for the other cases पञ्चमी, तृतीया, चतुर्थी, सप्तमी and षष्ठी and excludes the first two cases. All other cases except सप्तमी are treated with specific reference to a list of words. The scope of Loc. is much greater than all others. Hence instead of putting a condition of a specific word and include it in that list, Pānini's normal practice would be to give the form itself while defining fas. Hence 2014TH. To set up an adverb of a homophonous nature of the Loc. sg. of the word out as the later grammarians do, is due to purely theoretical considerations. And there is no cogent reason to set aside its simplest interpretation that the word ", means the smallest unit of grain, particularly food grain. The difficulty lies in tracing the evolution or historical origin of the idiomatic use to which it is put, which can only be the same as the other idiom.
There is an idiom in Marathi to the effect that the idea of the depriving some one of his meagre means of maintenance is expressed by the phrase R HRU (Lit. to strike on the belly). There is a similar idiom 412
1 (killing the belly i. e. hunger). What we now have to concede is a figurative sense for tour which will stand for the belly, and which can be considered as natural in view of the fact that the filling of the belly is the normal expression for taking food. If this is conceded then we have good reason to associate these two idioms into a normal group based on the fact that, what is suppressed in the interest of something to be done, are the needs of the body and the conviction or belief of the mind, against which may stand the demand or compulsion of something to be done or expressed by the main verb, in which these idioms can occur.

Page #281

272
Amrita
Much of this speculation could have been avoided if we had a couple of early examples of these idioms in a natural linguistic and semantic context. But all that we have about their use is found in grammatical works based on the rule and the meaning assigned to it by them.
C. G. Kashikar Felicitation Volume

Page #282

Systematics of Panini's Aṣṭādhyāyī
The systematics of Panini's Aṣṭādhyāyī gives an impression of being somewhat enigmatic due to the extreme brevity of expression, interdependence of the various sūtras or rules, the order of rules and their exceptions, the presence of an anuvṛtti in an unpredictable manner and the numerous mnemonic devices used in a complex way. This gives rise to a question. can Panini be understood by himself or must we rely on the traditional interpretations not only as a help for grasping him but as being an essential part of the system and as such originating from the author himself, may be only in an oral tradition which was later fixed in a written form by successive writers as and when need for it arose? Once this apparently reasonable position is accepted, more doubtful methods of interpretation like a discontinuous anuvṛtti, the principle of s and dictums like a f Haf force on us to contend with when this attitude is coupled with another which denies any change in the accepted language and therefore any change in the rules of grammar, more unnatural procedures like, व्यवस्थितविभाषा, नियमार्थत्व of a rule and many others come into play to make the system cover many forms and constructions which came into vogue at a much later date. Besides grammar such a development is also seen in many other disciplines in Sanskrit. A proper analysis and evaluation of such methods may help us to get at least an imperfect glimpse into the historical development which lies behind them. One may even make a guess that in all such cases the historical facts have been at best reinterpreted though fortunately not completely destroyed.
But we also note another tendency of a limited extent, in which the historical development is almost completely lost. As a result of a discussion about the correct formulation of a rule, changes in its wording or arrangement are suggested which are considered as more satisfactory or more elegant. This new tendency simply substitutes this change in place of the

Page #283

274
Amrita
original formulation which thus completely disappears leaving to trace behind. The puzzling systematic of Pāṇini is, in part at least, due to such rewritings of his text. Once they become the accepted text, the implication on which they are based become a part of the system without which it cannot be understood. In such cases, the only method available for a modern student of Pānini is to try to find out. relics or traces of the original state of affairs left over by oversight or neglect. Sometimes inherent probability or statistical evidence which was beyond the powers of the redactors does help. The thoroughness with which such changes are carried out by later writers on Pāṇini leaves only a marginal scope for such research. The results in this field can only be more or less probable. Any rigorous proof is out of question.
A good example of such a rewriting of a sūtra is found in P. vi.2.42 which is generally given as follows: kafi'I HERC-37C CIEG yrasal तैतिलकद्रूः पण्यकम्बलो दासीभाराणां च । In this form it contains six distinct words, the first of which consists of four compounds forming a song with a singular ending, followed by three independent compound words each in Nom. Sg. and ending with the word crir ending in Ger. pl. and the particle a. The original formulation of the sūtra is preserved in the महाभाष्य of पतञ्जलि which runs as कुरुगार्हपतरिक्तगुर्वसूतजरत्यश्लीलदृढरूपापारेवडवातैतिलकः. पण्यकम्बलोदासीभारादीनाfufa athy I. which itself is a modified form of the original which must have been without a fart after afastone and w'thout a Nom. Sg. 37 after your chien and ending in दासीभाराणाम्. Without such a supposition the remarks of पतञ्जलि which follow तत्तर्हि वक्तव्यम्। न वक्तव्यम् । योगविभागः करिष्यते । कुरुगार्हपतरिक्तगुर्वसूतजरत्यश्लीलदृढरूपा पारेवडवा तैतिलकद्रः पण्यकम्बल इति। ततो दासीभारणा । चेति । तत्र बहुवचननिर्देशाद्दासीभारादीनामिति विज्ञायते | makes no sense at all.
The reason given for the addition of 3fc after R is to include such words as defa: tarifa: arifa: 311919, 5h: in the scope of the rule which enjoins पूर्वपदप्रकृति- स्वरत्व. The initial impetus to cut the Sutra into separate forms of words was given by Patañjali as a suggestion to avoid the addition of the word off to the sūtra and yet leave scope for the inclusion of the additional words by making use of the device of taking the plural at the end as indicative of a Tu. This is one of the traditionally accepted devices for this purpose which does occur in other places as well. But the procedure of Mayr also needs the addition of a Nom. Sg. ending to quel chest and तैतिलकद्रू Neither पारेवडवा nor अश्लीलदृढरूपा require any phonetic change to be . taken as Nom. singulars, being feminine words ending in ā. Then why not carry the procedure a step further and separate असूतजरती and अश्लीलदृढरूंपा as independent words as well ? And this is precisely what the 2014 and 484329

Page #284

275
Systematics of Panini's Aṣṭādhyāyī
do and go a step further and regard the first two compounds as also separate words. Both write तत्र कुरुगार्हपत रिक्तगुरु इत्येतौ अविभक्तिकौ इतरे तु असूतजरतीत्येवमादयः प्रथमैकवचनान्ताः । सा च प्रथमा सुब्व्यत्ययेन षष्ठ्याः स्थाने वेदितव्या । All this is done to avoid the use of af at the end and the interpretation of the Gen. pl. as indicative of a . At this stage only the last word is in the Gen. pl. and hence an addition of is essential to make it stand for a . A trace of the original position is preserved in the remark of कैयट दासीभाराणामिति बहुवचनं द्वन्द्वसमासत्रयं मत्वा a4 to justify the whole discussion. To justify the presence of case-suffixes in a compound he now adds विभक्त्यलोपस्तु सौत्रत्वान्निर्देशस्य स्यात् ।
Before we conclude that what is represented as a hypothetical argument is really a reversal of historical facts, it is necessary to ascertain the strength of the principle on which it is based. The TP of both the faчf and the g are indicated by three different methods in the viz. by the use of af in Mas. and Neu. gender, of fe in both these genders and the plural of the word with which the TT is supposed to begin. But the statistical distribution of these three is not such as can be attributed to a random choice. While the use of af appears to be the most frequent and later on the only choice, the use of fa and pl. appear to be relics of an older method or usage. According to tradition, the plural of a noun is used to indicate a in the following six sūtras ii.1.40; ii.2.38; .4.31; ii.4.61; v.4.139 and vi.2.42. It is now necessary to ascertain the significance of the choice of this method in these limited number of Sūtras. There is no af or भाष्य on ii. 4.61 न तौल्वलिभ्यः and v. 4.139 कुम्भपदीषु च But in all the other four cases, there appears to have been a pre- suggestion that these should be read as शौण्डादिभि:, कडारादयः, अर्धर्चादय: and दासीभारादीनाम् probably made by the af on the basis of either greater clarity, conformity with the current usage or refusal to admit the pl. as indicative of a T. In all these cases पतञ्चलि sticks to the principle बहुवचननिर्देशात्.... विज्ञास्यते । If अर्धर्चादयः and बहुवचननिर्देशात् सिद्धम् are real वार्त्तिक on P.ii. 4. 31, all others can also be taken as वार्तिक' and the non-acceptance of the principle of बहुवचननिर्देश can be considered as older than कात्यायन. All the other वार्त्तिक of कात्यायन make use of आदि to indicate a गण and only once is प्रभृति used for this purpose but never a plural of a word.
A closer examination of the few in which the plural appears to indicate a may help us to determine its exact significance. As stated by Pat., occurs in P.ii 1.40 and the plural is made to include other forms like अक्षधूर्त, अक्षकितव etc. and कैयट refers to a view that only synonyms are to be included (f). His next suggestion that the plural may be

Page #285

276
Amrita
26994441hef all is neither justified by usage in Sanskrit nor any other similar case. The original गण thus confined itself to words like शौण्ड, धूर्त, कितव which were synonymous, later extended by the addition of other words like 2015, yaitu, Haita, 371, 379, ug (differing from one school to another) and ufusa, कुशल, चपल and निपुण synonymous words of cleverness in general as against specific skill of the original group. In ii.2.38 Tafsi adds the words "TgM and Gus while others add the words TM, as, cis, 8, ara words indicative of bodily defects and still others गौर, वृद्ध, भिक्षुक etc. words having no semantic unity. Most probably the use of all such words in this context was pejorative and कैयट's remark साहचर्यात् आरोपितकडाररूपगण्डुलादि शब्द is sufficiently suggestive of this fact, and it is similar to the use of words like the, 294 etc. to indicate the excellence. In ii.4.31 अर्धर्चाः पुंसि च. Pat. gives the words कार्षापण, गोमय and He in both the genders. Later writers have taken this sūtra as the basis for indicating as many Sanskrit words as are used in two genders, a virtual लिङ्गानुशासन which could not have been the intention of the सूत्रकार. The necessity of defining a gender in a grammar of Sanskrit can only arise if the usage of a word differs from the gender which is taught for a particular formation, either a compound or a derivative. This is obvious from the immediately preceding rule ii 4.30 39491 704064. In fact what the rule tries to do is to indicate the extension of a gender as an alternative from Mas. to Neu. and vice versa. Yasfat's examples are of this nature. Later additions are not based on any such consideration. The plural in v.4.139 9794cy 9 has a different significance. It includes an indefinite number of words ending in 47 and this usage is fully justified in grammar. It is only in the TRAHETcfg that a single case like शुचिपद occurs where all others give शुचिपदी. Not much can be said about ii.4.61 न तौल्वलिभ्यः in the absence of the महाभाष्य on it. It may refer to the descendents of atract or a 10 of words.
All these cases of the use of plural are sufficient to suggest the principles on which this usage is based. It indicates either a group of synonyms or words ending in the same morpheme. Now coming back to E vi.2.42 the plural ending of RM114 could not have the function of indicating a group as none of the original considerations can be traced in it. The rejected view that it is the plural of a compound is the original usage, and af rejected it under the influence of a generalised theory that the plural of a word in the can indicate a tot as a mere group of words.
Once such a view was accepted as valid and also intended by Panini the way was open to make further changes in the P to achieve greater brevity. In this respect i.3.91 and 1.3.92 are interesting. Later tradition sets up

Page #286

Systematics of Panini's Astädhyāyi
for them two UP, ga: consisting of roots 1.777-799 and a subgroup C4: The महाभाष्य on vii. 2.59 न वद्भ्यश्चतुर्भ्यः points out the significance of the word चतुर्भ्यः as excluding the root कृप् which is included in the group वृत्तादि Here the word af can be safely omitted because of the word : following it. Once it is dropped here it is also dropped in 1.3.92 वृद्भ्यः स्यसनो: where it is required for the optional use of 34 which is a condition for the operation of the later rule vii.2.59. Once this procedure is accepted it was possible to drop af in 1.3.91 on the basis of iii.1.55 giving rise to : in place of द्युतादिभ्यः. It is less likely that आदि is not used in the earlier सूत्र but has been used in a later one. There is no other case where the plural of a root is used as an indicator of a and it is inherently unlikely in view of the original conditions of its usage.
277
If these observations have any validity, it is clear that the systematics of Panini's grammar is not only due to consolidation of the works of earlier grammarians but is also the effect of a reworking of his text by later scholars. They formulated additional principles and observed further linguistic material and reworked on the text in the light of these principles so carefully as not to leave any trace of the original position. This had the further effect that the newer text could not be understood without the assumptions of these new principles and then it was as easy and convincing argument that they were intended by Pänini himself. This very tendency has given rise to the later non-Päninian systems of Sanskrit grammar.
000
Systematics of पाणिनि's अष्टाध्यायी
Proceedings of the International Seminar on Pāņini

Page #287

Uktārthānām Aprayogaḥ
The system of grammar as formulated by Panini and developed by Kātyāyana and Patañjali approaches an axiomatic system to a considerable extent. As such its understanding requires a good grasp of the presuppositions, assumptions or postulates which underlie the system. A mnemonic stanza enumerates the essential elements of this system as consisting of dhātu 'root', sūtra ‘aphorism', gana 'class', unādi 'underived words', vākya 'sentence', linga 'gender', agama 'augment', pratyaya 'suffix', and adeśa 'substitute'. Patañjali himself defines upadeśa (glossed as adyoccārana i. e. elements stated at the very beginning of the system and hence its postulates) as consisting of dhātu, prātipadika, base form of nouns', nipāta 'arbitrarily formed elements or particles', pratyāhāra 'abbreviations', sutra, pratyaya, adeśa and agama. Basically the two lists amount to the same thing and the guiding principle appears to be the amount of motivation which can be detected in the linguistic expressions of Sanskrit, all unmotivated elements being included in the upadeśa.
The list also contains the elements of grammatical description as well. If we now add to them the various types of sūtras like samjñā, adhikara, paribhāsā, vidhi, niyama and a few others, we can divide the whole list into two distinct groups, one dealing with the basic elements of language and the other belonging to the analysis and description of the language, elements of the mechanism of grammar or features of the metalanguage used. Here we are concerned with the second type of elements used by Patañjali, in his Mahābhāsya.
These elements of grammatical description in the system of Pānini differ from each other in two respects, precision and scope. A pratyāhāra has the greatest precision while the paribhāsās (including the laukikanyāyas similarly used by Patañjali) have the least precision. This scale varies inversely with the scale of extent or scope. Naturally these rules or maxims based on

Page #288

Uktarthānām Aprayogah
279
common sense have greater importance in understanding the nature of language than the more accurately formulated traffic rules of grammar in which we may also include the samjñās of Pānini. While the Paninian system is very well formulated as regards the second, the same cannot be said of the first. This is really due to the very nature of the material and one such case is studied here in some details to highlight the problems involved in them. . The list of paribhāsās in the system of Pānini, as also in the other systems of Sanskrit grammar, is a varying collection of maxims of diverse nature. The maxim uktārthānām aprayogah finds a place in the lists of pratibhāsās attached to other grammatical systems like those of Vyādi, Śākatāyana, Cāndra, Kātantra, Kālāpa and others. It is not formally included in the list attached to the grammar of Pānini. This does not, however, mean that it is not known to it. Patañjali makes use of it in his Mahābhāsya more than a dozen times. The other systems have simply given it a formal status. To judge from its different wording and the explanation of this maxim given in the commentary of Vyādi and those of Durgasimha and Bhāvamiśra on Kātantra, this paribhāsā appears to be very close to the principle of anabhihitatva of the Pāninian grammar. One is the converse of the other.
Unlike the other paribhāsās, this is not extensively commented upon and hence a survey of its use in the Mahābhāsya is essential. Patañjali uses it in numerous ways as a link in a long chain of arguments, which makes it necessary to follow them to some extent to ascertain its precise application. However valuable and intersting these arguments are, only the relevant portion is summarised here, neglecting much out of them.
P. i. 4. 83 lays down conditions under which an upasarga (pre-verb) can be given the designation karmapravacaniya (adnominal use). Almost all these conditions are semantic in nature because some of the upasargas have developed an adnominal use in some specific meaning, thus detaching themselves from the verbs. The theory of the origin of the karmapravacaniyayas based on the explanation of the name is substantially correct. These are said to have expressed an action earlier but now do not do so (karma uktavantah). This amounts to saying that originally they were used as pre-verbs with some specific verb, the meaning of which they modified, but in course of time, this verb was no more used and another took its place. This broke the bond between the pre-verb and the verb and the upasarga got attached to the noun, giving rise to an adnominal usage. Thus śākalyasya saṁhitām anu prāvarsat came from saṁhitām anuniśamya where

Page #289

280
Amrita
anu was a second pre-verb of the verb nisam — and with the disuse of the verb got attached to the noun samhitā because it would not be construed with pravrs-. Though we cannot now ascertain the actual examples which may have given rise to such changes, the mechanism is clear enough. P. i. 4. 93 states that in case of adhi and pari the condition is anarthakau producing a piquant situation because this amounts to no semantic condition at all. Patañjali attempts to solve the problem by assigning two different meanings to the word anarthaka. One way is to take it to mean 'having no meaning' and thus assign to these two the karmapravacanīya status unconditionally. The ultimate gain would be that they can be called prātipadikas even when they have no meaning (artha) which is necessary for such a base. The second interpretation takes the word to mean 'having no additional meaning than that of the verb to which it is attached. This means that they are superfluous and therefore militate against this maxim. They could not be used. Patañjali is thus driven to the conclusion that the maxim is not acceptable : words, the meanings of which are already expressed, can also be used. He justifies it by citing examples like apūpau dvāv ānaya 'bring two cakes, brāhmaṇau dvāv anaya 'bring two Brahmins' where the duality is expressed twice, once by the numeral two and again by the dual suffix. One can go a step further and state that it is expressed three times, another time by the dual suffix of the numeral dvi.
P. viii. 2. 83 states that in a conversation involving greetings, in the answer of the greeted person the final part of his utterance beginning with the last vowel (called ti) is prolated. This prolation is to be used only in the principal word when it comes at the end of the utterance' and not a subordinate word. Thus in devadatta kuśaly asi in the counter-greeting of the teacher, the final i is not prolated. This leads to a difficulty in a case like adheyo' gni3 rnādheya3, because the word agni is considered principal and adheya as sub-ordinate. Patañjali solves the problem by calling adheya qualified by agni as principal, the whole idea being 'the setting up of the fire'. This entails the consequence that the word agni must be used twice one for each alternative 'setting up' and 'not setting up', being an integral part of the meaning. The use of the word twice is avoided by invoking this maxim. A consistent use of the rule would also lead to the dropping of the word adheya as well. No, the truth is that the rule need not be always observed :uktārthānām api prayogo drśyate.
Both these examples bring out the fact that this maxim is used to avoid a preliminary objection of no linguistic significance. The answer points

Page #290

281
Uktärthānām Aprayogaḥ
out that this is not an absolute rule. This paribhäșă is anitya.
In other cases Patanjali admits the validity of the maxim and uses it to parry a point raised against a statement or adds restrictions and limitations on its application.
In P. i. 2. 51 a statement is made to the effect that when a derivative suffix is dropped with the use of the word lup, and the derived word remains identical with the base but adds to it the meaning indicated by the suffix, the gender and number of the derived word remain the same as that of the basic word. Thus by adding the suffix an to a word like sirisah (a kind of tree) in the sense of a place close to it and then dropping the suffix (by the use of the word lup and not another like luk) the word comes to mean a village near such trees. This new word, however, keeps the gender and number of the original word in spite of its new referent, a clear case of a semantic change. To explain that only the gender and number of the original word remain and not the other features like its being a plant, it is pointed out that in a compound like sirisavanam formed from the derived word the n does not change to n, which change requires the word to mean a plant.
A question is raised: Do other grammatical features of the basic word continue to remain in the derived word? When Katubadarī as a name of a village is explained as Katubadaryä adûrabhavo gramaḥ the Genitive suffix, will not be carried over to the derived word because it is neither a gender nor a number. Now an attempt is made to interpret the word vacana in such a way as to include the case as well. A distinction is made between its technical sense (päribhäşika) and its literal meaning (anvartha) deriving it as ucyate iti vacanam, 'what is expressed by the verb 'to utter', its object i.e. the concept of number'. The point made here is not quite clear. However we know that in Sanskrit the morphemes of case and number form cumuls and are not segmentable. If we now refer by vacana the suffixes themselves, cases. will also be included and the inference drawn that case suffixes should also be carried over to the derived word. Admitting this position Patanjali argues that because the derived word already includes the meaning of the genitive case in the process of derivation, it cannot be again expressed by its own. suffix on the authority of this maxim.
P. i. 2. 64 lays down the rule of ekaseṣa to explain the formation of the dual and plural forms of nouns, based on a semantic analysis, which, to say the least, is unusual. The dual of a word like vṛkṣa (a tree) is analysed as vrksaś ca vrkṣaś ca vṛkṣau and the plural as vṛkṣaś ca vrkṣaś ca vṛkṣaś ca

Page #291

282
Amrita
vrksāh. Applying the rule of ekaśesa, only one of these identical forms of the noun is retained and the others are dropped. This requires further rules for the dropping of the singular termination of the retained word and assignment of the dual and plural inflections to it so that we get the desired forms yrksau and vrksah. As the analysis is based upon the normal pattern of a dvandva compound, we must also bar that compound in favour of an ekabesa. Because a dvandva requires two independent words which are formally and semantically different from each other for compounding, for ekaśesa a kind of sahavivaksā (combined idea) or yugapadvivaksā (simultaneous idea) is suggested, which helps us to assign to it the required number. To emphasise this aspect the process of derivation is revised as vrksau ca vrksau ca vrksau and vrkṣāś ca vrkṣāś ca vrkṣāś ca vrkṣāḥ. .
To justify the preference shown to ekaśesa over dvandva, the rule of sāvakāśatva is invoked and it is shown that in the absence of such a preference ekaśeșa will have no scope left. This is objected on the ground that an ekaśesa can be formed of two identical verbal forms like pacati ca pacati ca pacatah where no dvandva can be formed. It is now argued that because the act of cooking, the meaning of the verb, is only one, we cannot use the same form twice due to the present maxim. Does this mean we cannot use the dual and plural of a verb at all? The answer is, we can, provided we consider the duality and plurality as not applying to the action expressed by the verb but belonging to either the subject or the object of that. verb.
But the rule has a wider scope. Whether it is the verbal forms or nominal forms, as long as they express the same idea, they cannot be used twice and hence no ekaśesa can be formed. Even in the case of 'dvandva, as long as we regard each word as expressing the idea of the other member, no usage is possible in the form of either a phrase or a compound. Patañjali suggests a number of devices to resolve the difficulty. For us what is important is to note that this maxim is not applicable to words which show any difference in the form or nuance in their meaning.
A well known feature of the IE. languages is that when a pronoun refers to nouns of different genders collectively, preference is given to the masculine form over feminine, and to a neuter form over the other two. Pānini has brought this phenomenon under the scope of ekaśesa in his rule i. 2. 72 tyadādini sarvair nityam. tyadādi is a sub-group of the larger gana defined by the sūtra P. i. 1. 27 sarvādīni sarvanāmāni. The peculiarities of this sub-group are stated in P. i. 1. 74, i. 2. 72; iii. 2. 60 and vii. 2. 102. This group consists of what are known as demonstrative pronominal adjectives and

Page #292

Uktārthānām Aprayogah
283
the numeral eka. Different situations are involved in the scope of the rule i. 2. 72. This has given rise to a difference in the interpretations proposed. What is meant by sarvaiḥ by Pāṇini cannot be now ascertained. To judge from his examples sa ca devadattasca tau, and sa ca kunde ca tāni, Patañjali appears to think that groups of nouns and pronouns are indicated. Later writers include groups of two pronouns as well. Patañjali's remarks on the Vārttikas 4 and 5 and his examples sa ca yaśca tau and sa ca yaśca yau would indicate that such groups do show the ekaśesa, but which one is optional.
A question is raised about the need of this aphorism. One such purpose would be to avoid a dvandva in the groups of a noun and a pronoun or a pronoun and a pronoun. This is not necessary because the situation is already covered by the more general rule that there cannot be a dvandva between words related as genus and species or a situation of hyponymy. Counter examples like sudrābhīram, gobalīvardam, trnopalam are explained away by giving unusual meanings to the second members. By such a procedure the situation is brought under the scope of this maxim, which thus is made to object to the collocation of two such words. Patañjali, however, rejects this scope of the rule on the basis of examples like tam brāhmanam anaya gārgyam, and the use of the principle that whatever can be expressed by a phrase can also be expressed by a compound. The upshot of this lengthy discussion is to admit some kind of restriction on the identity or similarity of meanings which should make the rule applicable to such words.
In this sūtra i. 1. 44 naveti vibhāsā Pānini lays down his definition of option in his grammar. Leaving aside the various implications of his diverse method of indicating it, we may take it that option in his grammar implies that two optional forms are acceptable and available to the user. Patañjali points out that this will be true if we take the view that words are constructed in accordance with a set of rules laid down, corresponding to the formation rules of an axiomatic system. The user is then free to apply a rule or not to apply, if an option is available, and two acceptable forms will be produced. But if the function of grammar is to assign a particular label or description to an already existing form, an optional form is not possible because we cannot assign to an utterance acceptability or unacceptability at the same time, nor at different times. This will make a given form sometime acceptable and sometime not, but will not give rise to two acceptable forms. To avoid this the option is now shifted to assigning a lable to an underlying expression from which some effects automatically follow and two alternative expressions become available.

Page #293

284
Amrita
In the use of productive operation a new difficulty crops up. Pāṇini lays down more than one suffix having the same function in the same sutra or adds a new suffix in the next sūtra with the use of the conjunction ca; both having the same effect, if we leave aside other advantages got from these two methods. Such a procedure is called yaugapadya 'simultaneous use of suffixes' and anvācaya addition to suffixes'. If we now begin to construct words with such suffixes, a question arises : 'Are we to use all such suffixes at the same time or one after another, to the same base?'
To take a concrete example, P. iii. 1. 96 tavyattavyānīyaraḥ prescribes three suffixes to form gerundives. Because of the linear nature of language we cannot add all three at the same time. The other alternative is to add them one after another to the same base which also is not desirable in this case even though the IE morphology allows a basic pattern like R + S + S + S, etc... The necessary limitation of this pattern is that S, S,, S, cannot be the same suffix with the same meaning. That is what the subscript numbers mean. The Indian Grammar states this restriction by saying that the suffixes come after the root and are immediately next to it. When the first suffix is added to a root, the resulting form generally ceases to be a root and the next suffix cannot be said to be immediately after it because of the intervening suffix. This however does not amount to making the first suffix a terminal or form-closing one. The first difficulty is solved by extending the scope of a dhātu to include also the derivative elements (sannantā dhātavah). and the second by allowing suffixes of different meanings to follow each other. The use of suffixes of the same meaning in succession is avoided by using the present maxim. If the first suffix satisfies our semantic need we need not use another for the same purpose.
The most favourable situation for the application of this maxim is offered by cases in which suffixes of the same meaning and even of the same form are sought to be applied to the same base. In this context a mnemonic stanza cited by Patañjali twice in his Bhāsya summaries three such situations in Sanskrit.
Saisikān matubarthiyāc chāisiko matubarthikah. Sarūpah pratyayo nestah sannantān na san isyate.
Patañjali makes use of this maxim while dealing with two out of these three situations, the verbal suffix san and the possessive suffixes. The saisika suffixes do not appear to have attracted his attention. Kaiyata gives the following examples for them.

Page #294

Uķtārthānām Aprayogah
285
Sālāyāṁ bhavah sālīyo ghatah tatra bhavam udakam
iti punaścho na bhavati. virūpastu bhavati. P. iii. 1. 7 enjoins the suffix san to express desire after a verb, where the action expressed by the verb constitutes the object of the desire and the subject remains identical. A question is raised whether we can add this suffix to a verb which is already having it at the end. For example can cikīrsa- as a base take another san after it, to express the idea cikīrsitum icchati ? The answer is no, because the maxim states that this cannot be done. But an exception is found in case of an expression like eșitum icchati giving rise to eşisisati, obviously a rare form and a rare situation. However the difference between the two examples is clear. While in the permissible case, one of the desires is expressed by a root-morpheme, in the prohibited case, both are intended to be expressed by suffixal morphemes. This is indicated by the condition sarūpa having the same form (i. e. status).
On P. v. 2. 94, it is pointed out that one possessive suffix cannot follow another. The possessive suffixes form a group and include in, mat (~vat), maya and others. Naturally their meanings also show differences in the nuances and hence one can find two suffixes from them which can follow one another. Patañjali gives examples like dandimati śālā (room having men carrying staves), hastimati upatyakā [foot of a mountain having elephants (i. e. animals having trunks)] which can be easily understood. Then he takes another situation where the maxim applies. A person possesses men who in turn possess cows. But we cannot get a form like gomanman because the suffix is the same with the same meaning and hence barred by this maxim. The maxim is applicable if it satisfies the conditions of sarūpatva and samānavrttitva.
Another example is furnished by the rule of agreement between a noun and its adjective as regards number, gender and case which is so widespread in the IE classical languages. Patañjali takes up such a case in dealing with P. iv. 1. 3 striyām, in which there is an agreement in gender. His example is gārgyāyaṇī kārīsagandhā kālitarā where the referent is a single woman. Various reasons are given by Patañjali to avoid the use of this maxim by denying that all three express feminine, that the gender is expressed only once and the others are due to specific injunctions without semantic contents, or that the feminine concept, being generic in nature is realised differently in different individuals and conditions. Modern linguistics has added another solution by postulating a long component by making the morpheme

Page #295

286
Amrita
discontinuous. Agreement is not a universal feature and this type of redundancy in language serves the purpose of clarity and ease of understanding more than any semantic purpose.
In P. iii. 1. 12 the suffix kyan is laid down after a number of words forming a gana called bhrśādi to form denominative verbs. The meaning of the suffix is 'to become and is used if a Cvi suffix is not used, and the final consonant of the word is dropped. The Vārttika defines the meaning more precisely by calling it abhūtatadbhāva, when a thing is treated as something else which it is not. The group of words contain simple words as well as words preceded by prādis as in sumanas, durmanas etc. A complicated and subtle argument leads to the conclusion that the meanings of the prādis can qualify and modify the meanings of the listed words or may qualify the meaning of the suffix itself, in spite of the fact that they are listed before the words.
If we decide to take them as qualifying the suffix kyan, their meanings will also be expressed by the suffix and hence they cannot be used in the denominative formation because of this maxim.
Patañjali's answer depends on the tail that many suffixes are used with the same noun as in the case of manas and the single form manāyate cannot express different meanings if they are not used. Therefore instead of postulating a large number of homophonous suffixes with different meanings, it is better to express the difference by using them before the nominal base. Avoidance of ambiguity is thus a reason for setting aside the maxim.
Patañjali appears to make use of this maxim without any restriction in his comment on P. ii. 1. 36 caturthi tadarthārthabalihitasukharaksitaih which describes a Dative Tatpurusa. Tradition splits the long compound into six elements viz. tadartha 'meant for the meaning of the word in the dative' and the words artha, bali, hita, sukha and rakṣita. Both Kātyāyana and Patañjali agree that the scope of tādarthya must be restricted to the relation of prakrti (the material) and vikrti (its modified form) in order to make the inclusion of the remaining words in the sūtra significant. This restriction then excludes a compound like aśvaghāsa 'fodder for a horse' which can be also taken as a genitive compound.
The word artha as a second member of the compound gives trouble. Hence Kātyāyana formulates two more rules one of which makes it a nityasamāsa preventing its resolution and the other assigns to it all genders thus making it an adjective. This will explain compounds like brāhmanārtham payah and brāhmanārthah sūpah as well. To avoid these rules Patañjali makes

Page #296

287
Uktärthānām Aprayogaḥ
a number of suggestions most of which are too artificial to be acceptable. The last one is to make artha a substitute of the second member of such compounds. This approach proves to be not tenable and therefore Patanjali follows another approach and argues that because the meaning of the Dative is to express purpose, a word like artha cannot be used with it because of this maxim. As the expression is not available, no compound can be thought of. But the explicit statement of Panini makes it possible. No difficulty on account of gender can be raised because gender in Sanskrit is, after all, arbitrary and based on usage.
The last use which Patanjali makes of this maxim is found in the course of a long discussion on P. ii. 1. 1. samarthaḥ padavidhiḥ. While discussing sämarthya of words having syntactic congruence, it is argued that synonyms like sakra, indra etc. cannot be used to form a karmadharaya compound by putting them in apposition. The reason is, when the first word conveys the intended meaning, a second word of the same sense cannot be used. In this context a distinction is made between perfect synonyms and approximate ones. In the later case a compound is possible as seen in bhṛtyabharaṇiya.
The maxim uktärthänäm aprayogah is nothing but a strict application of the principle of economy applied to language. It stands in glaring contrast with the other characteristic of language called redundancy, which is considered as essential for efficient transmission of a message in any semiotic system including language. An adjustment between these two principles is thus a necessity of a language and the grammarian has to take note of this situation.
If we start with the theory that language consists of two articulations, one based on the other, it is obvious that redundancy plays an important role in the first or lower articulation and the principle of economy is subordinated to it. Hence the present maxim has no relevance for the phonology of the language. But when we come to the second articulation in which meaning plays a prominent part and gets associated with the sounds of language, both principles have a vital role of play. The inevitable adjustment is expected to be different in different situations. Hence it is worthwhile to scrutinize the application of this maxim at different levels in language and in different types of morphemes. This approach is implicitly present in the treatment of this principle at the hands of Patañjali, which needs to be explicitly formulated.
Acceptance or rejection of such a principle in toto has no significance. A valid approach is to ascertain where it functions and where it does not.

Page #297

288
Amrita
.'
Patañjali's use of the maxim is casual and is applied to specific cases. We have to extract from his practice the way in which he would use it in relation to classes of linguistic elements of Sanskrit in order to generalise its nature.
By a close analysis of the Mahābhāsya on the twelve sūtras of Pāṇini discussed above we can summarise the use of this maxim as follows :
The principle is not applied in the following situations : (1) the concept of number expressed by a numeral and by the grammatical category like singular, dual or plural, (2) when two words express a generic and a specific concept of the same nature, (3) to the law of agreement if it exists in a language, (4) if a distinction of meaning cannot be otherwise expressed in a language, (5) when an emphasis is to be placed on one of the two occurrences of the same idea. Here we can also include what can be called as a pregnant use of a word as in the famous line ravikiranānugrhītāni bhavanti kamalani kamalāni, which is often discussed by Sanskrit rhetoricians.
The maxim is allowed to function in the following situations : (1) if the concept is expressed by a case and a derivative suffix, (2) the use of a root when it expresses an action which is considered as having an abstract unity, (3) different suffixes having the same grammatical meaning, (4) two uses of a suffix like san with verbs, (5) two possessive suffixes having the same form and meaning, 6) a case suffix and a word expressing the meaning of that case, (7) absolute synonyms.
Looked at from the point of meanings involved, it may be observed that the maxim is applied to words having lexical meanings with stringent conditions like absolute synonymity or unity of an abstract concept which cannot be split into parts without destroying its unity. If forced to occur together, such words tend to develop some difference of meaning or some overtones. As applied to grammatical meaning, the maxim is generally applicable, and a difference in the shape of such morphemes plays only a secondary part. It is generally observed that the development of a semantic difference in a grammatical category is rare and hard to conceive. When a lexical item comes in contact with a grammatical element, the rule is not generally applicable. It holds good among different types of grammatical categories.
We may conclude by saying that in spite of a difference in the methodology used and many decisions taken on the basis of expediency in polemic arguments, Patañjali's use of this maxim stems from a good grasp of the nature of language in general and of Sanskrit in particular. 22. Uktārthānām Aprayogah, (Dr. R.N. Dandekar Felicitation Vol.)8

Page #298

Pronunciation of Sanskrit
CLUSTERS WITH H
Pănini VIII. 3. 23-7 deals with the changes of final m in different situations and incidently treats of n as well. Thus we are told that the final. m of a word, when followed by any consonant (hal), changes into an anusvāra (23), but remains m before a vowel. Both m and n when not finals of a word also become anusvära when any consonant other than a nasal or a semi-vowel (jhal) follows (24). An exception to the rule 23 is noted in Sūtra 25 to explain forms like samrāt, sāmrājyam, etc. Then follow two Sūtras which state that m followed by a group like h+m optionally remains m or becomes anusvāra (26) and followed by h+n optionally becomes n or anusvāra (27). A värttika on 26 adds that m can also optionally become y, , and I when followed by h+y, v, l.
The examples given for the last two rules in Kaśikā may be stated as follows:
kimhmalayati kim hmalayati or kim hmalayati kimhyah kim hyah or kiỹ hyaḥ
kim hvalayati=kim hvalayati or kiv hvalayati kim + hladayati kim hladayati or kil hladayati
kim hnute=kim hnute or kin hnute
If we look into the distribution of the anusvara in the language we find that while it is liable to be assimilated to the following stops as regards the place of articulation and can become nasalized y, v, I before these sounds, it keeps its distinctive nature before others, namely the sibilants (ś, ş, s), r and h, which is thus its exclusive sphere of occurrence. Obviously Sanskrit did not admit nasalized forms of these sounds. The anusvära thus

Page #299

290
Amrita
shows only a twofold treatment before consonantal sounds. Either it is optionally assimilated to the following consonant if a nasalized articulation of that place is possible in the language or remains an anusvāra, if no such nasalized pronunciation is possible.
If we now look into the two Sūtra-s of Pāṇini, 26 and 27, and the vārttika bearing on them, we find that the same twofold treatment is operative in these cases as well, though the matter is masked to some extent by the mode of writing. Thus when the following word begins with h in turn followed by m, n, y, v and I, we get a set of forms in which the final m of the preceding word becomes an anusvāra and we have cases like kim hmalayati, kim hyah, kim hvalayati, kim hladayati and kim hnute. But when the final m of a word has become a nasal stop like mor n or a nasalized semi-vowel like ĝ, or 1, we expect the following word to begin with a '' consonant which admits of a nasal articulation at the same place, i.e. m, n, y, v, and l. In other words the second set of forms should sound kim mhalayati, kin nhute, kiy yhah, ki v vhalayati and kil thadayati. This means that consonant-clusters written as hm, hn-, hy, hv- and hl- were pronounced in two ways : as h followed by m, n, etc., or as m, n, etc.,, followed by h. We know that the second type of pronunciation became more and more frequent in the later stages of the Indo-Aryan languages and it is the usual one in most New Indo-Aryan languages which show such clusters. If we regard the vārttika as marking a later stage in the growth of Sanskrit than the Sūtra-s of Pāṇini we may conclude that the tendency began with clusters of h with m and n and later on spread to clusters of h with y, v, and l.
If we again look at the wording of the two Sūtra-s of Pānini we note that his expressions meaning makāraparahakāra and nakāraparahakāra refer more to the script of these clusters than to the alternative pronunciations which are implied by his rules. It is possible to avoid this result by taking his words mapara and napara in two different ways, when his followed by m and n' and 'when h comes after m and n', and thus argue that he intended both types of pronunciations in these cases. But we note that Jinendrabuddhi expressly denies the alternative 'h coming after m or n' and such an explanation looks too artificial to be really intended by the Sūtrakāra. This fact, thus, has some bearing on the use of the written form of a word in the description of the Sanskrit language by Pāṇini.
25. Pronunciation of Sanskrit: Clusters with H
The Adyar Library Bulletin, Vol. XXV

Page #300

Traces of Short ĕ and Ŏ in Ṛgveda
The problem of the short values of ĕ and Ŏ suspected of being found in RV. is intimately connected with that of the abhinihita sandhi, and both must be treated together. For a long time, the facts of this sandhi in RV. are well-known, and scholars have come to the conclusion that, in spite of the writing of this sandhi in the traditional text, it was nearly non-existent in the oldest period of the language, and came into vogue only gradually so as to become finally the usual practice in the Classical Sanskrit. The RV. Samhitā reveals an apparently conflicting procedure of sometimes eliding or merging the initial a in the preceding -e and -o, and more often keeping them apart with a hiatus between the two. No general principle for this divergent treatment is easily perceptible and the Ṛk-Prātiśākhya finds it necessary to formulate a few general rules based on empirical observation and then to enumerate exceptions to them as found in the Samhita. The rules of the Prätisakhya' may be summarised in brief. It lays down that the abhinihita sandhi is the rule between the Pädas of a stanza, which are treated as forming a unit in the Samhita. This rule is observed in the traditional text, except for 3 cases, which are duly noted by the Prätiśäkhya. In the interior of a Pāda, the sandhi occurs under definite conditions, otherwise absence of sandhi is the usual practice. The sandhi is observed in the following cases: (i) when the initial a- is followed by a light syllable beginning with y- or v; (ii) when the first word ends in -avo and the following word begins with afollowed by any light syllable (iii) after words ending in aye, ave, -ayo, - avo with initial a- irrespective of the nature of the following syllable (to this rule the Pratiśākhya notes 12 exceptions); (iv) after vo, preceded by the words a, na, pra, kva, citra, suvitä, eva or ka and followed by the initial a. After these general rules the Pratiśākhya notes the cases of 16 words, the initial a of which is elided; 10 words after which the a- is lost and finally some 62 cases of specific combinations in which a- is elided2.

Page #301

292
Amrita
Modern grammarians have also pointed out the lack of uniformity in the orthoepy of RV. and the predominant absence of this sandhi in the earlier phase of the language. Thus Whitney has noted that out of 4500 cases in RV. only about 70 show this loss of a- actually occurring in the pronunciation, while the writing shows it in three-fourth of the cases. From this he concludes that there is 'no accordance in respect to the combination in question between the written and spoken form of the text. The cases of its real occurrence are more in AV., and the sandhi gains in practice.
Oldenberg“ has discussed the problem in detail and has arrived at some definite conclusions, which are generally accepted by later grammarians. By an analysis of such cases as occur in the second part of the Xth Mandala (from 90 to end), he points out that in 19 cases the traditional text shows the Sandhi correctly, and in 211 cases the textual tradition is equally correct in not showing this sandhi and preserving the hiatus. There is only a single case (X. 108.5) where the sandhi is not shown even when metre demands it, while in 22 cases, the sandhi is written though the initial a- must be pronounced, as far as the metrical evidence suggests. From this he concludes that the Vedic tradition of expressing this sandhi is generally correct and is therefore genuine. The cases, where the incorrect writing of the sandhi is found, he explains as due to the working of the redactors, who were following the rules, which are also stated in the Prātiśākhya. These rules like the loss of a- when followed by y- or v- were the guiding principles of the redactors, but of the origin and reasons of these rules, he is sceptical. He frankly states that he has 'not succeeded in explaining the origin of these principles and he calls them 'principles which are arbitrarily invented'. According to him, some such principles were either current or invented, which were used by the redactors in remodelling the text of the RV., and which were further deduced from it by the Prātiśākhya. This explains the remarkable agreement between the text of the RV. and the rules of the Prātiśākhya.
Later writers have added little to the conclusions of Oldenberg. Thus Wackernagels states that according to the evidence of metre, the final and initial vowels form two distinct syllables in 99 per cent cases of RV. and in about four-fifth of the cases of AV. and the metrical parts of YV. Sporadic cases occur in the old prose and the verses of the Brahmanic period. He further makes the important observation that the final syllable of the first word is metrically short. Arnold has pointed out that 'the combination of the two is rare in RV. proper, but is occasionally found in all parts of it; in the

Page #302

Traces of Short ě and o in Rgveda
293
popular RV. it is considerably more common. Moreover, as he finds the loss of a- even after the caesura, he thinks that the poets did not regard the loss as equivalent to combination of the syllables. Macdonell' has simply summarised the statements of Wackernagel, and in his later work A Vedic Grammar for Students, has included the short values of ě and o in the list of metrical necessities. Lastly Bloomfield and Edgerton' have discussed the cases of the abhinihita sandhi, as far as they are revealed in the variants of the Veda. They consider that the redaction of the Vedic texts was carried out when the rule of elision was established in the language. But in view of the metrical necessity of the Veda, the redactors did not elide the a- where the pronunciation demanded it. In prose and at the end of a Pāda, the elision was always marked, probably because no confusion of pronunciation would arise, as a- must be read in all these cases. From the consideration of cases in which the writing is consistent with the actual pronunciation and those in which it disagrees, they have found out that the writing is consistent with the actual pronunciation in two-thirds of the examples which proves that 'an attempt was made to conform spelling to original pronunciation of the metrical texts', or more cautiously ‘an attempt was made to make the writing conform to certain definite standards, which had some relation to actual pronunciation.
As regards the actual pronunciation of the final -e and -o followed by Q- which remains metrically unelided, there is some amount of vagueness in the current opinion of the Vedic scholars. From the very beginning it has been noted that the final vowels in such cases demand a metrically short value, though the exact sound which they possessed is differently assessed. Bloomfieldo discussed the value of these -e and -o before a- and gave them the short values of ě and ->. His explanation of the fact is based on the supposition that in these particular circumstances, there survived the original short values of -ě and ->, which normally became -a in the Aryan Branch. The original endings -ěs, -os lost their final -s before the following vowel, and the new final vowels -ě and -> had a two-fold development. Before vowels other than -a they lost their vowel-quality, but preserved their quantity and became -Q. When they were followed by a- they kept their quality but disregarding their quantity were written as -ě and -7; for the Indian alphabet had no signs for short -ě and -ě. These sounds later attracted others, which were really long 7 and 7 as coming from a+i and a+u, to themselves, first before the following a- and then before other vowels as well. Finally the original e was absorbed by the more frequent o and there remained three vowelso (out of

Page #303

294
Amrita
-ěs, -os) o and e (of original diphthongal nature) all written as long but pronounced short before a..
Oldenberg has advanced very cogent arguments to reject this suggestion of Bloomfield. He points out that besides the improbability of preserving the original short values of -ě and -o in such circumstances, the theory make too great a use of graphic confusion at this early period of language to make it probable. After a close examination of the facts of the case, he himself has arrived at a different conclusion. He starts with the undoubted supposition that the metrical value of the final -e and -o in all such cases, is short, and proceeds to point out that in case of original long ē, 7 the sandhi was a-y, a-v before vowels other than a., and as in the case of the long diphthongs, the hiatus was left after the loss of the final y, v. Nothing is more natural than to suppose that the same sandhi was originally observed even before a- thus giving rise to -a a- and -a a- and later on for -(from -as) a- to -as a -as well. This will explain the short value of the final syllable in all cases where both syllables are kept apart. Now remains the question of explaining the use of 7 and 7 in place of -ay, -av and -as, when they are followed by a- but not when other vowels follow. Here Oldenberg's argument gets a little confused, but he appears to suggest that the abhinihita sandhi, which is found in a few cases in the RV. itself, must have influenced these cases also. The crasis of -e+a- etc. over -ay+a- etc. is proportionately as frequent as the other of -e+i- or -0+u- over -ay-ti and -as+u-, and the presence of y and v may have coloured the quality of the preceding vowel resulting into -e- and -o- with a double crasis. This usage in turn, must have affected the other group where both the vowels are kept. Why the redactors preserved the two syllables here and did not do so in many other cases like ata or i+a, cannot be made out. It is thus clear that Oldenberg regards that the phonetic value of the finals in the cases showing absence of the abhinihita sandhi was only -a, which was originally followed by a sound corresponding to the spiritus lenis in cases of -ay and -av and some kind of glottal stop in case of -as, which prevented further crasis.
Wackernagel has virtually accepted the phonetic value -a a- assigned to them by Oldenberg, but he differs from him in two minor points. Against Oldenberg, he explains the writing of e and o for the regular a by supposing that the form before the consonants was used before a as well, and it soondeveloped into the classical sandhi, which later influenced the writing of the Veda. He points out that the reduced coefficients y, v of the original diphthongs ē, 7 could not give the timbre of e and o to the original a, as this

Page #304

Traces of Short ě and o in Rgveda
295
explanation fails to satisfy the case of as from -ah. In other words, the abhinihita sandhi is no continuation of the original sandhi of e and o before a
Though Oldenberg believed that the original value of the finals in such cases was -a a., in which Wackernagel agrees with him, he usually used a simplified notation of writing them as e' and o' in order to indicate their short value. This has led Macdonell to commit a strange mistake. Accepting the explanation of Oldenberg that the sandhi of final e and o before a- was the same as before other vowels, he yet regards the syllables to be pronounced as ě and o, which is thus explicitly stated by him in his Vedic Grammar for Students. This is obviously self-contradictory. Bloomfield and Edgerton have nothing to say about the phonetic value of the finals and they simply point to the discussion of Wackernagel, thus presumably accepting his views.
The examination of the relevant facts however reveals a different principle at the basis of all these cases, which has somehow escaped the notice of the earlier scholars. We can begin our investigation, in the first place, with material supplied by the RK-Prātiśākhya and then an attempt may be made to control the results by an analysis of one or two Mandalas of RV. In case of deciding the peculiarities of a text like RV. where no external control is available, it is of some importance that an investigation carried on with a part of the text be controlled by a similar analysis of another part of same text. Besides using the cases collected by the Prātiśākhya, Oldenberg and Arnold, I have utilised the III and IX Mandalas of RV. fully.
A fruitful approach to the problem would be to determine above all the possible cause of the two-fold divergent procedure followed by the Rgvedic poets themselves and later on by the redactors in dealing with the cases of the final e and o followed by a-, which Oldenberg remarks, he cannot find out. Nor has the Rgveda Prātiśākhya given us any definite rule suggesting the reason of this two-fold treatment of either preserving or eliding the initial a- in such cases. Generally the Rgvedic poets have kept the two syllables apart and intact, and we can reasonably suppose that they absorbed the initial a- in the preceding vowel, whenever the metre demands it. The redactors have also followed the same procedure. But we also find numerous cases of discrepancy between the writing and the actual pronunciation. One of the most striking cases of such discrepant treatment is found in the observance of the abhinihita sandhi between the ends of the Pādas, which form one unit in the redacted text. This rule is invariably observed in the RV.

Page #305

296
Amrita
.
and is duly noted by the Prātiśākhya. Oldenberg simply leaves aside all these cases, calling them as merely a whimsical activity of the redactors and then comes to the plausible conclusion that in the majority of the cases of the latter half of the X Mandala, the writing shows a fair approximation to the actual pronunciation. The writers of the Vedic Variants suggest that the sandhi was indicated here, as there was no possibility of confusion in the pronunciation. Obviously both the explanations are not satisfactory and the attempt to bring together the pronunciation and the writing in agreement, by neglecting this majority of cases of the discrepancy, cannot be acceptable. In the I Mandala there are 60 cases of this sandhi in the interior of a pāda, while there are as many as 108 cases at the Pāda-end. In the III Mandala the cases are 12 in the middle of the Pāda' and 40 at the end. In the IX Mandala 2 the cases are 14 to 68. Even among the cases where the loss of a is indicated in the middle of the Pāda, we find in the I M. the initial a. read in 41 cases as against 19 where it is lost in pronunciation 13. The figures for the III M. are 11 where a- is pronounced as against 1 where it is lost in pronunciation as well : for the IX M. q- is pronounced in 11 cases, while it is silent in three cases. As against this we may set the figures of Oldenberg for Mandala X 90- end; where in 19 cases a- is correctly lost, while in 22 it must be pronounced in spite of the writing. The explanation of this difference can only be found in the fact that in the X Mandala this sandhi had come into greater vogue.
If it is admitted that the guiding principle in these cases, at least at the beginning, was not the actual pronunciation, we must find it in the nature of the final vowels e and o, and naturally in their length. It is quite obvious that these vowels at the end of the Pada must necessarily be long, and thus require the elision of the following a- irrespective of the fact that it is to be pronounced or not. In other words, the original principle which guided the observance of this sandhi in some cases and not in others must have been the quantity of the final e and o, whether long or short. As their value is uniformly long at the end of the Pāda, the elision of the a- in such cases is naturally the rule, and the question whether a- in these cases is to be pronounced or not is immaterial. If this is true, we should expect no exceptions to it, and except for three isolated cases already noticed by the Prātiśākhya, the rule is uniformly observed. Even the three exceptional cases - show some peculiar features which explain the absence of the sandhi. The cases are :
.

Page #306

Traces of Short ĕ and ŏ in Ṛgveda
297
VII. 103. 3 cd. na putro. anyo anyamupa vadantameti Tristubh
VIII. 9. 15 ab. parake. arväke asti bheṣajam Brhati
IX. 39. 5 ab. paravato. atho arvavataḥ sutaḥ Gayatri
In all these cases, it is easy to see that immediately after the Päda end, an exactly similar case follows, where the absence of Sandhi is justified, and this must have influenced the redactors in not setting the Sandhi earlier.
Taking our suggestion from these numerous cases of the abhinihita sandhi, which were so long regarded as purely arbitrary and of no account, we may proceed to investigate others where the sandhi is justified by the metre. It is reasonable to suppose that the Vedic poets followed this sandhi in all those cases where metre requires it, though it does not follow that it was not used in all those cases where a is required to be pronounced metrically, for we have just noted that this sandhi may be one way of indicating the long value of the preceding vowels and not merely an indication of absorption of the following a-. But wherever the poets themselves have followed this sandhi, it is obvious that the metrical value of the final e and o must be long. Now, it is well-known that, though we are not in a position to decide the metrical length of all syllables of the Vedic metres, there is a fair chance of doing it in the case of the last few syllables forming the cadence, and in other places the surrounding syllables at least indicate the preference of the Vedic poets for short or long syllables. Thus to take the most frequent types, the 6th of the octosyllabic lines and the 8th and the 10th of the eleven and twelve syllabled lines must be long. The last syllable of all these lines is anceps. Other less obvious cases of long syllables are the second and the third of all these lines when the third and the second are found to be short.
According to the enumeration of Arnold", there are 84 cases in RV. where the initial a- is metrically lost after the final e and o. There are actually 89 such cases, out of which 13 show the written a though the metre requires its elision, but a few cases remain doubtful. The remaining 76 cases can be distributed as follows, according to the syllable in which the final e and o occur.
I. In the eight-syllable lines of metres like Gayatrī, Anuṣṭubh, Pānkti, Bṛhati and Satobṛhati

Page #307

298
Amrita
Remark
Long
Doubtful
Syllable 2nd
No. 4
3rd
4
2 with 3rd syllable
short U - V -) 2 in later Anustubh
(--- with the 2nd syllable
short (-u-u before the break
GU--) (U-U-)
4th 5th 6th
1
II. In the lines of even syllables mostly Trişțubh:
2nd
6
3rd
4th
6 with third short 1 with second short followed by caesura
followed by caesura with an early caesura
5th 7th 8th 9th
X. 49. 5. c.
10th
11
48
III. In lines of twelve syllables, mostly Jagati : 1st
(---U) 2nd
third short (U - U -) 7th
with early caesura
8th
10th
Total
76

Page #308

Traces of Short ĕ and ŏ in Ṛgveda
This analysis of the cases where the abhinihita sandhi is observed and metrically attested leaves no doubt that in all such cases the final e and of were long and metrically felt as such. It thus confirms the suggestion to which we were led by the observation of the sandhi at the end of the Päda.
299
The 13 cases in which the a- is written and yet, according to Arnold, metre requires its elision, are difficult to decide. They are:
I. 70. 4b, visām na visvo amrtaḥ svadhiḥ| Dvipadā Viraj. The no. of syllables of 10 may have influenced the redactors.
I. 190. 3d. mrgo na bhimo arakṣasastuviṣmân | Tristubh. There may be confusion of the caesura after the 4th and 5th syllable.
I. 53. 2c, sikṣānaraḥ pradivo akāmakarśanaḥ | Jagati.
I. 89. 6c. svasti nas tärkṣyo aristanemiḥ | Virātsthänä. The metrical elision may not be intended as can be seen by comparison with the next Pāda svasti na bṛhaspatir dadhātu |
I. 186, 11a. iyam sa vo asme didhitir yajatrā | Triṣṭubh. It is possible to take the first two short syllables as one long.
III. 59. 2d, nainam amho aśnotyantito na durat | Tristubh.
VII. 61. 3d. ṛdhag yato animiṣam rakṣamānā | Tristubh.
VIII. 50. 5b, iyano atyo na tosate | Bṛhati
VIII. 58. 1c. yo anūçāno brāhmaṇo yukta āsīt | Tristubh.
IX. 9. 4b. nadyo ajinvad adruhan | Gayatri. The loss is doubtfull.
X. 72. 4c. aditer dakṣo ajayata | Anustubh. More naturally the first two syllables must count as one.
X. 89. 13d. anvapo ajihata jāyamānam | Tristubh. The number of the syllables may have influenced the writing.
X. 108. 5b. pari divo antän subhage patanti | Tristubh.
Many of these cases are metrically defective, but as far as our present problem is concerned, in no case do we find the final e and o occupying a place where a long syllable is necessary and in all cases the syllable can well be short.
We may now proceed to observe the practice of the redactors of RV. as it is revealed in the cases where the loss of a- is not justified by metre.

Page #309

300
Amrita
It is quite apparent that the Prātiśākhya has followed no principle on which to explain the numerous cases where the a is elided in the interior of the Pāda. Yet it is obvious that the majority of the cases of this elision were collected by the writer of the Prātiśākhya and brought under some empirical rules, which, however, show no inner connection and are mostly descriptive. A verification of the rules of the Prātiśākhya with the actual cases in RV. gives the following results, which go to show that the description of the Prātiśākhya is accurate and fairly exhaustive. (i) Under the rule, where a- is followed by y and v of a light syllable we may include cases like I. 7. 4a; I. 17. 2a; I. 32. 15a; I. 45. 5d; I. 48. 14b; I. 51. 3a; I. 77. 4b; I. 81. 8c; I. 85. 10a etc. with a total of 130 cases. (ii) Under the rule of -avo followed by a light a- followed by any light syllable, come cases like I. 92. 1d; I. 9c; I. 163. 8b; V. 30. 10a; X. 12. 4c = in all 5 cases. (iii) The rule about the endings -aye -ave -ayo -avo includes cases like I. 24. 8c; I. 47. 8a; I. 51. 6b; I. 103. 7b; I. 104. 1c; I. 121. 13d; I. 122. la; I. 126. 2d; I. 164. 5c; I. 186. 7a; II. 23. 16b etc. = 67 cases in all (iv) The rule about vo preceded by a, pra, etc. explains cases like L. 107. 1c; I. 110. 3a; I. 168. 5a; III. 57. 2d; IV. 34. 3c; IV. 34. 11a; IV. 55. 1d; V. 54. 10c; V. 61. 2a; X. 32. 5a; X. 63. 6c; X. 76. 5a; X. 166. 3a; X. 166. 4d; = 14 in all. (v) The rule about the initial a- of words like adāt, arvattah, ajanayanta etc. covers 28 cases. (vi) The loss of a- after words like vāsovāyo, abhibhuve etc. covers 9 cases. (vii) Finally the Prātiśākhya enumerates 62 cases of specific combinations in RV. The total of all these cases thus explained by the Prātiśākhya comes to 315 in all, leaving behind a small number unaccounted for. .
This way of analysis and classification, though accounting for the majority of the cases cannot be accepted as representing the principles of either the poets themselves or of the redactors of the Samhitā. Not only does it fail to explain cases like I. 168. 1c. ā no rvācah; III. 60. 6b no' smin; X. 90. 4b pădo syehābhavat etc. but in some cases the loss falls under two distinct rules as in V. 31. 5c pavayo rāthāh (rules 3 and 5); VI.47. 24c payave' dāt (rules 3 and 5) etc. Moreover we meet here with the formulation of general rules which are so in mere appearance, as they are applicable to single cases only, for instance with the words arvattah, avyatyai, aviratā, hiranyaśrngo etc.
If we now try to examine the cases where a- is elided in the interior of the Pāda in RV. but where metre requires its pronunciation and classify them according to the metrical value of the finale and o, we get the following results.

Page #310

Traces of Short ě and ò in Rgveda
301
* 1. In the lines of eight syllables :
Syllable
No.
Remark
Long
Doubtful
2nd 6
7
5
2
3rd"?
5 have the third short 10 have the 4th short 1 has the second short
anceps cf. the note below
111
4th 18
3
5th
6th 19
i 45
42
Note :- Apparently the 18 cases in the fifth syllable go against the rule that the 5th should be short. But a careful consideration of these cases reveals an important fact. Of these, 16 have the scheme (- (U) -) : I. 17. 2a; III. 13. 2d : V. 22. 3c; V. 35. la; V. 65. 3a; VI. 14. 13a; VIII. 11. 6a; VIII. 12. 19a; VIII. 23. 21b; VIH. 27. 13a; VIII. 27.15d; VIII. 30. 3a; VIII. 33. 6a; VIII. 33. 10b; VIII. 53. 7a; VIII. 91. 7a. Two have the scheme (- (-) U -); I. 37.13b; V. 6a where the elided a- is long by position. In actual writing, of course, the foot gives the appearance of (-U - ) which is the normal ending of the octosyllabic lines. A careful reading of all these lines makes it highly probable that the end was so pronounced that the a- was really absorbed in the preceding long syllable, which thus occupied the place of both the 5th and 6th syllables of the lines.
II. In lines of eleven syllables :
1st20
3
2nd21
3rd 22
6 have the third short 1 has the second short
before caesura before caesura
4th 23
5th24
6th 25 7th 26
8th27

Page #311

302
Amrita
III. In lines of twelve syllables :
1st28
2nd 29
4th 30
before caesura
5th31
before caesura
7th 32
8th 33
9th 34
cf. II. 1. 9c; X. 49. 5c.
51
11
The facts are sufficiently clear and decisive. In the vast majority of the cases, the final e and o before the elided a- is long, while in a few cases the value remains doubtful. It is only in two cases of the 9th syllable that they appear to show a short value, but both the lines are metrically defective. Like the earlier cases where the loss of a- is metrical and hence attributable to the poets themselves, these cases also reveal the fact that. the redactors have followed the same principles of dropping the initial a- after the long ê and 7, thus preserving the memory of their long value in a definite number of cases. It is just possible that in a few individual cases, they may have made mistakes or confused the real values of the finals, which is not surprising in view of the same writing of e and o every where. In any case they never intended to indicate by the loss of a., a loss of that syllable in actual pronunciation.
If the above suggestion is correct, it should be possible to verify it by the large number of cases, where the initial a- is not elided after the final e and o in RV. Naturally we should expect in all such cases a short value for the finale and o, justified by the metre, wherever we are in a position to decide it. Indeed, this fact, viz. the metre normally requires a short value of final e and o before a- was already noted by the Vedic scholars long ago, and they have given it even a wider scope than is justified. According to the current opinion of scholars like Oldenberg, Arnold, Macdonell and others, wherever the a- of the Samhitā is to be read, whether it is written or not, the preceding e or o must be pronounced short, and we have seen above, how this short value is explained by Oldenberg by the supposition that it was really the sound a in all such cases, later changed to e and o. Our discussion

Page #312

Traces of Short ě and o in Rgveda
303
so far reveals that this view, in this extreme form, cannot be maintained. Apart from all the cases at the Pāda-end where q- is read, the cases of the sixth syllable in the octosyllabic lines, and those occurring in the 8th syllable of the lines of eleven and twelve syllables, can never be regarded as showing a short value of e and o. To appreciate fully the value of these 33 cases, it must be remembered that these occur in the cadence which is metrically most rigid and stand out of a total of 35 such cases, leaving aside the 18 cases of the octosyllabic lines for which we have found a special reason for their occurrence. The two exceptions are found in lines which are metrically defective.
But there are ample indications to show that the final e and o were pronounced short before the following a- which was not elided. From the bulk of such cases, I have collected and analysed all such cases from the and IX Mandalas of RV. in order to verify the hypothesis. The cases even in these two books are ample enough to give definite and reliable results and there is no doubt that the results obtained by an analysis of the other Mandalas would prove to be similar. Altogether the tabulation gives the following results : .
I. Lines of eight syllables : Syllable No.
Metrical scheme with e and o as short.
(U.-) 2nd 36
17 = (-U-V) 5 = (-UU.)
26 = (yuy) 1 = (UUU.) 4th 38
38 = (uu.U) 2 = (U-UU) 5th39 95
92 = (U-u) 3 = (U..U)
cf. IX. 18. 1b; IX. 53. 1a; Ix. 98. 3a. 189
1st 35
:
5
3rd 37
Out of these 189 cases 92 have necessarily the short value of e and 0. It is not possible to determine with the same amount of certainty the value of the other cases, though it can be seen that the neighbouring long syllable suggests a preference for a short value of the syllable in question.
II. Lines of eleven syllables :
1st 40
3
2 = (U -- -) 1 = (U-U--)

Page #313

304
Amrita
2nd 41
3rd 42 4th 43
5th44
11 = (- U - -) 3 = (-U ...) (..u(...) UUU - U) caesura after the 4th (U.) 8 = (- UU) 2 = (UU) ( UU --) (-u -)
6th45
7th46 8th47
9th 48
50
195.
Here also the 50 cases of the ninth syllable must show a short value. The 11 cases in the 8th syllable appear to go against the expected short value, but we must note that in all these 11 cases there is the word-ending, in which case a short syllable in the 8th is permissible. In the 15th cases of the 6th syllable, the e ando occur in the second place after the caesura and is preferably short. The shemes are 5 = (-u) 10 = (UU -). The cases in the 7th syllable are not so uniform. Their schemes are : 2 = (UU) 4 = (UU U) and 4 = (UU). Even here the preference for a short syllable is visible. The 65 cases in the 4th and 5th syllables all occurring before the caesura remain doubtful. In the 27 cases of the third syllable, the value is decidedly short as both the 2nd and the 4th syllables are long. The 14 cases of the second syllable show a long third syllable and in view of the prevailing iambic cadence, they should be preferably short. The 3 cases in the first syllable remain indeterminate. III. lines of twelve syllables :
(
UUUU 2nd 50
fu.-) 3rds
(U-U-V) 4th52
1 st 49
5th 53
6th 54 7th55
UU.) 3 = (
UU) 5 = (-UU) (U, U-U) (-u -u -)
8th 56
9th 57
106

Page #314

Traces of Short ě and o in Rgveda
305
Here the picture is essentially the same as for the Tristubh lines. The 23 cases of the 9th syllable are all short : all the five cases in the 8th syllable show the necessary caesura after the word-ending. The five cases of the 6th syllable occur after an early caesura with the schemes : 4 = (U U -) 1 = (- -). The schemes for the cases of the 7th syllable are : 3 = (UU) 3 = (UU) 2 = (
UUU). The 48 cases of the 4th and 5th syllables, all occur before the caesura. The 10 cases of the 3rd syllable have the second long and the 2 of the second syllable have the third long58.
Thus we see that whenever the redactors of RV. have kept the two vowels -e a-, -o a- side by side, the final e and o occur in a place in the metre, where either a short is required or is preferred. We thus feel justified in assuming them to have a short value.
We will have now to discuss more closely the values of final e and o in such cases. The long values of these sounds are well-known and need no justification. But the case of their short values is more difficult. It is probably the unwillingness to admit such short values, which led Oldenberg to suggest his explanation discussed above. The only place in Sanskrit grammar, which mentions the short values of e and o is to be found in the Mahābhāsya of Patañjali on the Sūtras e- o- n and ai- au- C (1. 1. 2). While discussing the necessity of having an indicatory t after these vowels, he points out that according to Satyamugrarāyaniyas of the Chandogas there is pronounced a short e and o in some cases. He illustrates them with examples like : sujāte eśvasünrte; adhvaryo adribhih sutam; sukras te enyat, yajataṁ te enyat : all of which are cases of original e and o followed by a., which latter is represented by short e and o and to which the name ardha is given. In an earlier sentence Patañjali himself says that ardha-ekāra and ardha-okāra mean the short (hrasva) values of these sounds. It is true that finally Patañjali points out that this short value is not found either among the people (loke) or in the other Vedas (anyasmin vede), yet the cases preserved are important as marking an intermediate stage of the abhinihita sandhi. They show that e and o followed by a- developed into -ě+ě and o+ő finally becoming -ě- and -- with the absorption of the following vowel. The stage of development revealed by the RV. cases of -ě a- and -o q- may be regarded as the very first step before the assimilation of the vowels began.
Some evidence of a different type to confirm the short value of o before the following unabsorbed a- may be found in a few curious cases of RV. itself. The locative singular of the word sānu is found in RV. 28 times.

Page #315

306
Amrita
Of these the form sānuni occurs once, while the forms sānavi, sānau and sāno occur 9 times each. Here in all the 9 cases of sāno (followed by avye 8 times and by avyaye once) in the IX Mandala, the Padapātha uniformly gives the original word as sānau, regarding the sandhi as irregular. This use is duly noted by the Prātiśākhya (II. 70) where Uvvata says that sano avye takes the place of the regular sānāvavye. Macdonell59 suggests that the regular form sānavyavye would give rise to a disagreeable sequence of syllables and therefore the endingless form is used, but without the lengthening for the sake of the metre. In all these 9 cases, the o occupies the 9th place of the Tristubh and Jagatī and is metrically short. In all the cases it is followed by a-. Now, if the demands of the metre are to be satisfied for a short syllable, one fails to see how the Guna form in place of the expected Vrddhi form could help, if both o and au are regarded as long. If on the contrary, the tradition, which regarded o followed by a- as metrically short, was still alive, one can readily understand that the use of o before the following a- without its elision would give the required metrical scheme. Other cases like go-agrayā IX. 71, sd; tiro-ahnyam III. 28. 3b; III. 28. 6c, dūre-ante, III. 54. 7a etc. should be judged in the same manner.
It is well-known that the Prākrit dialects show the short values of ě and , mostly before conjunct consonants and sometimes final e and o are regarded as short in the metrical scheme. These facts are noted by writers on Prākrit grammar and metre. Their frequency is much greater in the Apabhramsa language and is probably a part of the general tendency of the shortening observable in that language.
Two conditions circumscribe the short values of ě and 7 in Prākrit. Either they occur before a conjunct consonant as in ottha, chetta etc. or they are found as the final syllables of words like mālāe, dhammāð etc. In the first case, the metrical value of the syllables remains the same, both Sanskrit ostha and kşetra having the same scheme as the Prākrit ottha and chetta. In the second case, it is only under the influence of metre that we can be conscious of the short value of these vowels. It will, therefore, be more correct to say that these final e and o are pronounced short for the convenience of the metre. In both cases, the phonetic variation is not significant, for with the following conjunct the long values of ě and o cannot remain side by side with their short values in Prākrit; and in case of short ě and o at the end, there are no doublets with long ē and 7 with a difference of meaning. In other words, whether short or long e and o form the same phonemes. ..

Page #316

Traces of Short ě and o in Rgveda
307
• If this is their real nature, it is easy to bring them in line with other cases of short and long vowels like a ori, which are found in the RV. Samhitā. It is well-known that these vowels are often short or long in RV. and the variation is rhythmic and metrical but never significant. This difference is indicated in writing as well, because these vowels as short and long also form different phonemes in the language. Final e and o must have been subject to similar variation in the Vedic language and this fact is traditionally indicated by the presence or absence of the absorption of the following a-. This scope of insignificant variation of a purely metrical nature is characteristic of the poetic language and can be seen in the poetry of many New Indo-Aryan languages also.
Annotations : 1. II. 34-50. 2. Cf. Max Muller SBE. XXXII pp xlviii-1. 3. Sanskrit Grammar 135. 4. Hymnen des Rgveda : Prolegomena p. 389 foll. 5. Altindische Grammatik I. 272b. 6. Vedic Metre 127 7. Vedic Grammar 72. 8. p. 437. 9. Vedic Variants II. 890-912. 10. JAP. III. 11. III. 4. 10a; 13. 2d; 16. 5a; 19. 5c; 20. 3b; 29. 3d; 29. 16b; 30. 16a; 33. 6c; 33.
7d; 57. 2d; 60. 6b. 12. IX. 7. 3a; 71. 9c; 73. 4a; 74. la; 74. 6a; 86. 11d; 86. 27b; 91. 2d; 96. 4a; 97.
20a; 97. 33a; 97. 41b; 107. 2a; 107. 6c. 13. I. 24. 8c; 30. 16d; 33. 13b; 51. 3a; 51. 5a; 52. 9d; 59. 2c; 59. 3b; 79. 11a; 81.
1c; 85. 7a; 88. 6a; 103. 7b; 118. 7a; 122. 1a; 162. 7a; 167. 2a; 168. 9c; 186.
8c.. 14. Vedic Metre 127. 15. [ Annals, B. O. R. I. ] 16. V. 35. 3a; V. 61. 2a; VI. 48. 12b; VIII. 2. 36c; VIII. 30. 4d; X. 102. 1b X. 171.
2b.

Page #317

308
Amrita
17. I. 45. 5d; 48. 14b; 81. 8c; 91. 9c; VI. 54.3b; VII. 96. 5c; VIII. 96. 5c; VIII. 8.
6b 8. 20d; 27. 4; 79. 3b; X. 25. 4b; 190. lb. 18. I. 105. 17a; III. 16. 5a; X. 26. 6c. 19. I. 7. 4a; VI. 47. 24c; X. 22. 7c; 102. 1d; 185. la. 20. I. 88. 2a; X. 109. la; 181. 3a. 21. I. 107. 1c; II. 12. 8B; III. 33. 6c; 57. 2d; IV. 34. 3c; VII. 35. 13b; X. 15. id;
77. 8c; 99. 8d. 22. VI. 26. id; 50. 14a. 23. I. 32. 15a; 71. 9a; 77. 4b; 112. 24c; 121. 13d; 126. 2d; 163. 8b; 181. 3a; 190.
7a; III. 4. 10a; 33. 7d; IV. 2. 5a; 3.8c; 4. 12b; 13. 5d; 14. 5d; 20. 2b;21. 10d; 24. 4c; 25. 1c; 25. 6d; 25. 86; 25. 8a; 43. 6b; V. 41. 5b; 57. 7d VI. 4. 8a; 9. 7c; 9. 7d; 17. 5b; 21. 6a; 21. 3a; 23. 26; 26. 1c; 36. 1b; 63. 7a; 66.4b; VII. 1.19a; 18. 14a; 24.4b; 33. 2d; 36. 8c; 38. 6c; 38. 7c 38. 8a; 53. 3a; 98. la; 104. 12d; IX. 91. 2d; 96. 4a; X.28. 11c; 30. 3a; 34. 5b; 61. 7a; 95. 5b; 95. 5c; 95.
6c; 99. 4a; 101. 1d; 177. 2b. 24. I. 104. lc; 163. 9a; 164. 5c; 180. 4b; 183. 5b; II. 12. 6c; 28. 6b; III. 19. 5c;
20. 3b; 29. 16b; 30. 16a; IV. 1. 4b; 1. 5a; 16. 20d; 17. 16d; V: 30. 10a 76. 2c: 83. 10d; VI. 3. 5b; 23. 9d; 25. 2b; 50. 9d; 63. 1b; 75. 1b; VII. 19. 10d 21. 9c; 24. lc; 48. 4b; IX. 71. 9c; 97. 20a; 97. 33a; 97. 41b; 'X. 7. 7a; 12. 4c; 78. 4a;
87. 3b; 107. 11c; 120. 7a; 183. 2a. 25. IV. 33. 3d; 34. 11a; X. 15. 5d. 26. I. 186. 7a; VI. 20. 4b; 22. 8b; 26. 3a; 36. 2c; X. 39. 10a; 54. 3b. 27. I. 51. 15a; IV. 2. 12a; 34. 3a; V. 31. 5c; 41. 11c; 25. 1c; VI. 50., 4b; 50. 15c;
66. 4a; 67. 7c; VII. 3. 2a; 58. 2b; 60. la; 64. 5b; -65. 5b; 86. 7b; 97. 2a; X.
46. 2d; 48. 7d; 61. 12c; 99. 4c; 115. 9b. 28. II. 24. lla. 29. I. 168. 1c.168. 5a; V. 24. 10c; X. 32. 5a. 30. I. 51. 6b; 92. id; 112. 2c; 140. 6a; 152. 9c. II. 21. 2a; 31. 6b; IV. 36. 5d; V.
54. 2d; VII. 83. 5d; IX. 86. 11d; X. 48. 5b; 63. 14a; 113. 7c. 31. I. 85. 10a; 85. 11a; 110. 3a; 132. 4b; 138. 3b; 155. 3c; VI. 61. 3c VII. 75. 1c;
VIII. 25. 20c; IX. 73. 4a; 74. la; 74. 6a; 86. 27b; 107. 2a; X. 43. 2d; 48. 3b;
56. 6a; 66. 14d; 76. 5a; 94. 4b. 32. I. 47. 8a; IX. 107. 6c; X. 78. 7a. 100. 9a. 33. I. 36. 12a; II. 1. 7a; VI. 25. 3a; VII. 104. 200; VIII. 21. 1b; X. 48. 61:

Page #318

Traces of Short ě and o in Rgveda
309
34. II. 1. 9c; X. 49. 5c. 35. III. 10. 3c; IX. 21. 5c; 52. 4c; 62. 8a; 65. 22b. 36. III. 16. 3b; 27. 2a; 27. 14a; 41. 8a; 51. 11a; 51. 12a; 62. 14a; IX. 2. 4b; 23.
3b; 23. 5a; 34. 6b; 34. 2c; 39. 5b; 44. 6a; 42. 2a; 62. la; 63. 4a; 63. 96; 63.
16b; 64. 12a; 65. 20c; 66. 13b. 37. III. 13. 5c; 44. la; IX. 2. 5a; 6. 3b; 7. 7c; 8. 1c; 11. 7c; 23. 5c; 24. 4c 26. 1b;
26. 1c; 26. 2a; 27. 1b; 30. 4a; 31. 5c; 35. 3c; 40. la; 48. 3c; 51. la; 52. 1c;
61. 3a; 61. 21a; 62. 9b; 66. 7c; 98. 6b; 102. 3c. 38. III. 8. 7a; 9. 3d; 24. 2b; 27. 7a; 29. 10b; 42. 9c; 45. 2b; IX. 2. 7b 3. la; 3.
2c; 7. 5a; 11. 2b; 13. 9a, 18. 7b; 21. 2a; 21. 7a; 24. la; 24. 2a 26. 4c; 27. 5b; 29. 5a; 30 4b; 32. 5a; 34. 6c; 37. 4b; 38. 3b; 49. 5a; 59. 2a 61. 16a; 61. 25b; 62. 11b; 63. 5c; 66. 22a; 66. 29a; 67. 10a; 67. 19a; 67. 20a 101. 16b; 102. 7a;
103.4b. 39. III. 10. 7a; 10. Sc; 13. 4d; 16. 6d; 22. 4a; 24. 2c; 27. 4a; 37. 11c; 40. lc 40.
7b; 42. 1c; IX. 2. 5c; 3. 9c; 5. 2b; 6. 2c; 7. 2a; 7. 3a; 7. 3c; 7. 4b; 9. 2b; 10. lc; 12. 8b; 13. la; 13. 1b; 16. 4b; 16. 6a; 16. 7c; 17. 2c; 17. 3b; 18. 1b; 19. 7c; 20. 5c; 21..7b; 32. 2b; 24. 6a; 28. 2a; 28. 6b; 30. 1b; 31. 1b; 31. 5b; 32. 1c; 34. 4b; 34. 1b; 36. 4c; 37. 1b; 41. 1b;42. 5c; 45. 5a; 50. 5b; 51. 3a 52. 1b; 53. la; 53. 1b; 54. 1b; 55. 1a; 56. 1b; 57. 2b; 62. 2c; 62. 23b; 62. 25a; 63. 3b; 63. 5b; 64. 4b; 64. 5c; 64. 24a; 66. 5a; 66. 11b; 67. 1b; 67. 3a; 67. 4a; 67. 15c; 67. 22a; 67. 32a; 98. 2a; 98. 3a; 98. 3c; 98. 11b; 100. la; 100. 7b 101. la; 101. 4c; 101. 5b; 101. 7b; 101. 14a; 101. 15c; 102. 5b; 103. 2b; 103. 3b;
105. 2b; 105. 4a; 107. 10a; 107. 12b; 107. 22d; 113. 7d. 40. III. 39. 4b; 55. 17b; IX. 91. 4d. 41. III. 1. 11b; 5. 4a; 5. 4c; 5. 9c; 8. 2c; 19. 2a; 29. 7a; 33. 6a; 51. 4d; 57. 6a;
IX. 89. 6d; 91. 5d; 97. 21c; 97. 35d. 42. III. 1. 1d; 1. 3d; 1. 4d; 1. 20a; 5. 2d; 6. 5a; 14. 5a; 15. 2a; 15. 2d;15. 4a; 17.
1d; 18. la; 19. 5c; 21. 4b; 22. la; 32. 15a; 54. 21c; 54. 21d; 55. 2a; 55. 2b;
55. 3c; IX. 88. 6d; 89. 5b; 96. 10b; 97. 15d; 97. 17a; 97. 51a. 43. III. 1. 10b; 1. 18a; 5. 96; 6. 5c; 6. 3a; 20. 1d; 22. 3a; 23. 1d; 30. 10c; 30. 14d;
31. 9b; 32. 10a; 43. 7d; 46. 3d; 54. 18c; 54. 18a; 54. 20c; IX. 69. 10b; 88. 8d;
97. 12d; 97. 13a; 97. 39d; 97. 55c. 44. III. 7. 3c; 7. 5a; 8. 2b; 14. 7b; 14. 7d; 15. 3b; 18. 2a; 29. 9d; 29. 13c; 30. 6c;
32. 9b; 35. lc; 38. 4c; 38. 4d; 43. 5d; 51. 7b;53. 7b; 53. 17d; 54. 12d; 56. 2a; 57. 4b; 58. 1d; 58. 3d; 58. 4b; 59. 4b; IX. 87. 1b; 87. 6a; 87. 7b; 89. 1c; 89. 7a;91. 1c; 92. 5c; 93. id; 96. 13b; 96. 23b; 97. 4c; 97. 5c; 97. 7d; 97. 21a; 97.

Page #319

310
Amrita
31b; 97. 37b; 97. 40c. 45. III. 5. 5a; 7. 9a; 17. 5a; 32. 3d; 36. 3d; 53. 23d; 54. 2a; 54. 3c; 59. 2a; 61.
6a; IX. 90. 4d; 91. 3d; 96. la; 97. 21b; 97. 30d. 46. III. 20. 2d; 30. 9c; 34. 7c; 46. 3b; 55. 5c; 55. 8a; Ix. 89. 3b; 96. 16; 97. 11b;
97. 35c. 47. III. 1. 21b; 4. 4a; 6. 4c; 7. 7c; 14. 7b; 15. 1b; 18. 2c; 19. 4a; 38. 4a; IX. 89.
3a; 97. 54d. 48. III. 1. lc; 1. 12d; 1. 14c; 4. 2b; 4. 11a; 5. 9a; 6. 8d; 7. 2a; 8. 5a; 14. 1c; 14.
1d; 14. 5a; 17. 13b; 23. 4b; 29. 16a; 30. 2d; 30. 19c; 30. 20d; 31. 7d; 31. 11b; 32. 13a; 35. 1b; 35. 1c; 35. 5b; 36. 6d; 46. 2b; 48. 3c; 50. 4d; 53. 5b; 53. 21a; 55. 12c; 55. 15a; 56. 5c; 56. 6b; 57. id; 58. 4d; 61. 2c; 61. 5c; IX. 85. 12a; 85. 12b; 87. 4b; 87. 5d; 89. 1b; 89. 6b; 93. 2b; 96. 10b; 96. 24c; 97. 3b; 97.'
6c 97. 31d. 49. III. 2. 7c; IX. 70. 3a; 86. 15a; 86. 15b; 86. 42a. 50. III. 26. 3c; IX. 86. 19b. 51. III. 2. 3b; 3. 3b; 9. 2c; IX. 71. 3d; 72. 4a; 74. 2b; 83. 2b; 84. 1c; 85. 8c; 85.
45b. 52. III. 2. 4b; 26. 1b; IX.71. 7a; 77. 3c; 86. 8a; 86. 12b; 107. 6c; 107. 13a; 73. 1c. 53. III. 26. 6c; 29. 11d; 60. 3b; 60. 3c; 60. 7d; IX. 68. 6b; 70. 2a; 70. 4c; 71. 5b;
71. 7b; 72. 3a; 73. 3b; 75. 1b; 75. 5b; 76. lc; 76. 2d; 78. 1b; 81. 5b; 82. la; 82. 2b; 82. 3c;85. 5c; 86. 3a; 86. 3c; 86. 6c; 86. 21b; 86. 24b; 86. 25b; 86. 25c;
86. 26a;86. 27d; 86. 34b; 86. 36a; 86. 39c; 106. 8c; 107. 17c;108. 2c; 108. 4c. 54. III. 29. 15d; IX. 75. 2b; 81. 3d; 86. 3b; 107. ila. 55. III. 2. 14c; IX. 68. 9c; 69. 1c; 69. 2b; 79. 1c; 83. 3c; 107. 1c; 107. 13a. 56. III. 2. 7b; 2. 9b; IX. 60. 3a; 86. 42b; 105. 6c. . 57. III. 9. 4c; IX. 69. 4d; 71. 2b; 71. 8a; 72. 1b; 73. 2c; 73. 5b; 73. 7c; 73. 8d;
78. 5a; 80. 5d; 82. 3d; 83. 1c; 85. 4a; 86. 8c; 86. 12a; 86. 18b; 86. 44b; 86.
45c; 86. 48c; 107. 2c; 107. 13a; 107.22a. 58. [Annals. B. 0. R. I. ] 59. Vedic Grammar p. 297.
Traces of Short ě and õ in Rgveda,
ABORI. XXIX. 1948. 11

Page #320

Pāņini 1. 4. 32
This sūtra of Pānini, defining the HIGH Rich is of some methodological interest and hence is treated here in some detail. The line of argument started by the Vārttikakāra on it and followed up by Patañjali and his commentators is probably the first place in Sanskrit grammatical literature where a strict ordering of rules is envisaged, in the sense of their sequential application and not in a purely mnemonic or abbreviatory sense, for whic there is ample scope in the Astāddhyāyī of Pāṇini. In fact the division of the work into 44GHTER and fruit is of this second type and the same can be found in numerous other places on a smaller scale.
The basic function of ordering rules is to make the application of one dependent on the other, so that the effect of the first gets modified by the second or the scope of the second gets restricted to cases which are outside the scope of the first, depending upon the nature of the input to the second rule. Of these two types, the first is the normal situation in the use of Sanskrit grammar and rules in it can be both of the context-sensitive and context-free types. In the second alternative two situations can be distingished. In one, a statement is made or a rule is formulated and then its scope is extended by the addition of more specific cases. This is what Pānini does when he wants to define a Kāraka. In the present case, he defines HYGA in 1. 4. 32 to mean a person who is aimed at by the doer of the action through the intermediary of the Karman. Then the scope of this Kāraka is extended in the following sūtras so as to include a person who gets satisfied when verbs like 5 are used (1, 4.33), a person who is intended to be made aware of, with the use of the verbs like a real and 4 (1. 4. 34), to include the creditor when the verb ere is used to mean to owe someone something (1. 4. 35), things which are desired when the verb pe is used (1. 4. 36) and the person against whom anger is felt when verbs meaning anger, hatred, jealousy and envy are used (1. 4. 37). Similarly in the scope of YGF are

Page #321

312
Amrita
included a person about whom enquiry is made with the use of the verbs and (1. 4. 39) and the person who does an act which elicits a promise or consent expressed by the verbs like fag 3 or encouragement with the use of the verbs or f (1. 4. 40, 41). With these extensions, one can then form a composite concept of or indirect object, however difficult it may prove to formulate a unified idea or the image to cover them all. Once the concept with its specific range is formed the use of the fourth case i.e. Dative becomes available to it by the rule II. 3. 13 all along its range. All such rules can be thought of as either context-sensitive or context-free. according to convenience. We may state:
(I) The desired object with the use of the verb स्पृह, is called संप्रदान, is called
(II) The desired object in the context of the verb
The distinction between the two, however, is trivial, because they are really rules of semantic contents and nearer the lexical rules of transformational grammar, where a set of semantic features are assigned to a lexical item. In the case of these technical terms we can use a kind of a dummy symbol A, which will be later replaced by the appropriate case morpheme.
The other alternative arises when a rule is stated in the form of a very wide generalisation and then its scope is restricted with the addition of specific cases as its exceptions, either as obligatory or optional. Thus 1. 4. 38 restricts the scope of by replacing it within the case of the prefixed verbs 3 g etc. An optional restriction is found in 1. 4. 44 where, in the use of a verb like fat to hire for a fixed period of time, the payment is treated as optionally, being also a according to 1.
अभिक्रुध्,
4. 42.
In all these cases, the ordering of rules can be considered as merely an external feature. Whether the main rule of His stated first and then extended with additional rules or is restricted by assigning other to it as in 1. 4. 38, the order in itself is not important. The same holds good in the other case where the extending or restricting rules precede the main rule. This can be seen in the case of where the main rule given in 1. 4. 49 is restricted by rules like 1. 4. 38, 43, 46-48 all of which precede it in order." In all such cases the rules restricting or extending the scope of the concept can be formulated as either context-free or context-sensitive. What is true of them all is the fact that the place in which they occur relative to others does

Page #322

Panini 1. 4. 32
313
not affect their interpretation or scope of operation. Their order is based on matters of convenience like brevity of expression or ease of reference.
A strict ordering of rules, however, is something different. In it, the application of the second rule depends on the first in deciding its scope of operation, which is what is left over with the application of the preceding rule. A change in the order of the rules will give a different result as the ranges of the reordered rules become different. Only in such cases can we speak of a strict ordering of rules. It is a matter of some doubt whether Panini's grammar permits such an ordering. His technique appears to allow the use of both the rules with their normal ranges and then an attempt is made to eliminate the undesirable results by changing the output of one into that of the other, replacing one by another, making only one, usually the later, as operative or deny the existence of the one in favour of the other. This can be seen in his traffic rules like 311 CACRIGT I 1. 4. 1, fayfare i orth 1. 4. 2, and fHG4 VIII. 2. 1.
A genuine case of a strict ordering of rules is, however, found in the alfarah on 1. 4. 32,and Patañjali's discussion on it. While explaining the use of the word 4701 in the sūtra, it is pointed out that in its absence, the direct object of a verb like to give will become HUGH being what is intended by the subject. This can be avoided by stating that what is aimed at should be through the agency of the direct object. As explained by Kaiyata, in the sentence 3Y&RIP Ticgifa what is intended by the giver is to consider the cow as the means of reaching the teacher. Thus : becomes the 47 and 34214 the HUGH. The indirect object is related to the main action through the direct object. The use of the words (3f49fa) H (HYGHH) is intended to avoid the subject himself getting the HIGH-E1. The use of the two pre-verbs 3719 and ensures that the use of the dative is not confined to the present time but can be used in the future and past as well. The main verb indicates the present, the use of 3714 includes future and 7 the past. Of course, Kaiyata is willing to admit the artificiality of this explanation and consider the senses of number and time as not being intended at all.
Now follows the Varttika क्रियाग्रहणम् or क्रियाग्रहणमपि कर्तव्यम्. There are many usages in Sanskrit wherein the dative case is used with verbs without the use or possibility of the direct object. Patañjali gives three examples Pl निगर्हते. युद्धाय संनद्यते and पत्ये शेते. It is worth nothing that the first verb is transitive, the second reflexive and the third intransitive. The meanings of these three expressions are given by Kaiyata as stade frafa 'he denounces the

Page #323

314
Amrita
funeral ceremony' युद्धविषयकं संनहनपूर्वकं निश्चयं करोति 'he makes himself ready to fight' and funny ta 'she sleeps, approaching the husband.' That these are idiomatic expressions is implied by the remark of हरदत्त in his पदमञ्जरी on काशिकाः नैतत्सार्वत्रिकम्, किं तर्हि, प्रयोगदर्शनवशेन नियतविषयम् ।
To include such usages in the scope of HGH the Vārttika tells us to add in the sūtra a word like also. Thereby we can consider these cases as ich because they are aimed at by the subject through the agency of the action he performs. Possibly the meaning is, 'he aims at the face through the action of blaming', 'he aims at the fight through the act of making himself ready for it' and 'she aims at her husband through the act of sleeping'. Patañjali rejects this addition, not on the ground that some other explanation should be sought for these datives than HUGT but by arguing that the intended meaning of fall can be got from the word of in the sūtra, on the basis of the popular usage that af means fl. The argument is carried a step further. An objection is raised that the word 647 has two meanings, 'object which is a technical or stipulated sense and 'action' which is a natural or usual meaning. Now by a general rule the technical meaning is to be preferred to the popular meaning in a scientific treatise. Hence 044f here means anglia044 and not merely a fel. The technical sense cannot be made applicable to a fare because one action cannot be made most desired through another. Patañjali tries to justify even this with the argument that with the help of one action, expressed or implied, we desire another. The application of such an analysis to the given cases is explained by नागेश in the words : एवं च पत्ये इत्यादेर्दर्शनादिपूर्वकारम्भकर्मपत्यद्देशकं IRR intef: What is meant is, the lady sleeps aiming at her husband through the actions beginning with seeing and followed by desiring, deciding; beginning, concluding and then acquiring the result.
The extreme artificiality of the last explanation apart, the intention of All in adding the area is clear. The sūtra 1. 4. 32 defines the indirect object with the help of the direct object, whether expressed or implied, but fails to define it with verbs which are intransitive or lack the direct object, though a dative case can be used with them. To account for this, he suggests the addition of the word fonul to the sūtra to act as an intermediary in place of the the action being implied by the verb itself. We may, following नागेश, paraphrase the three expressions as निन्दनक्रियया श्राद्धं, संनहनक्रियया युद्धं and शयनक्रियया पति अभिप्रैति. नागेश adds one more subtlety to the discussion with the " remark एव च बहुलक्ष्यसंस्कारानुरोधेन कर्मशब्देनोभयं गृह्यते ।
What the situation requires is one definition of High for the transitive verbs and another for intransitive or reflexive verbs. The first can use the

Page #324

Pāṇini 1. 4. 32
315
word ofon, the second 41. To put these two together, we can visualise a couple of possibilities. One is (1) to consider the aferch to add one more definition of HIGH, fruita afufa # HIGH4, as phrased by the church. The other (II) is to add the word क्रियया to the first making it कर्मणा क्रिययापि यमभिप्रैति # HYGF74. In this case the two conditions can be taken either in a (a) conjunctive or (b) disjunctive sense. The results of these alternative ways of taking the rules differ and may give us a clue to deciding how the Vārttikakāra intended them to be taken. To take the alternative Ila is clearly inadmissible because it does not improve the position of the sūtra 1.4.32. by including the intransitive verbs and the addition becomes redundant. In the second case IIb the position becomes the same as I, with the two rules contracted into one. In both cases the use of her to the exclusion of 440 leads to the undesirable result which happens when the word ou is simply dropped from the sūtra. This is due to the fact that there is implied in the context of 3fafa and becomes prominent when Anfo is dropped. As ka puts it : यदि तहि क्रिययाभिप्रेयमाणस्यापि भवति कटं करोतीत्यादावपि प्राप्नोति वचनाद्धि कर्मसंप्रदानसंज्ञयोः
fa: PG. The addition of the words : tarifah for afe! GUTAS Freafare does not solve the problem, but certainly offers a hint to that effect.
The only way out of the difficulty is to consider the rules as being strictly ordered. We may call these rules for and fh4f1919 rules. We may then arrange them in this order and apply them one after another. After the 064ff 1914 rule has been applied, which will deal with transitive verbs, we apply the
479 rule which will be then applicable to intransitive or reflexive verbs and such other verbs which have no object expressed or implied. In such a case the distinction between I and IIa becomes operationally of no value. After we have applied the rule of fufftype we may apply the rule which lays down both the conditions in a conjunctive sense. But the result will be the same; because, as a consequence of the strict ordering of the rules, the first condition becomes vacuous and only the second condition remains operative, now being applicable to the verbs left over after the effect of the first rule i.e. intransitive and reflexive verbs or verbs without a direct object. Thus a strict ordering of the rules alone will satisfy the needs of the case and we may be justified in thinking that the Vārttikakāra intended such an ordering of his rule vis-a-vis the sūtra of Pānini 1. 4. 32. He probably intended to indicate it with the use of the word 319. In this sense, the form of the Värttika fMETHY onder is to be preferred to the shorter one क्रियाग्रहणम्.
Pānini 1. 4. 32
Indian Linguistics 1968
14

Page #325

Some Etymologies in Manusmộti
ETYMOLOGY is one of the potent means of explanation with Indian writers and very often they make the etymology of a word serve the function of an explanation. This inordinate desire to etymologise a word is very old and must have started from the days of the Brāhmaṇas which offer a number of such explanations of words, many of which are, however, quite fanciful. That this school of the etymologists grew very rapidly is evident from the remarks of Yāska about his predecessors and his own contributions in the field of etymology. In all such efforts, these writers were helped by the extraordinary transparency of the Sanskrit language and whenever the source of the word could be ascertained by reference to the material available in the language itself, the Indian school of etymologists and following them the grammarians have hit. upon the correct origin of the word. But their ignorance of any of the cognate languages made them impossible to guess the correct source of so many words which have lost all connection with other words of the language or available roots and in all such cases they have satisfied themselves and their craving for etymology by a fanciful process of word-analysis in which they attempted to find out the known roots of the language. That such attempts were made quite early is apparent from the remarks of Yaska and sākatāyana's analysis of the word satya given by him. This was the result of the formulation of the extreme view that all nouns must come from verbs or roots, a view which modern linguistics has proved to be far from true.
The Dharmaśāstra attributed to Manu has indulged in a number of etymological explanations of a few words in course of discussing law proper, which are interesting both as regards their fanciful nature and the use made of them by the writer. Thus there are seven such cases which may be discussed in detail in order to ascertain their real origin and in what light the medieval commentators took them.

Page #326

Some Etymologies in Manusmrti
317
. The most famous of them is the etymology of the name of the Supreme Being nārāyana which is made to serve as a kind of proof for the cosmogonical speculations of the Smrti in the first chapter. The Creator first Created the waters and was himself born in them. Then the author says 1.10. äpo nārā iti proktā āpo vai narasūnavaḥ /
tā yad asyāyanaṁ pūrvam tena nārāyanah smrtah //
He was called Nārāyana because the place of his origin was the primordial water which was called nārāh (nārā + ayana) which in their turn, were so called because they were the sons (i. e. first products) of nara the creator himself. The commentators go into greater details and explain how the formation is to be analysed. Thus Medhātithi says : apo vai narasūnavah / sa bhaved, bhagavān naraḥ puruşa iti prasiddhaḥ, āpaś ca tasya sūnavo 'patyāni atas tā naraśabdena ucyante / drsto hi pitrśabdo patye / vasisthasya apatyāni vasisthah / .. iti abhedopacarena tā āpo naraśabdavācyāh / yad yena prakāreņa asya prajāpateḥ pūrvam ayanaṁ prathamaḥ sargaḥ āśrayo vā garbhasthasya tena hetunā nārāyanaḥ smrtaḥ / narā ayanam asyeti narāyana iti prāpte "anyesam api drśyate" (Pān. VI. 3. 137) iti dirghah / pūrusa iti yatha / athavā Sāmühiko 'n // This means that the word is made up of nara + ayana = narayana which is further changed to nārāyana where nara refers to the first creation of the Supreme Being i. e. water and ayana may mean the first creation or the Substratum. In the first case the word nārāyana comes to mean nara by a roundabout way, the waters are called narāh because created by nara, and the creator is called nārāyana because the waters were his first creation. The second alternative appears to be the intention of the Smrti, as the preceding verses point out. For the lengthening of nara into nāra Medhātithi has given two suggestions, that it may be either arbitrary as in pūrusa or it may be the collective suffix an (= a) which causes Vrddhi and replaces the final vowel -a giving the form nāra. The explanation of Govindarāja is nearly identical except that he reads Manu's verses as 'āpo narā iti proktāh' thus making the lengthening of nara into nāra, according to the first explanation of Medhātithi, a part of Manu's statement.
Kullūka has followed the second explanation of Medhātithi and only points out that as the word āpaḥ is feminine the feminine form of nāra formed by an (Pņ IV. 1. 83) should take nīp but here it takes irregularly tāp, giving the form nārāḥ. The remaining commentators have nothing new to add.
This etymology of Manu is probably copied by Visnupurāna (27) and

Page #327

318
Amrita
Mahabharata (III. 189. 3) which reads naränäm ayanāc cäpi tato nārāyaṇaḥ smṛtah/ The grammar of Panini has explained the word in a quite different manner. The word nara occurs in the gana nadadi of the Sütra naḍādibhyaḥ phak (Pan. IV. 1. 99) which would make it a taddhita derivation in the sense of a descendant, thus making närayana a descendant of nara. Later grammarians give both the explanations side by side. Thus for example Bhānūji' says naräṇāṁ samûho nāram / tasya samühaḥ (Pän. IV. 2. 37) iti an/ tad ayanam yasya narasya apatyam iti va // naḍādibhyaḥ phak..
This word nārāyaṇa is purely an Indian formation and its etymology is obscure. There arise two questions in this connection whether the reading of Manu should be naraḥ or näräḥ and whether the word really meant 'water' as is asserted there. That, in spite of the reason given by Manusmrti that waters are so-called because they are the sons of nara (narasûnavah), it is not improbable that the word may have preserved its primitive meaning which was lost to classical Sanskrit. In Rgveda we do find a compound word svarṇara besides svarnṛ which appears to show two distinct meanings. In a number of cases it is used as an attribute of Agni where it may mean the 'heavenly man'. Thus II. 2. 1 samidhānām suprayasam svarnaram the enkindled, well fed, the heavenly man.' VI. 15. 4. dyutānam vo atithim svärnaram agnim (I decorate) Agni, the bright one, your guest, the heavenly man'. VIII. 19. 1 täm gürdhayā svārṇaram/ Extol that heavenly man.'
In one or two cases the word occurs in company of other proper names. VIII. 3. 12. krpam indra pravah svarnaram / "as you helped Krpa and Svarnara O Indra." VIII. 12. 2 yéna daśagvam adhrigum vepayantam svarnaram, yénā samudram avithā tamimahe "We desire that by which you protected irresistible Daśagva, the Svarnara who moves all and the ocean."
But in the majority of cases the word svarṇara appears to express the meaning of heavenly water or a fountain of heavenly water. This is the case when the word is used with the verb mand- or mad- (causal). VIII. 6.39 mandasvä su svarnare "Take delight well in the heavenly water". VIII. 65. 2. yad vā prasravane divo mādayāse svarnare, yad va samudré andhasah "and when you become exhilarated in the heavenly stream, in the heavenly water, and when in the ocean of Soma drink." VIII. 103. 14 mädayasva svarṇare/ "take delight in the heavenly water". IV. 21. 3 a yātu .... svarṇarād avase no marutvān / "Let (Indra) come from the heavenly fountain to help us, along with the Marut." V. 18. 4. stirnam barhiḥ svarṇare śravämsi dadhire pari /

Page #328

Some Etymologies in Manusmộti
319
"the sacrificial grass is spread around the heavenly fountain and they have placed their glories there.” X. 65. 4 svarnaram antariksāņi rocana skambhuḥ / "they supported the heavenly fountain, heaven and earth and the bright regions."
In all these cases the description appears to suggest a heavenly place where the gods can take delight as they do in the juice of Soma here. In fact it appears to be a heavenly counterpart of the Soma Juice offered to the gods in the sacrifice. In any case the compound expression svarnara obviously contains the word nara which distinctly refers to some liquid and the whole word, some place filled with it. If this is so we can readily believe that the word nara expressed the idea of water and was used in the plural like its synonym āpaḥ and the statement āpo narāḥ iti proktāḥ may be accepted as true. The word nārāyana would be then regularly derived from nara in the sense of 'descended from water'. The only modification, which the Manusmrti has introduced, pertains to the extension of the chain of evolution backwards by tracing the origin of the water to the First Man and thus drawing the natural conclusion that the waters were the 'sons' of that nara. Naturally we should accept the reading narāh as being the original one and even from the wording of the comment of Medhātithi, it may have been before him. But very early this word nara meaning 'water' must have got confused with the other word nara, the thematic form of nr meaning 'man'; which led to further speculation and the creation of the myth about the sages nara and nārāyana and the way of Manu's explanation.
In II. 76 Manu gives an analysis of the word om, the three constituent sounds of which a, u and m were milked by the creator from the three Vedas. Here the Smrti is simply following the views of the Upanisads, particularly in Chandogya 2. 23.2. which is obvious from the mention of the three Vyakrtis in both the places. But the real origin of this interjection has been clearly pointed out by J. BLOCH who draws attention to the fact that in Indo-Aryan languages a final long vowel, much more a pluta one, shows a tendency to get nasalised. Thus the simple interjection o was nasalised into
which later became (om). This agrees well with its original meaning of an assent, an affirmation which it has kept long into the classical period. The analysis of the word into a, u and m, based upon the Sandhi of a + u = o is quite fanciful but may be pretty old.
While discussing the duty of honouring guests Manu (III. 102) gives the etymology of the word atithi in the words: anityam hi sthito yasmāt

Page #329

320
Amrita
tasmäd atithir ucyate / "Because he does not live for long he is called atithi". Here again the commentators try to justify the explanation in various ways. Medhātithi says: anityam hi sthitiḥ/tisthateḥ atipūrvasya ayam sabdaḥ / caṇādikaiḥ kathamcid vyutpattih / This means that he derives the word from atistha-. Here ati naturally will have the meaning of 'beyond', 'surpassing" etc. But it is obvious that he cannot fully explain its formation and, as will be seen, there cannot be a Uṇādi suffix to justify this etymology. Kullūka takes the word to be made up of a and tithi and explains na vidyate dritīyā tithir asya iti atithiḥ/ Herein he is followed by Raghavananda. The explanation of Nandana is quite fanciful: antyaśabdāt akāraṁ ikäraṁ takāram ca adaya sthitisabdāt thakāraṁ ikāraṁ ca ādāya saṁniveśya atithiḥ iti ucyate/ In all probability the author is imitating the citation of Yaksa on the word satya. How can anityam give ati ? This etymology of atithi is similar to the one found in the commentary on Khuddakapātha which says: natthi assa thiti yamhi va tamhi va divase agacchatiti atithi /
The Unādisūtras derive the word from at- 'to wander' and the suffix ithin (IV. 2) which appears to be probably the correct explanation. This word is the same as the other one atithin (RV. X. 68. 3) which means 'wanderer'. That the word is of Indo-Iranian age is evident from the Avestan word artis corresponding to Sanskrit atithi. But the relation of the 'word with the root at was soon forgotten because the dental of the root became cerebralised and classical Sanskrit knew only of at- 'to wander'.
Manu's explanation of the word māmsa 'flesh' is quite fanciful. He says (V. 55) that because the eater of flesh is to think that as a punishment he will be eaten by his own victim in the next life, flesh is to be called mämsa. It is thus made up of măm (Acc. sing. I pronoun) referring to the person who eats flesh and sa (= sah) the demonstrative pronoun in the third person singular masculine referring to the victim which is eaten. That this could not satisfy even the commentators is obvious and Medhātithi makes the usual remark nāmadheyanirvacanaṁ arthavādaḥ suggesting thereby that all such attempts of etymology are worthless.
The Uṇādisūtra (III. 64) maner dirghaś ca / derives māmsa from man'to think' and the suffix sa (cf. III. 62). A similar formation is to be found in the verb mimämsate and mimämsä derived from the root man- according to Panini III. 1. 6.
That even this etymology is wide of the mark can be seen from the fact that mamsa is a word, representing a thematic form of the more

Page #330

Some Etymologies in Manusmrti
primitive mams. just as mäsa, is from mäs. Thus to find a nominal suffix -sa in mămsa is out of question. That the word is of Indo-European origin will be evident from parallels in other cognate languages. Thus Greek has the word mérós 'thigh' Latin membrum of doubtful etymology 'limb', Gothic mimz 'flesh' and OCS. meso 'flesh'. This leads us to postulate an Indo-European word like memsro- in the sense of a limb of an animal, flesh'.
321
Explaining the word vṛṣala which has come to mean a low caste man, Manu remarks that gods call him so because he obstructs right or justice. He analyses the word into vṛşa which means dharma and the particle alam in the sense of enough (VIII. 16). Medhätithi explains how vṛsa comes to mean religion vṛṣasya kamavarṣinaḥ dharmasya 'of Vṛşa, religion, which showers all desires. His derivation of vrsa is correct but that of vrsala is linguistically
useless.
The Uṇādisūtra (1. 111) vṛṣādībhyāścit derives the word from vṛs-+ala. From RV. onwards the words vṛṣan and vṛṣala are frequently used and form a kind of contrast, the first expressing the idea of a vigorous bull often applied to Indra and the second meaning a castrated bull and used with Vṛtra in a derogatory sense. That vrs- means to shower and vṛsan a vigorous bull is evident from comparisons like Gr. 'erse dew' Lat. ros 'moisture' all from the primitive ueres.
Manu's explanation of the word jāyā (IX. 8) wife is so called because the husband is born in her' is copied from the statement in the Aitareya Brāhmaṇa VII. 13. tajjāyā jāyā bhavati yad asyām jāyate puṇaḥ / Leaving aside the fiction that the husband is reborn as his own son, the etymological part of the explanation is quite sound. The word is to be split as jā (= jan- to produce) and the nominal suffix -ya. The Unädisutra (VI. 112) janeryuk explains it thus and the other word bhäryä from bhr-+ ya would furnish a good parallel. The word means 'one who gives birth' (jan- jana-yati). That the author should explain it as yasyam jāyate and not ya janayati is again due to the fact that the root jan- has undergone a change in its meaning in course of its history. The primitive root gene- must have both a transitive and an intransitive meaning as in Veda jajäna 'created' and jajñe 'was born'. When the transitive sense becomes obsolete in Classical Sanskrit it was not possible to take its derivative jäyä in an active sense and hence the fiction of the husband taking birth in her. This was helped by the weak grade of the root ja- which was also to be found in the passive form jayate.
Finally Manu gives the explanation of the word putra 'a son' (IX. 138).
*

Page #331

322
According to him the son is so called because he protects his parents from falling in a hell called put. Thus the word is made up of put + tra- 'to protect'. The commentators have nothing important to say except that Raghavananda wrongly calls the hell pum which could not explain the word
putra.
The Unādisūtra (IV. 164) puvo hrasvaśca explains the word as pū'purify' = tra (nominal suffix). The difficulty here would be the gender of the noun which should have been neuter as in case of mitra and kalatra in spite of their meanings.
Here again the word is old. We have parallels like Avesta puthra, Lat. puer 'boy' Gr. pais 'child'. The Latin word pullus (from putslos) means 'a young one of an animal', so also the Lith. putytis. This may suggest an IndoEuropean base like putlo which would become putra in Sanskrit by the usual change of -- to -. That this was a diminutive of the word pum 'man' is also probable and would naturally explain how the word came to mean 'a
son'.
Amrita
From this review of the few etymologies offered by the Manusmrti it should become clear that the contention that whenever the material in the Sanskrit language was insufficient to arrive at the correct source of the word the Indian etymologists have simply satisfied their desire for derivation in a fanciful manner with obviously wrong explanations. But when the language afforded them the slightest help as in the case of atithi and jāyā they have rightly hit upon the real source of the word.
Annotatios :
1. Nirukta, I. 13.
2. Mandlik's Ed. p. 18.
3. Vyakhyasudha on Amarakosa, p. 13.
4. Cf. GELDNER Rigveda in Auswahl, I. p. 209.
5. Lindo-Aryen, p. 45.
6. Ibid., p. 57.
7. BRUGMANN; Kurze vergleichende Grammatik, p. 111.
ᄆ
Some Etymologies in Manusmrti
BV. W. 1. 1942
32

Page #332

A Note On Utkalāpaya
BÖHTLINGK and ROTH have the following note in the first Volume of their Dictionary "utkalāpa (ud + kalāpa) adj. with upraised (spread out) plumage (of a peacock). Ragh. 16. 64. Mrtśćak. 76. 3. From this the denominative utkalāpay- to cause the peacock to spread its tail, fig. to bid someone to be proud; to recognise someone's merits; to pay thanks; (?) vayaṁ sarve vidyāpāre gatāḥ / tadupādhyayāmutkalāpayitvā svadese gacchāma / tathaivam kriyatāmityuktā brāhmānā upādhyāyamutkalāpayit- vānujñāṁ labdhvā pustakāni nitvā pracalitāh / Pañcat. 244. 24."
In the Verbesserungen und Nächträge to their Vth Volume they add "utkalapa, utkalāpay- is according to BENFEY causal of kal- with ud- : it means (1) to take leave of someone (Acc.) Pañcat. 244. 25. ed. orn. 53. 15; Vet. in Gott. gel. Anz. 1860. p. 736. (2) to carry the wife from the house of the father to one's house; Vet. in LA(II) 17. 14; Gott. gel. Anz. 1860. p. 736. cp. utkalāpana." M. WILLIAMS and APTE have nearly the same to say on this word.
Many things about this word make one suspicious both as regards its form and meaning. It will be seen that it occurs only in two books the Pañcatantra and the Vetālapañcavimśikā, and it is taken either as a denominative or a causal. The form utkalāpayitvā occurring in Pañc. with the termination of the Gerund as tvā in spite of the pre-verb ud would suggest some kind of Prākrit influence. The meanings derived from it, by regarding it as a denominative of utkalāpa are not found in literature and the other two (1) 'to take leave' and (2) 'to take the wife away from the father's house' are not agreeing with both the derivatives. Neither do they show any inter-relation between themselves. One is naturally forced to see some kind of misreading or some other confusion as regards the origin of this word. It is evident that the word as used in the two popular books has nothing to do with the noun utkalāpa as found used by Kālidāsa.
Both the Pañcatantra and the Vetālapañcavimśikā are popular works and the commonly used recension of the former comes most probably from the Jain sources. As such we should expect that the word is some kind of Sanskritisation of an original Prākrit or Vernacular expression to be found in

Page #333

324
Amrita
Jaina writings. In the Bharatakadvātrimśikā edited by HERTEL we find two places where a similar word is found used : p. 28. An ascetic takes away a few sugar canes from the field of a wealthy person without his permission; but to avoid being called a thief he asks permission of the field itself and himself gives the answer as he wanted. The owner one day catches hold of him and asks him why he has stolen the canes, to which he answers, "aham sarvadāpi mutkalāpyaiveksudamdan grhītavān" suggesting thereby that he has taken the permission of the field. p. 36. "sacaturdaśa vidyāḥ sāngopāngā adhitya....taṁ gurum mutkalāpya svakiyapuraṁ prati cacāla”. From these two places it is clear that the word mutkalāpya is used in the sense of 'to take permission or leave and the second passage closely agrees with the one in Pañc. We can naturally think that both the words are in reality one and the same and the form was originally mutkalāp- with the meaning to take leave'. .
About the origin of this word itself we find that Hemacandra in his DEŠĪNĀMAMĀLĀ gives the explanation 'mukkalamucie saire' VI. 147. according to which the Deśī word mukkala means both 'proper' (ucita), and 'free' (svaira). Dhanapāla further confirms this statement when he states in his PĀIALACCHIĀNĀMAMĀLĀ 'sacchamdā uddāmā niraggalā mukkalā vissaṁkhalaya' 13. From this Prākrit word mukkala, a denominative, can be formed by the adding of the usual suffix ave in Prākrit as mukkalāvei to make free, to be free, to do the proper things' according to the meanings of the original word. It was, it appears, wrongly Sanskritised as mutkalāpayati from which the forms in use were derived. The Prākrit word mukkala continues to live in the modern languages in the form of mokalā 'free' in Marāthī, mukalāvo 'to make free in Gujarāti. From its original meaning 'to free' it is easy to see its other meanings, the student at the end of his studies freeing himself from the authority of the teacher, and the carrying of the bride from her father's house is to free her from the authority of the father who was supposed to be her guardian upto the time, not only till her marriage but the formal departure from the paternal house.
The Prākrit word mukkala itself appears to be derived from the past passive participle mukka from the Sanskrit root muc to release from a form like *mukna as suggested by PISCHEL. The verb mukkaï has given rise to the Marāthi verb mukane 'to lose'. The form extended with the addition of the syllable la we find used in the Apabhraíśa verse quoted by Hemacandra vaddappaņu pari pāviai hatthiṁ mokkaladena "but greatness is obtained by loosening the hand (by giving many gifts)”. VIII. 4. 366.
A Note On Utkalāpaya
NIA. 1. 5. 1938
34

Page #334

Group, 4 : Linguistics and
Lexicography

Page #335


Page #336

The Text of the Tatvarthādhigama-sūtrāņi
Most of the philosophical schools in India have a Sutra work of their own which summarises and arranges their philosophical tenets. The only Sutra work of Jainism of any importance is no doubt the Tatvarthadhigama-Sūtrāņi1 of Vacaka Umāsvāti, which is their most authoritative book. Besides being the clearest epitome of the Jain philosophy, it has the further advantage of being recognized as authoritative by both the sects of the Jain community, the Digambaras and the Svetambars. Both the sects accept the text of the work with slight changes in the Sutra Pathas. Its popularity and importance can be further seen by the great number of commentaries and subcommentaries on it written by the greatest scholars of both the sects. Tradition would like us to suppose that the whole of the work in its present shape belongs to Umāsvāti, and embodies the philosophical principles of Jainism as preached by Mahāvīra, their last Tīrthankara. But the work presents us many interesting problems about its text and its development which it is worth our while to consider.
The Sutra of Umäsvāti has come down to us in two different recensions each belonging to one of the two sects of the Jain community. The Śvetāmbara recension differs greatly from the one accepted by the Digambaras both as regards its wording and interpretation. Another important difference between the traditions of these two communities is the fact that the Digambaras own only the Sutras as the work of Umāsvāti and the commentators of this sect write their glosses on the Sutras only, while the Śvetāmbaras add a Bhāṣya to the work which they attribute to Umāsvāti himself and call it a Svopajña Bhāṣya. Thus both the Sutras and the accompanying Bhāṣya form the text to the Svetambara commentators to comment upon. This difference has caused many other divergences between the two sects. Of them the question of interpretation and the consequent differences of opinion on theological points is of less importance in

Page #337

328
Amrita
determining the textual problems of the present work and are of no farreaching results. They turn on points like the question of regarding living beings embodied in the elements of fire and wind as movable or immovable, the number of the heavens and hells and the number of the Nayas which last again is only a difference in the process of division more than in the final result. But the other two differences of the wording and the authorship of the Bhāsya are important to us and must be taken for a detailed consideration.
One of the differences which strikes one prominently in comparing the two recensions of the Sūtras is the difference of wording that is found in the two Sūtra Pathas. One thing can be easily seen in it, that is, the difference mainly pertains to the technical terms of the Jain philosophy and theology used in the work. In Sūtra VII, 4 the word Apāya is found used which is also to be met with in the same form in the corresponding Sūtra of the Digambara version. The meaning of this word used in this context is the usual one of danger. But in Sūtra I, 15 the same word Apāya which is used in its technical sense of one of the four divisions of the Matijñāna is found transcribed in the Digambara version in the form of Avāya. This leaves no doubt as regards the exact scope of these verbal differences. Similarly where the svetāmbaras read Manahparyāya (I, 24, 25), Upapāta (II, 32, 35), Aupapātika (II, 47), Parāpare (III, 17) and Visarga (V, 16) the Digambara Version is found to use Manahparyaya, Upapāda, Aupapādika, Parāvare and Visarpa. Now the difference between these two sets of the terms is such as to suggest an obvious explanation. These terms, mostly technical in meaning, were originally current in their Prakrit form in Jain philosophy and are found in abundance in the Jain canon. The two sects translated them in Sanskrit in two different ways and produced these two sets of words which was helped by the ambiguous and flexible nature of the Prakrit language. This fact, if true, reveals naturally another interesting point about the language in which the Jain canon must have been written. It shows that in that language both t and d were represented by one and the same sound, a conjunct was the cause of shortening a long vowel before it and p and v were reduced to one sound.
This fact, that the technical terms in the Sūtras have a Prakrit origin, has led some to believe that the whole of the Tatvarthādhigama-Sūtrani was originally written in a Prakrit dialect and was later rendered into its present Sanskrit form. There is however not sufficient evidence to conclude that the work was in Prakrit and the adduced parallel of the Mokkhamaggagai, the 27th chapter of the Uttarādhyayana is certainly inadequate. There are no doubt a few facts which would give colour to such a theory. The work is

Page #338

329
The Text of the Tatvärthädhigama-sütrāņi
called in the Sambandha-Kärikäs by the name Mokṣamärga and there appears to be an apparent similarity of form between them as both begin with a definition of the so-called Mokṣamärga. But more than this similarity the differences between these two works are more important and far-reaching.. The work of Umäsväti is better known by its other title TatvärthadhigamaSuträni which is also mentioned in another of the Sambandha-Kärikäs. The word Akhyā used there leaves no doubt that the author intended it to be the real name of the work, while the words Moksamärga in verse 33 are intended to introduce the first Sütra. About the authorship of these verses we have to conclude that they have a greater chance of being the work of Umäsväti himself and that would give it a greater support. Similarly a comparision of the work with the Uttarådhyayana chapter reveals many points of divergence between the two. While the Tatvarthādhigama includes in its conception of the Mokṣamärgȧ only the three jewels of Jainism, Samyag-Darśana, SamyagJñana and Samyak-Caritra the Uttaradhyayana adds Tapas to it. The treatment of Samyaktva between the two works is also completely different, and while the Tatvärfhädhigama' gives only the seven Tatvas the Uttarädhyayana gives the nine Padärthas as the basis of treatment".
More important is the question about the authorship of the Bhasyal which, as said already, is the bone of contention between the two sects of the Jain community. In the present state of our knowledge it is very difficult to decide the question one way or the other. Usually three different arguments are advanced to prove the common authorship of the Sutras and the Bhasya. It is pointed out that the writer of the Bhasya, while introducing the Sūtras uses the first person forms like Pravakṣyāmi, Vaksyämah" which would go to prove that the author of both the Sutras and the Bhasya is one and the same person. To this is added the fact that unlike other Sütra works the present one has not suffered much in interpretation which can be explained by supposing that the Sütrakāra himself explained his work in his Bhasya, putting all innovations in interpretation to silence. And lastly it is pointed out that at the end of the Bhāṣya the writer refers to himself by the name Umäsväti.
But on a closer scrutiny it will be seen that the arguments are not very convincing and leave many loop-holes in them. Against the few instances of the use of the first person occurring in Sutras v, 22, 37 and 42 there are a host of others in which the third person is used throughout, as Atrocyate, Atraha, and Uktam bhavatā and similar expressions 12. The only conclusive argument in such cases would be a reference in the Sutra to a passage or topic dealt with in the Bhasya only, as is the case with the Kärikäs of

Page #339

330
Amrita
Kavyaprakāśa. But such a reference we find nowhere in the present case. Further we have the instance of Jinabhadra who in his Viśesävaśyakabhāṣya on the Avaśyaka-Niryukti introduces a Gāthā of Bhadrabahu with the use of a first person form. On the other hand we have indications to show that the Sūtrakära was different from the Bhāṣyakāra. Very often in the beginning of chapters three and four, and in some other places the Bhāṣyakära uses the words Uktam bhavată in referring not only to the Sütras that went before it but also to those that are yet to come1. The full force of this use of past tense the commentators appear to have missed as they were led away by their own preconception to regarding the two works as the writings of one and the same author. But in a similar case Abhinavagupta in his Locana on the Dhvanyaloka where he appears to regard the author of the Vṛtti as different from the author of the Kärikäs remarks on the words Daraśitam eva agre, darśitameveti kārikākāreņa iti bhütapratyayah's. This leaves no doubt as to the significance of the past tense. At least it is clear that all the Sutras were ready, before the writer of the Bhasya began to write down his own work. Nor is the argument from the lack of divergent interpretations sufficient to prove the mono-authorship of the two works. It will at the most give credit to the author for writing clear and unambiguous Sutras leaving no scope for the commentators to distort and misinterpret them. The verses at the end of the Bhāṣya would at first lead one to suppose that the Bhāṣya is the work of Umāsväti himself. But a careful reading of them gives no hint to such a view. They can very well be supposed to be the work of the Bhäsyakära who is supplying us information about the author of the Sutras, and it is difficult to see how they confirm the view that the Bhäsyakāra was Umäsväti himself. On the other hand it is inconceivable that he would have praised himself in those terms found in them. The case of the Sambandha-Kärikäs at the beginning is a little different. I see no reason to say that they do not belong to the writer of the Sutras. Beside the use of the words Vakṣyāmi and Pravakṣyāmi the style of these verses is certainly superior to that of the Bhāṣya and it is not an impossibility that the writer of the Sūtras should have prefixed a few Kärikäs to his Sutras by way of introduction. In fact they discharge the function admirably. This will also explain the quotation of these verses by later writers as the words of Umäsväti even though no great weight can be placed on it, because very early there arose the tradition of the monoauthorship of the Sütras and Bhāṣya. The oldest available commentary on them, the Sarvärthasiddhi of Pujyapada" only comments on the Sūtras, and neglects the Bhāṣya altogether, which procedure is inexplicable on any other supposition than the fact that Pujyapada regarded the Bhāṣya, if at all he

Page #340

The Text of the Tatvārthādhigama-sūtrāņi
331
knew it, to be not the work of the writer of the Sūtras.
About the development of the present Sūtras it is not possible to state anything with confidence and with an air of certainty. It is only a few suggestions and traces of such a development that we can hope to glean in the existing work itself. If we look at chapters three and four of the present work we find many traces to conclude that they may be later additions to the original work. One fact that strikes us at the beginning is the presence of the two names that are given to these chapters, the Lokaprajñapti and Devagatipradarśana, while the names of the other chapters are not found or did not exist. The names of these two chapters are such that they indicate the nature of these sections as independent works, and other works bearing similar titles are found in Jain literaturel. Moreover, the two recensions of the Sūtras are found to differ unusually at a greater length in these two chapters only. In the third, the Svetāmbara recension has only 18 Sūtras while the Digambara one counts as many as 39, leaving a difference of 21 Sūtras. The difference of the fourth chapter is 11 while the next greatest difference is only 5 found in the 7th chapter. This is a clear indication to show that these two chapters are of a later origin. One more argument that is to be noted in this connection is the defective arrangement of the fourth chapter. We usually find Umāsvāti very consistent and accurate in arranging his material. But such is not the case with this chapter. It begins with the statement that gods are divided into four groups, and what we next expect is the names of these four Nikāyas. But the next Sūtra states that the third group is characterised by the yellow Leśyā. Another fact pointing to a similar conclusion is the style of the Bhāsya on these two chapters which differs from the remaining portions.
Another group of Sūtras that intrude on the general arrangement of the work is found in chapter fifth and includes Sūtras 29-31. In fact Sūtra 3220 should naturally follow without a break on Sūtra 28 as being of the same train of thought. An external indication to show the interpolatory character of these Sūtras is the difference which the comments of both Siddhasena and Haribhadra show in accepting the extent and wording of the text. Moreover the general discussion of the nature of existence should either come at the end of the chapter as is the case with the general discussion about Drāvya, or if possible at the beginning. Again the enumeration of the four Nayas based on sūtra 31 should have been dealt with in the first chapter where the topic was once opened. The Nature of these Nayas and their novel names and the great pains that Siddhasena takes to make them square with

Page #341

332
Amrita
the usual seven Nayas would point to a similar conclusion. If another suggestion can be made of the original scheme of the Sūtras, we expect that chapter nine should have been divided into two at Sūtra 18, each part being devoted to each of the two principles, Samvara and Nirjarā.
We have again a suggestion or two in the book23 itself which would throw light on the difference of number of the two recensions of the Sūtras, the Svetāmbara one containing 344 and the Digambara recension having 367 Sūtras. At the end of chapter 8 we have in the Bhāsya on Sūtra VIII, 26 the concluding words "ato anyat pāpam" which the Bhāsyakāra deduces by implication from the Sūtra itself. But these very words form a Sūtra in the Digambara version as Sūtra VIII, 26. Similarly a portion of the Bhāsya from “syādetat lokāntāt” upto “mukto niskriya iti” on Sūtra X, 6 in the Svetāmbara recension is certainly not a Bhāsya on Sūtra 6 as is supposed by the commentators. The passage runs as follows. "syādetat lokāntādapyurdhvam muktasya gatih kimartham na bhavatiti/Atrocyate / Dharmāstikāyābhāvāt / Dharmāstikāyo hi jīvapudgalānām gatyupagrahena upakurute / Sa tatra nästi/ Tasmāt gatyupagrahakāranābhāvāt parato gatir na bhavati apsu alābuvat/ Nādho na tīryak ityuktam / Tatreva anuśrenigatir lokānte avatisthate muktas niskryas iti /". From this it is clear that the words atrocyate introduce a new Sūtra which is given as Dharmāstikāyābhāvāt which is further explained by the Bhāsya that follows it. And in fact this Sūtra is found in the Digambara recension as Sūtra X, 8. It is also not very difficult to explain the conclusion in which the commentators have fallen while explaining the previous Sūtra. The Bhāsyakāra again uses the simile apsu alābuvat but in a slightly different sense, which can be easily explained with the help of the context. It means that just as the alābu cannot go further on its course when it reaches the surface of the water so is the case with Jīva which also cannot proceed further on its course. But misled by the simile as being adduced from Sūtra 6 where it is already used they take the whole passage as the words of the Bhāsyakāra on the earlier Sūtra.
This point raises the important question about the origin of the two recensions. Even though we have already explained the major portion of it as due to later additions or misrepresentation, a few still remain which it is difficult to explain with the present available materiai. But one thing becomes clear and that is the Svetāmbara recension must have suffered greater changes in its course than the Digambara one. This is further corroborated by the fact that the Sūtras and the Bhāsya being regarded as the work of one and the same person it was not possible for them to keep these two works

Page #342

The Text of the Tatvārthādhigama-sūtrāņi
333
fully distinct. It can also be gathered from the remarks of Siddhasena in his Bhāsyānusārini where he very often complains of previous writers who appear to have confused the text of the Sūtra and the Bhāsya with each other.
Annotations : 1. It is to be noted that of the Sūtra works of the heterodox systems of philosophy
we know only two, the Brhaspati Sūtras of the Cārvākas and the present work
belonging to Jainism and have only the later one. 2. T.S. II, 13 and the corresponding Sūtra of the Digambara recension. 3. T. S. I, 34. 4. No. 31.
5. No 22. 6. Kundakunda and others add Tapas in such a context. 7. T. S. I, 1. Utta. XXVII, 2. 8. Utta. XXVII, 16-27. T. S. I, 2 foll.
9. Utta. XXVII, 14, T: S. I, 3. 10. Cp. Pt. Sukhalalji in Anekānta. 11. Sambandha-Kārikā 31; V. 37. 12. cf. T. S. I, 25, 26, 27, 31 and so on; II, 25 etc. 13. I am indebted to Prof. Upadhye for this information. 14. Cf. Bhāsya on IV, 1. 15. Kävyamālā No. 25 p. 138. 16. Pujyapāda Akalanka and others do not give these verses. 17. Ed. by Nitave at Kolhapur. 18. Cp. Jambudvipaprajñapti, Suryaprajñapati Candraprajñapti and other works. 19. T. S. IV, 1. 20. Cf. 28 Bhedasamghātābhyām Cāksusah, and 32, Snigdharuksatvät bandhah. 21. Ed. by Kapadia Vol. I, p. 261, 320, Vol. II. p. 101 etc.
28 The Text of the Tatvārthādhigama-sūtrāņi
The Journal of the University of Bombay 1935

Page #343

Phonetics and Phonemics in Historical Linguistics-I
During the last few years the methods and techniques of descriptive linguistics are more frequently used in the historical study of languages. They are replacing the traditional methods of the neogrammarians which were so prevalent at the end of the 19th and the beginning of the 20th century. The changes which have resulted from the use of the new methodology in the field of historical and comparative linguistics are on the whole beneficial and have brought greater precision and rigour than was known before. The phonemic theory, which was the earliest to develop fully in the descriptive field, has produced a radical change in the aims and procedure of reconstruction of the earlier stages of languages based on comparison.
It is now generally accepted that what matters in the history of a language is the phonemic change and not a mere phonetic change, which may or may not amount to phonemic change. It is possible that a phonemic change may occur in a language even when the phoneme shows no phonetic change affecting it. It is also agreed that when we compare two or more related languages, what we aim at and succeed in doing is a reconstruction of the phonemes and the phonetic system of the parent language or an earlier stage of the language studied, and not its exact phonetic makeup. In fact, it is quite possible to refer to the phonemes of the parent language with purely arbitrary symbols like x, y, or z, which would not commit us to any phonetic values to be assigned to them.
These are some of the essential ideas of the new approach to historical linguistics and being essentially true, the comparativist has to admit them whole-heartedly. But can the consideration of the phonetic values of the reconstructed phonemes be quite irrelevant to historical linguistics? Will it not be necessary for the historical linguist to try to ascertain their phonetic nature. with all the evidence available to him? This will depend on the nature and amount of evidence available and at best he will be able to reconstruct the phonetic values of the phonemes only approximately. This is generally con

Page #344

Phonetics and Phonemics in Historical Linguistics-I
335
ceded -by all. Most linguists would choose such symbols for their reconstructed phonemes as would indicate their phonetic values also, rather than work with purely arbitrary signs. What is important is to know how far the phonetic aspects of the phonemes are essential, in theory, for reconstruction as such and, in a wider context, for the proper understanding of the history of the language.
: The only way in which we can trace the history of a language or a group of related languages is to compare the phonemic systems of the language at two different stages in its growth, which is admittedly continuous. In the comparative field, we compare the phonemic systems of a language with the phonemic system of the lost parent language which is reconstructed with the help of other languages of the same family. Naturally, the main purpose of this procedure is to make understandable the continuous change that has occurred in the language, as far as our sources allow us to do it. Obviously to trace the gradual manner in which the sounds of the language have changed and its sound-system is modified, is more likely to help our understanding of the change than ideas of a sudden replacement of one phoneme by another or the transformation of one phonemic system into quite a different one. What we do, when we compare the phonemic systems found at different stages in the life of a language, is to compare the initial point and the end result of a number of changes (may be wholly or partly phonetic) that are going on all along throughout the period, and we do not follow the changes themselves. Even in the case of a phonemic change involving no phonetic change, it is patent that somewhere else in the language a phonetic change has occurred in that some phoneme is either added or lost. An explanation of this type of change would naturally require an insight into the other phonemic changes which are based on real phonetic changes.
An actual example may help us to understand what is meant by the contention that phonemic changes are made more comprehensible by insisting on the continuous phonetic changes underlying them. This is also likely to bring out one peculiarity of phonemic analysis, which is probably the cause of concealing the continuity or gradualness of change in language.
The phonemic analysis of the vowels of the Old Indo-Aryan (DIA) offers no great difficulty. We have three vowels i, u and a which occur both short and long, and length is phonemic. /iti/ contrasts with siti/; /uru/ with /ūru/ and /vara/ with /vāra/. We have two mid vowels e and o which occur always long and two diphthongs ai and au. A number of alternatives are

Page #345

336
Amrita
available for their phonemic analysis. Some of them like considering i ua and length as phonemes or i i, u ū, a, ā as distinct phonemes differ only trivially. The only important point to be decided is whether to consider e and o as (1) unit phonemes or as (II) vowel clusters. If they are considered as unit phonemes we have to consider whether ai and au should be now regarded as also (la) unit phonemes and think them as long forms of e and o or (b) think them as clusters leaving the system assymetrical. If we consider e and o as vowel clusters we get a symmetrical system and a neat relation of short and long between e and o and ai and au. A number of considerations would favour such an analysis. A close parallelism of this type can be seen in the Sandhi rules of the language : eka + eka = ekaika, when analysed as /aika/ + /aika/ = /aikāika/ is exactly like /deva/ + /avatāra/ = /devāvatāra). The pattern of derivations also suggests a similar interpretation: né + ana : = nayana when analysed as /nai/ + /ana/ = /nayana/ is exactly like /vad/ + /ana/ = /vadana/. The ablaut variations are also similar: padya is to pada as vaidya is to veda which can be phonemesised as /vāidya/ : /vaida/. We have the further evidence of the Prātiśākhyas which considere and o as dipthongal in nature. Hence the phonemic system of the vowels of the OIA can be represented as follows:
short
long monophthongs /i/ /u/
lū/ /a/
// diphthongs /ai/ /au/ ai/ rāu./
There are some six vowel phonemes and four clusters acting as syllabic peaks.
A typical Middle Indo-Aryan dialect (MIA) would show the following vowels : i, u, a, e, o all of which are phonetically both short and long. But we also know that e and ē and o and 7 do not contrast because the short vowelse and o occur only before a geminate consonant or a consonant cluster (in a closed syllable) while the long ones can be found before a single consonant or in an open syllable. They are thus allophones of each other : velli but kēli, pomma but lõma etc. The other vowels show a phonemic contrast based on length : vaya and vāya; piya and piya; suya and sūya. Hence we have a phonemic system of the following type :

Page #346

Phonetics and Phonemics in Historical Linguistics-i
337
337
// / /u/ lül /e/
10/
/a/ // eight phonemes on the whole.
This system assymetrically because the two mid vowels are indifferent to length. We have a choice to put them with the short or the long series, leaving a gap in the other. Because the long allophones have greater freedom of occurrence we may choose to put them along with i ū and á and present the system as follows:
short
long // /u/ W /ū/
lē/ //
/a/
// We now compare the phonemic systems of the OIA and MIA. We note a great change between the two and a sudden jump or a drastic modification of the vowel system: While the system of the short simple vowels remains the same, the whole system of the diphthongs disappears. The system of the long vowels gets modified by the addition of two more phonemes lē/ and /o/ and there is a corresponding gap in the series of the short vowels. One gets the impression of a good deal of sudden change in the system, which is likely to remain unexplained.
We, however, know what has happened to the vowels of the OIA when they pass on to the MIA stage. Very few phonetic changes have occurred, and all of them are the result of the same basic tendency which is perfectly understandable provided we are willing to look at them in their concrete phonetic forms. Following a drift, which has its beginnings even in pre-Indo-Aryan stage, the changes are all in the way of reducing the length of vowels by one mora under certain definite conditions, thus reducing the long diphthongs to short ones and the short diphthongs to monophthongs in the OIA stage. The same tendency has changed the short diphthongs of OIA to monophthongs in the MIA and all vowels are shortened before a consonant-cluster or in a closed syllable. We can tabulate the changes as follows: Pre-OIA OIA
MIA
o oro au
o oro
au

Page #347

338
Amrita
10
ē or e o oro
au
10
If we state the changes in a phonemic form we get:
OLA
/ai/ Māu/ /ai/
MIA /e/ 70/ /e/ /0/
/au/
What has happened is clear. The phonetic pattern of the OIA has been shifted to the earlier stage much nearer to the I-I of the traditional view. Naturally there is a greater gap between it and the system of the MIA. This is so because the phonemic analysis tries to reveal the system or the structure of a language which is sometimes better represented by an earlier or a later stage in its actual history. It thus implies some amount of shift in the stage in the course of history of the language which is likely to make the understanding of the gradual change difficult.
A similar result follows a comparison of the two vowel systems of the OIA and MIA if we trace the development of each vowel separately. In phonemic terms: OIA
ΜΙΑ -/i/
-/u/
/u/— lū
-/a/
//
-
/a/ lā/ /ai/y /āi/ /au/
//
lo/
lus
Here three types of changes are involved : long vowels have become short; short diphthongs have become monophthongs and long diphthongs have also become monophthongs phonemically of the same type.

Page #348

Phonetics and Phonemics in Historical Linguistics-I
339
In phonetic terms:
OIA
MIA
if ara a e e
au
Here we have only two types of changes : long vowels have become short and diphthongs have become monophthongs. In the phonemic statement we have to keep apart the changes of the vowel clusters from those of the simple vowels, while in the phonetic statement we can say that ai becomes ē and au becomes 7 and along with other long vowels like i ū and a they are shortened in a closed syllable. There is thus greater economy and a more accurate picture of facts.
OOO
Phonetics and Phonemics in Historical Linguistics-I
Indian Linguistics 1958

Page #349

Phonetics and Phonemics in Historical
Linguistics-II
Phonetic distinctions often admit of gradations and a phone can be said to possess more or less of a particular phonetic feature. This makes a . clear-cut distinction between two phones often difficult and consequently a systematic and schematic statement becomes either impossible or uncertain. This is the basic reason for the uncertainty which attaches to the principle of phonetic similarity between sound segments. To the extent one invokes this principle in phonemic analysis, this indefiniteness is likely to find a place in the final result as well.
On the other hand, the phonemic distinctions always admit of a decision of the either or type. Two phones are either the same phoneme or are phonemically different. The question of more or less is either impossible or meaningless. This has the advantage of making a clear-cut and precise statement of the phonemic system of a language possible, and as shown by Trubetzkoy, dialect boundaries based on phonemic comparisons can be definite and clearly separated.
This insistence of phonemics on the systematic analysis of a language, however, produces some difficulties in historical linguistics, which aims at following the concrete changes in the history of a language. It is well known that the different phonetic changes which a language undergoes in the course of its development can be classified as either phonemic or non-phonemic. This means that some phonetic changes amount to phonemic change in a given language, while others do not and therefore remain as mere phonetic changes. Phonemics either takes no note of this latter type of change or cannot do so for lack of evidence to ascertain them. It is even suggested that as long as the phonetic changes do not produce any phonemic change, they are unimportant and linguistically irrelevant. Minute changes in the actual production of the phonemes in a given language are of no value for the

Page #350

Phonetics and Phonemics in Historical Linguistics-II
341
structure of the language and the same is true of phonetic variations which . may have occurred in the production of phonemes at different times, provided they have no effect on the phonemic system taken as a whole, in a wider sense, so as to include the distinctive features of phonemes, their number and arrangement.
By its insistence on a systematic analysis, phonemics thus takes note of only such facts as admit of a systematic analysis to form part of it and neglects other facts which cannot be so handled. This is likely to lead to the neglect of some amount of linguistic material which essentially forms part of the language and plays an important role in its history. It is bad theory and practice to regard a priori something as irrelevant or outside the scope of linguistic considerations. Only by working out its effects and following its nature can we conclude that a given thing is either relevant or not relevant.
The procedure of following the history of a language by noting the phonemic changes which occur in the development of a language, naturally uses the phonemic systems of that language at different stages in its growth and indicates such changes as are revealed by the comparison of these phonemic systems. And in most cases we expect this procedure to give us a fairly accurate picture of the history of that language. One important aspect of this history pertains to the question of the relative chronology of such phonemic changes. A good account of the history of a language ought to make clear this aspect, as far as the evidence permits us to know it.
In the majority of cases, phonetic changes are the basis of phonemic changes which result from them. While these phonetic changes are gradual and continuous, the resultant phonemic changes may appear, at a given moment, as a sudden jump or shift due to either a loss of a phoneme, its merger with some other phoneme or its split into two or more phonemes. The procedure of phonemic analysis requires that these changes be clearly separated from the earlier stage where the necessary conditions for such changes are not present. In all cases the loss or rise of a new contrast is the necessary condition for phonemic changes to occur. This situation is likely to lead, in some cases, to obscure the relative chronology of the different changes, which is, thus, a distinct disadvantage of this procedure. We shall consider a few concrete cases to see what is involved in this limitation.
In the development of the Indo-Aryan languages from the IndoEuropean base, we have two phonemic changes which are closely connected with each other. The three original short vowels a, e, o have all become a

Page #351

342
Amrita
in the Indo-Iranian stage and continue to be so in the old Indo-Aryan. Moreover, the original IE series of velars (including the labio-velars) has developed two types, one palatal and the other velar, the first of which is later represented by affricates in Sanskrit. The explanation of the rise of the palatal series (č, j) from the original velars (k, g) is sought in the effect produced by the following front vowels like e, as against the back vowels like a or 0. It is also known that the change of the velars into the palatals preceded the merger of e with a along with a similar merger of o with it. This sequence is necessary simply because, if the vowels are regarded as merging earlier, the very reason for the development of the palatal sounds would be lacking. The question of the relative chronology of the mergers of o and e with a need not concern us here.
The phonemic statements about these changes and their comparison will bring out an important limitation of the procedure. The two divergent sounds of the velar stops, one more advanced before e (say k”) and one more back before a and o (say k) would not be phonemically distinct as long as the difference between the vowels following them remains phonemic. The phonemic split between these two as // and /k/ can be regarded as having occurred only when they contrast in the same environment, which in this case would be the same phoneme /a/ coming from all the three sounds of the earlier stage. We may graphically refer to these changes in their phonetic concreteness as follows:
III [ke]
[k'e-1] [ko]
[ko-]
1
[ka]
When we try to put these changes in a phonemic form, we have an alternative analysis available for the second stage. One way would be: /ke/ ke/
/ca/
/ko/
/ko/
/ka/
in which the first two stages are identical. Another possible way of analysis would be :
/ke/
/ca/
/ca/
/ko/
/ka/
/kd/
in which case, the last two stages become identical. The difference

Page #352

Phonetics and Phonemics in Historical Linguistics-II
343
between the two ways of looking at the facts is something as follows. In the phonetic analysis we consider the change of palatalisation as preceding the change of the merger of e and o, while in the phonemic analysis we find that both the changes, the split of the stop phoneme and the merger of the vowels, occur at the same time, either between stages II and III or between stages I and II according to the manner in which we choose to phonemicize the second stage. In other words, the relative chronology of the two changes gets obliterated or obscured.
The development of the Slavic languages offers us with another similar case. As far as the contrast between palatal and non-palatal consonants at the end of the word is concerned, we may distinguish three stages in the phonetic development in the Slavic group. We may set up a simplified picture of the following type : Pre-Slavic Common Slavic individual languages
[-ti] [-tu] .
[-tu] It is obvious from this picture that the development of the palatal consonants is anterior to the loss of the two jers coming from IE short vowels . *i and *u as the effect of the vowels on the previous consonants must
presuppose their presence. The relative chronology of these two changes is fairly obvious and certain.
[-]
When we phonemicize the data, we can represent it in the following manner :
/-ti/ /-tu/
/-tů/ In this analysis it is clear that both the changes become synchronous and we cannot antedate one as against the other. The contrast between the two phonemes // and /t/ is dependent on the loss of the vowel phonemes and in any case not earlier to it. It is possible to analyse the stage of the common Slavic in a slightly different manner :
7-ti/
/-ta/
/-tu/ /-ta/
/t/ Here we make the distinction between [i] and [ŭ] allophonic and that between /t/ and /t/ phonemic. But even in this analysis the two changes,

Page #353

344
Amrita
emergence of contrast between /t/ and /t/ and the merger of [i] and [u] into one phoneme become simultaneous. It must be noted that the merger of the two, i and ů as one phoneme say /a/ or with zero is a matter of indifference as far as the relative chronology is concerned.
We may take a third example from the Romance field. We have both direct and indirect evidence to attest the following phonetic facts about the vowels of Classical Latin, early vulgar Latin and the early stage of continental west Romance.
EVL
CWR
[i:]
11)
[e:)
[e:)
[e]
[E]
(a:)
[a:)
[a]
回 回 回 回 回 回 a
[o]
[o:]
[o:)
[u]
[U]
[u:]
[u:)
The sequence of changes is clear. Classical Latin vocalism developed a further distinction of quality for the front and back vowels in such a manner that shorter vowels were lower than the longer ones. Later on the quantitative difference was lost and only the qualitative difference was kept along with the merger of i and e and u and o. Even here the length must have been lost before the phonemes merged as is seen in the case of a : and a. Still later a quantitative difference developed according to the place of occurrence of the vowel, whether in an open or a closed syllable. Thus the probable sequence of changes is (i) development of the qualitative difference in vowels other than of maximal opening, (ii) loss of quantitative difference, (iii) merger of vowels, (iv) emergence of new quantitative difference.
If we put the data in a phonemic shape we get the following picture :

Page #354

Phonetics and Phonemics in Historical Linguistics-II
CL
I
/i:/
/i/
/e:/
/e/
/a:/
/a/
/0/.
/0:/
/u/
/u:/
Ila
/i/
/i/
/e/
/ɛ/
/a:/
/a/
/0/
/0/
/u/
/u/
EVL
IIb
/:/
V
/e:/
/e/
/a:/
/a/
/0/
/0:/
/u/
/u:/
CWR
III
345
V
/e/
/ɛ/
/a/
/a/
/0/
/u/
If we phonemicize the EVL as stage Ila we get the relative chronology as (i) loss of length and qualitative difference together (ii) merger of i and e:, a, and a and u and o:. The still later development of length, being allophonic, will not find place in a phonemic statement. In any case the first two changes become simultaneous and their relative chronology is obscured. If we accept analysis IIb, we find no change between stages I and II, but the two simultaneous changes, loss of length and development of qualitative difference, occur between stages II and III, thus again obscuring their relative sequence. From these examples it is clear that the effect of phonemic analysis, as the basis of comparison, tends to make some successive changes simultaneous and thus obscures their relative order to some extent.
If we apply the procedure of internal reconstruction to the phonemic system of a language, there may arise a similar difficulty in our conclusions and statement of changes, obscuring the real facts to some extent. The past participle of Gothic furnishes us with an example of this type.
In the case of weak verbs of Gothic, we find a participle which shows the suffix with t or which alternate with each other: @aurfts, but kunes, nasies etc. It comes from the Germanic forms ta, ea and *da, all traceable to IE to Gothic possesses /t/ and /8/ as two phonemes and as they show in this morpheme a regular alternation dependent on the nature of the preceding phoneme, we can reconstruct a single original morpheme with the help of the procedure of internal reconstruction. As e appears to be restricted phoneme of the two in this alternation, not occurring after a fricative like s, for h, we can set up the original morpheme as ea and make a statement

Page #355

346
Amrita
of the change that becomes t after these fricatives while it remains unchanged everywhere else. The change of Germanic to d before s may be left out of consideration here. It is also clear that the absence of after these fricatives leaves no scope for contrast between 0 and t in this environment and hence it is equally possible to reconstruct the phoneme o in all the places, which will leave no occasion for a morphophonemic alternation. It is presumed that on the basis of phonetic similarity, the sound after the fricative is identified as phoneme /t/ occurring elsewhere in the language. The relative order of changes would be then:
Germanic
Gothic *to
*(a) after fricative
>(a) elsewhere But other comparative evidence shows that the actual changes were something as follows :
IE
--**a
Germanic
Gothic *0a or*da after vowels (a) *to
*ta after fricatives — +t(a) Here the split is shown as occurring between IE and Germanic. In the earlier statement, with the shift of the split to a later stage, we have to postulate a change of 0 to t after fricatives, which has obviously no basis in facts.
We may therefore conclude that a phonemic analysis, by its insistence on the system of the language alone, may in some cases obscure the relative order of phonemic changes, which is due to the fact that it takes note of changes not until they find an indication in the phonemic system of the language.
ODO
Phonetics and Phonemics in Historical Linguistics-II
Indian Linguistics 1960

Page #356

General President's Address
Let me begin by thanking you all for electing me the general president of the 32nd session of the All-India Oriental Conference, an honour which is as much pleasing as demanding for me. The choice of a topic for an address to be delivered to a learned gathering like this, in which each one is an expert in one or more branches of Oriental Learning, thus covering a vast field, proved to be a very hard task and to satisfy you all even to a small extent will prove harder still. A topic which will interest most of you was not easy to get and after some deliberation I took courage to formulate a new subject to which I would like to give the name of 'interpretation' which, along with categories, Aristotle made the basis of his logic.
All branches of Oriental Studies; and more particularly the classical ones among them, involve two types of approach which are known as philological in a wider sense and linguistic in a more restricted sense. They are otherwise known as classical philology and historical linguistics, which together cover nearly all the usually pursued Oriental Studies. The subject of interpretation is common to both and forms their basis, being a necessary prerequisite. Without a valid theory and a careful practice of the process of interpretation, neither classical philology nor historical linguistics can make progress and will not have any acceptable validity.
Both these studies were carried on from early days in India as well as in Europe, first without any sharp demarcation between them, but becoming more and more distinct in course of time. Today they show difference in their scope, aims and methodology. Classical philology has a larger scope and deals. with all aspects of culture which developed in classical times and is limited. only by the cultural unity which it represents. This unity is also reflected in a given classical language and hence we speak of Sanskrit, Greek or Latin philology and sometimes even in a reconstructed language as in the case of Indo-European philology. An intermediate position is occupied by studies like

Page #357

348
Amrita
Germanic or Slavic philologies. On the other hand linguistics is not limited to a given language or even to a group of languages but deals with language in all its forms. Its scope includes all languages and even some semi-linguistic systems because it aims at an explanation of the nature of language as such and to clarify the basic assumptions underlying it. If philology leans more towards historical studies, linguistics is inclined more towards axiomatics.
The aim of philology is the study and description of the cultural aspects of civilization like religion, ritual, philosophy, literature, law and other scientific subjects based on the oral or written records bearing on them and then to explain their peculiarities and trace their origins with the help of other classical civilizations, mostly by the process of comparison and contrast. On the other hand, the aims of linguistics are to ascertain the systematic nature of language, describe the past and present languages of the world, study their growth with the historical material and comparative method and clarify its nature with the help of sciences like psychology, anthropology, logic, acoustics and others in so far as they relate to the formation of thought and its expression in language. The more recent tendency to formulate universally valid principles of language has made it necessary for the linguist to utilise all possible languages without any restriction..
The methodology of linguistics is well developed both in the field of synchronic and diachronic studies. It takes for granted that the meanings of utterances are already known or given and concentrates its attempts on the way in which they are expressed in words, in other words what their correspondances are. Of the physical and mental components of language, the first is studied in detail with the methods of modern science. But some amount of analysis of the thought-contents is also necessary before the two are related to each other in a systematic and coherent manner, in accordance with the basic structure of the language used. On the other hand, philology takes for granted that the principles of the grammar of the classical languages are already known and concentrates more on the nature, contents and the value of the classical works in order to clarify and assess the different aspects of its culture. Naturally it uses to a great extent the comparative method by drawing on all the cultural and historical disciplines.
These two methodologies are not mutually exclusive but depend upon each other and get intermingled at every stage. This is clearly the source of the antagonism which is seen between classical philology and linguistics during the last 150 years. With the rapid growth of linguistics, it was taken

Page #358

General President's Address
349
as a rival science by the older philologists. The situation was aggravated by the way the linguists handled the classical material with imperfect preparation and to the neglect of the long-standing tradition. But a little thought will show that cooperation more than confict is what is needed and is desirable. In fact both studies have their own proper fields in which they are supreme.
ey have also a large common field which must be treated uniformly with a methodology proper for it. To this we should now turn under the name of 'interpretation'. Without a correct and unbiased interpretation of the written documents, neither philology nor linguistics can do its work properly and arrive at valid results. .
The process of interpretation of a classical text demands the following preparation :
1. constitution of a correct text 2. thorough understanding of the language used 3. detailed knowledge of the vocabulary of that language 4. historical background of the text studied
5. its cultural milieu, and 6. its evaluation and explanation in the light of all available evidence.
These aspects of interpretation have been studied from early days and each one has developed into an academic branch with its own name. To obtain a correct text needs the use of the method of textual criticism. The science of grammar takes note of the second, while the semantic analysis of words develops into the science of lexicology. The evolutionary aspects is taken care of by the study of the history of that subject and the material contents of culture form a part and parcel of the cultural milieu. What is, however, not generally realised is the fact that each one of them draws on the results of the one or more of the other aspects and cannot carry out its task without constant help from them. Their interdependence is obvious. Hence the further necessity of adjusting their claims with each other by the process of proper evaluation which is the concluding aspect of interpretation.
Instead of building up a complicated theoretical scheme of the mutual inter-dependence of these aspects and the way in which they should be adjusted and revalued, it may prove more useful to analyse a series of examples where their apparent conflict is met with. This is also imperative because more than two aspects often get involved and hence a neat

Page #359

350
Amrita
correlation between any two at a time cannot be easily attempted.
A correct text is essential for interpretation because it is often possible to assign some plausible sense to an incorrect text which thus gives rise to a ghost word or a fictitious meaning. An unusually striking example is supplied by the Dictonary of Buddhist Hybrid Sanskrit in its entry andakāstha. The phrase asarataratvaṁ kāyasya can be understood to some extent by the idea of fragility and hence anḍakāṣṭha can be made to yield the meaning of an egg-shell which satisfies the condition. In the text, however, which runs as hirodakasikatäpindairandakästhebhyo'pi asärataratvam kayasyävekṣya the words are sikatäpinda and erandakaṣṭha and not sikatäpinḍaiḥ and andakaṣṭha. The eranda tree is well-known for being asara 'without a core.' The ending of the ablative plural also confirms that at least three objects are referred to, which are hirodaka etc. While hirodaka itself remains obscure, the other two are often used in such a context.
When a text gives no satisfactory meaning it is often emended so as to give some sense. Such an emendation cannot be regarded as certain as long as some textual evidence is not available. An emendation is suggested on the ground that at least the meaning intended is certain and the doubt pertains only to the wording. Hence in such a case the confict is really between two meanings-one of which is said to be intented and, the other not. But there is no unique way to resolve the conflict of such a nature and hence some external evidence must be sought to decide whether the unusual meaning of the words actually used is intended or not. A good example of this type is supplied by a passage from the Isvarasiddhi of Yamunācārya (p.77) which runs adhiṣṭhānakriyäkarmabhūtasya dehasya adhisthätṛdehanupraveśo 'nupapannaḥ, yugapadekakriyāyām ekasya karmakartṛtvavirodhät. Thinking that dehanupravesa, in its literal sense, does not suit the context, the editor of at recent edition (Madras 1972, p.96) change it to adhiṣṭhātrdaśānupravesa, the intended sense being in their own words 'cannot intelligibly be included in the very body (?) of the definition of the controller'. But this summary decision bypasses an important stylistic consideration.
A peculiar feature of the style of the late Nyaya works is the tendency to shorten words and phrases, particularly long compounds which are required to be repeated again and again with the addition of a word or two. The shortening is done in such a way as to remind the reader of the earlier statement which was already given and discussed in full. This is abbreviated as ending in the last word after which the addition begins. Another type of

Page #360

General President's Address
351
abbreviation is one in which only a part of the expression is used and the suppressed part has to be supplied, at least mentally, by the reader. A constant use of this procedure makes the word actually used get an additional meaning of the elements which are dropped. This may lead to some kind of misunderstanding and unnecessary correction. If one compares expressions like laksaņaśarīrapravistasya (Manidarpana 3.6), śabdapadasya lakṣaṇapraveśe (4.3), janakatvasya laksnāpraveśāt (Maņisāra 8.9), ata eva bahuviseșanapraveće’pyevam na syāt (10. 23) one can easily see that the full expression laksanaśarirapraveśa which means 'addition of a word to the scope of the definition given' gets shortened into laksanapraveśa, śarīrapraveśa or simply praveśa, all having the same meaning as of the original expression. Adhisthātrdehānupraveśa really stands for adhisthātrlakṣaṇadehānupraveśa which is clear enough and no emendation of deha into dāsā is called for. In fact, a traditional commentator, who has a feeling for such usage, paraphrases it as adhisthātrtāvacchedakaghatakatā, which means the same thing. Based on such stylistic considerations interpretation must make a distinction between the literal meaning of a word and its pregnant meaning due to the context. Such abbreviations also occur in the ordinary usage of a language though rarely. A descriptive name for a porcupine in Sanskrit is śvāvicchalākā 'whose darts can pierce a dog. It is shortened into either svāvidh or salākā which latter develops into Marāthī sālvī.
An example which involves the separation of words and also consideration of grammar and rhetorics is found in a verse attributed to a
poet Dharanidhara and cited in the Subhāsitaratnakośa 18.21 and · Saduktikarnāmrta 89.30 both early anthologies. In the edition of the HOS it
is printed as follows:
punaruktāvadhi vasaram etasyah kitava pasya ganayantyāh iyam iva karajah ksīnas tvam iva kathorāņi parvāni ||
Prof. Ingalls translates 'Look, deceiver, as she counts the days ever and again, the finger is as thin as she, the knuckles hard as you.' He has added no notes by way of explanation.
This stanza, apparently simple and straightforward, implies a number of problems in its accurate interpretation. The main point is how to divide the phrase punaruktāvadhi vāsaram with which it begins. The edition separates vāsaram from adhi. One can also separate it as punaruktau and adhivāsaram. The reading of the text punaruktāvadhi admits of two divisions : the whole as one compound word or two words punar and uktāvadhi. Interpretation has

Page #361

352
Amrita
to decide which one is valid and under what conditions.
Proceeding from what is clear to what is doubtful, we may first note that the speaker of the stanza is a Dūtī because it is included in the section called dūtīvacanavrajyā. It is addressed to a Nāyaka as is clear from the use of the vocative kitava. She says that her nail is worn out (or finger is thin) like this one and the finger-joints have become stiff like you (i.e.the Nāyaka). That the poet intends a comparison between the Nāyikā and her nail or finger is not acceptable because a comparison between a person and his limb is regarded as banal. Because the joints are compared with the Nāyaka and the nail is compared with the Nāyikā, it follows that they belong to the Dutī and not to the heroine. As these are the effects of constant counting it is also necessary to take the word ganayantyāḥ to refer to the Dūtī and not the Nāyikā. Once this is accepted, the gist of the stanza becomes clear. The Dūtī is speaking to the Nayaka and telling him that he should see for himself how her nails are worn out and the finger-joints have become stiff by constantly counting the repeated utterances of the Nāyikā ( at the same time spreading her hands before him). What is counted now depends upon the interpretation of the remaining words of the stanza. To take the word vasara as being counted goes against its singular number and leaves the remaining expression unexplained. Taking it avadhi as one word leads us to imagine that it is the avadhivāsara which her lover had often given to her but not kept. But it cannot be connected with ganayantyāh which refers to the counting to many days. We are thus left with the other alternative that adhivāsaram is an adverb meaning day by day and punaruktau is used as a locative of reference, the repetition of her lover's name uttered by the Nāyikā. Whatever interpretation is regarded as more appealing, there is no doubt that they are far apart from each other and a choice has to be made by taking into consideration the difficulties in grammar and poetics which each one encounters.
Compared with these small points, interpretation of books dealing with Vedic ritual is a matter of greater importance and also of nicer judgment. The present-day performance of the ritual acts may or may not agree with the ritual texts due to changes and developments towards greater elaboration and adjustment to newer practice and ideas. On the other hand the meaning of the ritual text remains to some extent uncertain due to changes in the meaning of words used therein. This situation is further complicated by the traditional interpretation which often attempts to adjust the meaning of words with a later practice and thereby conceals a change in the ritual. No simple

Page #362

General President's Address
353
mechanical procedure can be used to detect such mutually dependent changes and there is usually no external control available. Hence a careful assessment of the evidence must be made to get at the real facts in their historical perspective. A limited problem of such a nature is supplied by the use of the word gataśri often found in ritual literature. Its importance lies in its bearing on the question of deciding whether the Vedic Aryans used some kind of thanks-giving offering for favour obtained.
The word occurs both in the Vedic literature and in Classical Sanskrit but with quite different and opposite meanings. Examples from classical works like the Rāmāyaṇa 5.18.25 rāmo gataśrīr vanagocarah, Mahābhārata 1.192.23 gataśrīr api pārthivaḥ, Buddhacarita 8.3. gataśrīr iva tena varjitaḥ, Kumarasambhava 13.14 gataśrir vanam, Bhāgavata 3.2.7 gataśrīsu grheṣu and many others make it obvious that the word is used in an unfavourable sense 'one who has lost his prosperity, glory or beauty'. It is thus a bahuvrihi compound with a nisṭhā as its first member with an udätta on it, as per the rules of Panini 2.2.36 and 6.2.1. Only works dealing with the Purvamimämsä appear to use it in an exactly opposite meaning 'one who has attained. prosperity and consider it as a technical term in this sense. Khandadeva in his Bhāṭṭadīpikā (3.175.21) writes gatasrtitvaṁ tantraratne praptaśritvam iti vyākhyātam. Obviously they have taken the word and its meaning from works dealing with the Vedic ritual.
In the Vedic literature the word occurs from the Taittiriya Samhita onwards and is found in the Maitrayani and Kathaka, in the Aitareya, Taittiriya, Jaiminiya and Satapatha Brahmanas, in the Taittiriya Aranyaka and most of the Srautasütras. Its negative form agataśri, however, occurs once each in the Taittiriya Samhita and the Śrautasutras of Baudhayana and Äpastamba. The reading in the Tattiriya Aranyaka 10.39.1 is uncertain and is often given as āgataśrī which is against accent and must be corrected to gataśrī. No independent occurrence of agataśrī is noted.
Even with the reading gataśrī the commentators on the Vedic works and the Śrautasutras continue to assign it the meaning 'who has attained prosperity' by explaining it in a number of ways: gatā prāptā śrīr yenāsau, gatā śrir yam, gatä präptä śrir yasya as a bahuvrihi compound and rarely as śriyam gataḥ as a tatpurușa. Thus the most usual way to explain the word is to assign to gata the meaning of präpta, which is valid only for a few senses of gata and not all.
It is now necessary to examine the contexts in which the word is

Page #363

354
Amrita
Than
found in the ritual literature to ascertain whether its normal meaning suits them or not. Most of the occurrences of the word are confined to a few definite situations. Some specific oblation, or some alternative procedure is enjoined to one who is a gataśrī and as such they are not decisive in fixing the meaning. We are told that a gataśrī should milk the two front udders of the cow for the Agnihotra, twentyone sāmidhenī verses should be recited for him, he should offer a cake on eleven potsherds to Indra or one on eight to Agni, he should use a level Devayajana, he should maintain the Ahavanīya fire continuously, he should offer to Mahendra as his special deity, or the sacred waters should be brought in the sacrificial hall through the eastern entrance.
There are, however, a couple of passages in the Taittiriya Samhita which should help us to decide or at least to surmise what the original meaning of this word was : 2.1.3.4 reads indrāya vrtrature lalāmam prāsrngam à labheta gataśrīḥ pratisthākāmaḥ and 7.2.7.2 śukrāgran grhņīta gataśrīḥ pratisthakāmah. These are alternative practices and are meant to bring pratisthā if the performer is a gataśrī which here should naturally mean who has lost his prosperity and hence wants to regain it. A comparison of the second passage with that of Kathaka 30.3 further points out that very often the word pratisthākāmah is dropped in this context, being implied by the situation. Thus whenever an act is prescribed to a gataśrī it is meant that he will regain his prosperity which he has lost. It is the interpretation given by the later commentators which makes scholars take the word in the unusual sense. In the absence of any remark on the part of these commentators as in the case of Taittiriya Saṁhitā 2.1.3.4, KEITH naturally translates it as 'he whose prosperity is gone 'but in other places of the same Samhitā, where the remarks of the commentators are available, he translates who has attained prosperity : Even Bhattabhāskara is not quite unambiguous when he says gataśrīḥ kecid āhuḥ śuśruvān grămanī rājanyo veti. He is clearly referring to Tais. 2.5.4.4 where the context speaks of an agataśrī. What is meant is that a gataśrī should not mean any one who has lost his prosperity but only one who is either a learned Brahmin, a village headman or a petty chief, who will be entitled to the various alternatives to regain their lost prosperity.
This passage is important because it also indicates how the tradition has come to assign the unusual meaning to this word in the ritual. It states that one who is not a gatasri should not sacrifice to Mahendra. These three types of persons have Mahendra as their deity. He who over-sacrifices (atiyajate) will not reach prosperity but will come to harm. To explain this passage tradition makes two assumptions viz. gataśrī applies to these three

Page #364

General President's Address
355
types of persons only and that no one other than these is allowed to sacrifice to Mahendra. Both of them are of doubtful validity. In reality the word gatasri should restrict the members of these three types by qualifying them as those who have lost their prosperity. If any one of these three, who has not lost their prosperity, should again sacrifice to Mahendra he is not justified to do so and commits ati-yajana.
An external control in this case is supplied by the exact meaning of the verb ati-yaj-. The traditional explanation ätmiyām devatām atikramya yajate has no authority and is unique. There is no other passage in Vedic ritual where it is used in this sense. But another form of the verb yaj- with the prefix anu is used to mean to perform an anuyaga and with ati further added to it, it is used repeatedly in the Satapatha 3.8.5.1 where the context is quite clear to mean 'in addition, more, excessive', the other meaning of ati. Hence svām devatām atiyajate means 'he sacrifices to his deity excessively or unnecessarily because he is in no need of pratistha as he has not lost it. We thus come to the conclusion that gataśrī has the same meaning both in Vedic ritual and classical Sanskrit, and there is no trace of a thanks-giving offering in the Vedic ritual.
A change in practice can also be detected by the use of a careful. interpretation. The traditional ritual method of immolating an animal victim involves no use of a sharp instrument to kill it. In its place some kind of strangulation is used to achieve the same end. The description of the texts on this point of ritual are found in the TaiS. and SatBr. but are not sufficiently explicit to make the mode of immolation quite clear. Hence it has led to some amount of speculation which involves both historical and linguistic evidence for the justification of a given interpretation. Linguistic taboo and a shift in the practice are involved and hence a proper interpretation of the evidence becomes crucial.
Three important moments in this part of the ritual must be carefully considered. They are the typing of the victim to the sacrificial post, the way in which the animal is put to death and the cutting up of its parts for the purpose of offering them to the deities. As regards the first, the description of the texts like the TaiS and the Śrautasūtras of Baudhāyana, Āpastamba and Kätyāyana is not very precise and hence it is difficult to visualise the details of the act. What is clear is the fact that this method was different from the one used in the ordinary killing of an animal for secular use, which consisted of throwing a rope round the neck of the victim over its head when it faced

Page #365

356
Amrita
.
the butcher. The ritual procedure required that the rope be first tied to the right front foot of the animal which faced the Adhvaryu and then taken across the front of the animal, carried over the head and then encircling the right horn be fastened to the sacrificial post. What is not clear is whether the cord went behind the left horn or in front of it. That it was tied to the right foot and the right horn is clear and its purpose can only be to have the animal and its head remain attached to the post even when it is severed. The view of Sālīki was different, according to which the rope was fastened round the neck like a noose. This was however too close to the ordinary way and hence was rejected.
As regards the act of killing the victim, linguistic taboo appears to have led to the use of euphemistic terms for the act itself. The Vedic works use a number of verbs for it : alabh- to take hold of, anu-bandh- to tie down, samay- to pacify, sanjñapa- to make it agree and tamay- to make it suffer or languish. These works avoided the use of verbs like han- to kill, māray- to make die, chid- to cut etc. It must be emphasised that this refers to the avoidance of the use of words which directly mean to kill and does not refer to the act of dealing with the victim. When the word refers to the act independently, the texts use words like han-, visas- etc, Hence this is purely a linguistic taboo and not necessarily a taboo on the act of killing the victim by beheading.
The practice followed by the tradition is stated clearly first in the SatBr. 3.8.1. It lays down that the victim should not be killed by striking it with a frontal bone (kūtena), nor behind the ear, its mouth should be taken hold of and it should be made to suffer. It should not be allowed to make a noise and if it does an expiatory offering is prescribed. The text also states that the animal is loosened from the yūpa and taken to a place where the śāmitra fire is established. Kätyāyana follows this procedure closely. An alternative procedure is also stated and attributed to śālīki, in which a noose (veska) is used to kill the animal. It is thus a continuation of the mode of typing the victim to the yūpa. After the animal is killed, its various parts are cut off with a sharp instrument called śāsa and then offered to the deities. In this all authorities agree.
A change of practice appears to be involved. The alternative procedure of using a pāśa or veska appears to keep the original sequence of acts like tying the animal to the post and killing it with a noose, for the very. reason that this alternative was not favoured in practice and hence was not subject

Page #366

General President's Address
357
to change. The description of the SatBr. can be easily seen to have been expanded or modified by inserting a number of other acts between the preparation of the weapon for killing (section 5) and its use to kill the victim (section 14-16). Two linguistic peculiarities appear to indicate the original procedure. In the Samhitās and Brāhmaṇas this act is called krūrī-kr- and apahan- repeatedly, which refer to a bloody way of killing. Other texts also use the word krūra in the same sense. Secondly the use of the verb sam- in its causative or transitive form appears to show a difference in usage. While samayati is consistently used to refer to the act of appeasing or pacifying or expiating a lapse, the form śamāyate is confined to the act of killing. The PW is right in setting up two roots of the form sam- one in the Atmanepada with the meaning of to exert, to labour and the other in the Parasmaipada in the sense to pacify. However PW is wrong in deriving the noun samitr from the second, which really comes from the first. Samāyate is thus a denominative formation based on the activity of of the samity which is to kill the victim involving great labour. We must thus give up the idea of an euphemistic use of sam- to quieten and hence to kill, however attractive it may seem to us. One is reminded of the philosophical etymology of the word education from ex ducere 'to draw out' though really it comes from the commonplace verb edere to eat or nourish.
One
One of the peculiarities of Sanskrit literature is the way in which it gets divided into various schools and disciplines. A writer of a particular branch follows its tradition and even when he deals with different branches he keeps them carefully apart. The result has been that words used by one school are also used by others but often with different ideas associated with them. Thus the technical vocabulary of Sanskrit develops a bewildering polysemy and it is necessary to distinguish between its divergent uses. A word like pudgala in Buddhism has a different meaning from the one given to it in Jainism. Though the word keeps its basic semantic component the same, and in this it is helped by its etymological meaning, its associations differ and we cannot shift them from one school to the other. While giving a formal definition of sacrifice the Mīmāṁsakas and some srautasūtras discern three elements in it which are named as dravya, devatā and tyāga (cf. Katya Śs.1.2.1). The last, being a common word, has different associations in different schools. In the legal terminology as seen in the Dharmasūtras, it means giving up one's claim of legal ownership of an object when some gift is made. In this sense it is used in the definition of a sacrifice where the sacrificer gives up his ownership of the dravya which is offered to the deity,

Page #367

358
Amrita
so that it now becomes the property of the deity in the fullest sense. This is what the usual formula like agnaye idam na mama means. To read in it the philosophical idea of renunciation of one's desired fruit of an action done, the theory of nişkāmakarman as expounded in the Gītā as Prof. STALL does, is not justified. Even the grammar of the sentence, the dative of the deity and the genitive of the owner, is against it. Hence it is strange to read STALL's justification that 'the reader should not jump to the conclusion, therefore, that unfamiliar forms are mistakes, even if elementary rules of language seem to be disregarded, as for example in agnaye idam na mama this is for Agni, not for me.'
Interpretation of some details of Vedic ritual for the purpose of drawing conclusions of anthropological or ethnological nature is a matter of great importance but at the same time a matter of great caution. Two types of approaches are met with here : either the material is used to draw historical conclusions about their origins or an attempt is made to ascertain their structural value in the ethos in which it is embedded. This runs to some extent parallel to the synchronic and diachronic approaches in linguistics. Like them, these are not really opposed to each other and they must be synthesised by proper evaluation. The weakness of the early anthropologists like Bacchofen, De Morgan, Westermarck etc. was to draw hasty historical conclusions without properly interpreting the evidence in its structural context. More recently attempts have been made to interpret the Vedic ritual and sacrifice in such a way as to suggest that their origins are to be found in a series of political and social conflicts and violence among the early Aryan tribes. For this purpose the evidence of punarādhāna and samsava is used and interpreted in a peculiar manner.
The ritual of punarādhāna is explained by Dr. KRICK as an indication of the agnostic situation that prevailed before it. The establishment of the fire is made to represent the enthronement of a victor in a conflict and the punarādhāna is a kind of replacement of the earlier yajamāna and his hotr by their successful rivals by the process of elimination. The period between the two is one of long conflict or war. This appears to be nothing but an interpretation based on the famous custom in the shrine of Nemi in the Alban mountains near Roma as investigated by FRAZER in his famous work the Golden Bough. A statement about the destruction of a bhrātrvya by the utterance of the vasat-call, thought to be a vajra, can hardly bear such an explanation as against the explicit description of the practice at Nemi.

Page #368

General President's Address
359
The use of Samsava for such a purpose is equally unconvincing. Samsava is a situation which arises when two Soma pressings are performed in close proximity by two different sacrificers. It is briefly referred to in the Kathaka and in the Taittiriya and Jaiminiya-Brahmanas. But a more detailed. statement is found in the Pañcavimsa and it is summarised in the Drāhyāyana and Baudhayana Sutras. The gist of the whole act is that when two sacrificers want to perform the Soma pressing for Indra close to each other, it is thought necessary to win Indra over to one's side before the other party wins him. The means used are a very early recitation of the prataranuvāka in the midnight itself, the use of specific epithets to attract Indra, the use of special Sastras and Sämans etc. A more elaborate sacrifice is preferred to the sacrifice. of the other party. The Säman called Taurāśvasa and the Sastra called Vihavya preserve the memory of a priestly rivalry between Turāśvas and Pāravatas, and between Jamadagni and the other sages. As it is impossible to think that there could be only one Soma pressing in the whole country at a given time, samsava is defined as involving the use of two close places except when they are separated by a big mountain, a large river, a distance which a chariot can cover in one day, a strong wind, or a boundary line be.ween two friendly countries, or when no enmity is felt between the two sacrificers. All these details lead one to believe that it is due to the intention of having no overlap of two sacrifices and to avoid the resulting confusion. To think of a conflict between two kingdoms or tribes and on its basis to suggest that such conficts were the norm, which gave rise to all sacrifices and further to interpret such a situation between the rituals of Agnyädhäna and Punarådhāna is something which was never envisaged by the tradition. The details which. are preserved do not justify such a construction. Of course the question of the origin of sacrifice as an institution in conflict and violence or otherwise is at different one which may or may not be true. What is clear is that this detail of Vedic ritual has nothing to do with it.
Now and then interpretation faces a peculiar problem in that a well known word with a precise meaning is found used or explained by a reputed author in such a way that one has to consider it as incorrect or even wrong. In a situation like this interpretation has to take a further step of finding out the reasons which prompted the author to do so, because a lapse or wilful distortion is excluded. Such a case is supplied to us by a verse in the Samksepaśäriraka (1.259) and its explanation by Madhusudanasarasvati in his commentary.
While explaining the usual definition of Brahman, Sarvajñātman

Page #369

360
Amrita
discusses the various types of lakṣaṇas and then tries to prove that the highest reality cannot be considered as being different from either the material world or the individual souls. After having shown that the material world is not different from the Supreme Soul he states the prima facie view that it may be at least different from the numerous individual souls or Jivas because of the relation of adhiṣṭhātṛadhiṣṭheya between the two. All the other commentators and an alternative explanation of Madhusudana himself take it to mean the relation of niyamyaniyämaka in which the adhisthätṛ stands for the niyamaka and the adhistheya for the niyamya. This is in full agreement with grammar because the suffix tr can only be used in an active sense. However Madhusudana gives his first explanation in a different manner. He takes adhisthātṛ in the sense of adhiṣṭhāna with the suffix in the sense of karman-adhikarana or bhava. This is grammatically unjustifiable, because tr can never give such a meaning. It is under the influence of his philosophical views, that the relation between the Absolute and the Individual Soul is that of adhyasa, which is also a kind of ädhäradheyabhäva, that he flouts grammar and puts forth the equation adhiṣṭhātṛ is adhisthäna. For him the philosophically acceptable position must supersede grammar. But the task of interpretation cannot be confined to this judgement only, It must also explain this indifferent attitude towards grammatical correctness which is also reflected in the writings of Kumārila. In the present case two considerations may have weighed with Madhusudana in admitting such an equation. The word adhiṣṭhāna, though formally a bhava formation, has been used in other senses like that of karman or adhikaraṇa, both of which become identical with the use of adhi with the verbs sthä-, si- and äs-. But on many occasions. in the philosophical writings adhiṣṭhāna and its abstract adhisthänatva are used in the sense of a controller and control. In this sense the equation is valid and hence it is taken as valid for the other senses of adhisthāna also. The second reason is the peculiar construction of the stanza of Sarvajñätman. The two types of causes of the Jivas are given in the order yoni and nimitta. All agree in taking yoni 'source' to stand for the upädänakāraṇa 'material cause' and nimitta has the usual meaning of an instrumental cause including the karty 'doer'. All the other commentators who take adhisthätṛ in the active sense connect it with the nimitta and the adhistheya naturally stands for the yoni. Thus the order of the two pairs in the stanza is not the same. Madhusudana bases his interpretation on the supposition that the two pairs are in the same order, leading to what is known as yathāsamkhyānvaya. He feels justified in setting up his equation on his basis. In the final analysis, the order of words in a Sanskrit stanza cannot be allowed to set aside the rules

Page #370

General President's Address
361
of grammar and hence Madhusūdana's priorities are not correct. A lexicographer is bound to assign the word adhisthātr an active meaning as being intended by the author Sarvajñātman.
In some cases the situation appears to have been reversed. The context is so definite and strong that it compels us to admit a meaning of a word which appears quite unknown and unjustified. It is only a complex psychological association which can explain it. The negative particles a has to be admitted in a word like agha as it occurs in the Abhidharmakośabhāsya on I. 28 in the sense 'very solid, impenetrable'. Yaśomitra in his Sphutārthā tries to explain it as follows : ati atyartham hanti hanyate veti agham nairuktena vidhinā atyarthasabdasya akārādeśaḥ krtaḥ hanteś ca ghādeśaḥ. That gh can take the place of h is well known and is due to the original sound which was a velar voiced aspirated stop. But that ati or atyartha should be substituted by a is not easy to comprehend. Nor can we set up another a as a prefix in this sense.
A possible solution to this problem lies in some kind of association between the two, the negative a ( coming from the sonat n ) and ati, usually with the sense of ‘more? A clue to this association can be found in the works on medicine like the Carakasamhitā or the Astāngasamgraha, the language of which often shows striking similarities with Buddhist writings in Sanskrit. For instance in the Astāngasamgraha (I. 27) we read samsodhanausadhāni āgneyamārutāni atisaumyāni atimātrāņi ca alpasattvam āturam alpabalam vātipātayeyuh. Here, and in other places, ati as a prefix has the meaning 'not' and hence atisaumyani means 'not mild, rather strong'. This change of meaning of ati is explainable as due to the other meaning of this prefix atikrānta 'getting beyond and therefore not what is expressed by the following noun or adjective. For example atimtyu means 'beyond death, deathless'. Once this association is formed in the mind of the writers they feel that both can be used in other senses as well and a develops the meaning of ati 'excessive'.
Interpretation is involved in the very process of assigning meaning to a given word. This is a process which can best be described as a series of adjustments, each better than the preceding one, leading finally to an acceptable position. The usual procedure used for the classical languages is to fix the meaning of a word by examining the contexts in which it is used. But very often the context itself depends upon the meaning we assign to the word, or it is not sufficiently specific to decide the meaning to the precision

Page #371

362
Amrita
needed. Hence other principles like etymology or relation to other words must be brought into play. This often leads to a better result but involves a number of other aspects of interpretation. Such a situation can be profitably discussed with a concrete example. A curious word like adhyāma is often used in a number of contexts which are not specific enough to decide its meaning. Apparently it is a negative form of dhyāma which itself is of limited use. The Kośas assign to it the meaning like a kind of grass' or 'name of a plant' which cannot be explained on the basis of its derivation. We can only divide it into a root having the shape dhyai- or dhyā- and a primary suffix ma. But no semantic relation can be established between them. Dhyama is used as an adjective in the sense of black, dark' and its derivative verb dhyāmi-kr- means to darken. This use is confined to the Buddhist Hybrid Sanskrit works like the Mahāvastu, Lalitavistara and Daśabhūmikasūtra. But in this sense its etymology is not clear. Dhyai- in its usual sense 'to think or meditate being useless, we may venture to set forth a new root like dhyām- and assign to it the meaning 'to darken or blacken'. But this is not possible in view of the fact that this is the meaning of its derivative verbal form dhyāmi-kr- because of the force of the rule abhūtatadbhāva necessary for such a formation. Hence it must be some activity which can be connected with the dark colour in some indirect way or may imply some figurative usage. Uses like (Mahāvastu 1. 42) dhyāmāni ca abhūnsuh mārabhavanāni, (I. 68) kramau muneh dhyāmatām upāgatau, (1.314) māraḥ svakaṁ balaṁ dhyāmabalam samjānāti are not such as to allow us to guess its basic or original meaning. Nor is its use as an attribute of an army reconcilable with its use as an attribute of a house or a foot.
The negative form adhyāma is often used by the Jain writes and it is always found compounded with a noun which it qualifies. In the Prabhāvākacarita we come across expression like dhiyam adhyamadhāmabhuh (said of Siddharsi 121.1) dadhyau adhyamacaitanyam (209.210) in the Jinaśataka expressions like adhyāmatejā ya jinavarah (1.15), tattvārtham satvaraṁ yaḥ tvarayatu sa gurur boddhum adhyāmārūpam (3.11), in the Samarādityasamksepa, nītaś ca candikāvesmamadhyam adhyāmamānasah (6.148). In all these cases the context may lead to a number of meanings like big, great, bright, clear, composed etc. because of its connection with dhāma, tejas, caitanya, upaśama, mānasa and rūpa but their interrelation remains vague.
A secondary derivative from dhyāma is found in dhyamala which, along with its negative form adhyāmala, is used in the Pratyabhijñāvimarsini.

Page #372

General President's Address
363
Things are described as having a dhyāmala or adhyāmala nature in the two evolutionary stages called Sadāśiva and īśvara (II.229.6); dhyāmala is also used to qualify the bhāvacakra which is antahkaranaikavedyam iva and unmīlitamātracittakalpam (II. 219. 5) which Bhāskara explains as asphuţa ‘not clear' which will make adhyāmala to mean 'clear'. They are also traslated as 'soiled' and 'pure'. In the classical works like the Nalacampū and Viddhaśālabhanjikā, dhyāma and dhyāmala have variant reading syama and śyāmala with the usual meaning 'dark, black etc.
When we consider that this word is mostly used in BHS, Jain writings, and by Abhinavagupta, it is but natural to suspect that the word and its derivative may be influenced by Pali and Prākrits. Again it is exclusively used as an adjective, while ma-formations are either action-nouns with both abstract and concrete meanings, or rarely agent-nouns. Sanskrit grammarians, however, preserve a ma-formation as a past passive participle used as an adjective kşama from the root ksai in Pāṇini 8.3.53 kṣāyo mah. Its negative form aksāma has a range of meanings which resembles that of adhyāma in many ways. In the jain works aksāma is used in the very books in which we find adhyāma and 'nearly in the same contexts. In the Prabhāvakacarita we read akṣāmacaratrapavitraḥ (said of Sūra 160. 19), prāguktanijavrttāntam sa ākhyād aksāmaduhkhabhrt (204.23), yatir eko yuvā tasmai śiksām aksamadhir dadau (54.19). In fact at 202.31 both the words are used side by side in
sūreḥ śrīhemacandrasya gunagauravasaurabham ākhyād aksāmavidyaugham adhyāmopaśamasriyam
Additional examples are found in the Prabandhakośa (89.4) subhiksam aksāmam, (15. 3) sanghasamakşam aksāmasvarena babhāņa, in the Samarādityasaṁksepa (1.324) akṣāmabhāvanah, Pāndavacarita (6.783) aksamadhih, (7.248) aksāmanidra, (9.46) aksāmabhujasthaman, Udayasundarikathā (126.4) aksāmavikrama etc. All these cases give us the general sense of big, great, bright, clear etc. as in the case of adhyāma. This should also include the use of the word aksāma in the Astāngasamgrah (I. 27) and Carakasamhită (I. 10. 6) in the sense of a medicine which makes one strong or fat. There is one more occurrence of aksāma in EI II. 415.24 akşāmārāmadāmā dharaņisurataruh where it means 'not dried up', a meaning which must be set apart from the others.
A comparison of the meanings of these two words adhyāma and aksāma would easily show that both have a major semantic component which is identical, while each one has its individual component as well. This

Page #373

364
Amrita
situation can be reasonably explained on the basis of that a Middle IndoAryan word like jhama is at the root of both these words, which is also the source of the verb jhämeti or jhämei. But this is not all. The doubt raised by Edgerton about the lack of semantic link must also be explained. The original meaning of the root kṣai-(kṣāyati) is no doubt to burn and its participle kṣāma means burnt. From this developed the two meanings black and dried up. All these senses are found in dhyama as well. In the sense dark it is often replaced in classical Sanskrit by the words syama and yamala.
Kṣāma developed the meaning 'weak, emaciated' first in its negative form as found in the Carakasamhitä and was later extended to the simplex as well. Thus developed the other shades of meaning not weak, strong, big, complete, full etc. That this line of semantic development started with the meaning 'dried up' is possible but some amount of influence from the other word kṣiņa is also present. In the Niryukti of Bhadrababu on the Uttaradhyayana, we find an early attempt made to associate adhyayana with akṣina which suggests the way this influence must have worked with the Jain writers. The meaning clear and loud are due to this influence. There is a parallel to this in the meaning of the Latin word serenus 'clear, bright, fair' etc. from the verb seresco to dry, but is probably to be judged as an accident. When we speak of the influence of one word on the other we reach the boundary-line between what can be proved and what is of the nature of an imaginative suggestion.
Before I conclude I find it necessary to add a few general observations, if only to avoid possible misunderstanding.
When interpretation is called a science it must be emphasised that it is a classificatory or taxonominal science and not prescriptive or predictive. Of course the demarcating line between a classification and an explanation is hard to draw and one merges into the other. I am reminded of my bewilderment some fifty years ago when I started reading the Jain canonical work Nandisutra which states se kim tam paccakkam. paccakkham duviham pannattam. indiyapaccakkham noindiyapaccakkham ca and so on to the very end of work. Today it does not appear to me so very incongruous. The Nyaya system distinguishes betwen uddeśa, lakṣaṇa and pariksă but rarely follows it in practice. They are really not so different or better not mutually exclusive.
In the field of interpretation we do not speak of true or false, not even of accurate or inaccurate but only of better or worse. Here the scale of evaluation is a sliding one and the results asymptotic. It is best to avoid

Page #374

General President's Address
365
words which smack of finality.
By the very nature of the case I had to take up for discussion a few examples which have come to my notice. Naturally the view of some scholars have come for criticism, which may be a little improper. But let me emphasise that no disrespect is implied and no finality is claimed for the interpretations offered. To do so would amount to self-contradiction.
Not all or even the major aspects of the subject could be illustrated in an address like this. I had to confine myself to a few examples to get a glimpse of the topic. Let me frankly state that there is nothing very new in all this, and I know all of you are also engaged in interpretation in your research in Oriental Studies. Let me recall at this moment the words of Jayantabhatta in his Nyāyamañjari :
kuto vā' nūtanam vastu vayam utprekşituṁ kşamāh which is not merely an expression of modesty but embodies a profound truth. I add to it a metrically inelegant second line :
pūrvasürivicārāņām viśleșa evą nūtanaḥ.
Let me conclude, ladies and gentlemen, by thanking you for giving me a patient hearing.
000
General President's Address
All-India Oriental Conference 32nd Session November 1985

Page #375

Indian Linguistics
Let me, at the outset, express my sincere thanks to the members of the Executive Committee of the All-India Oriental Conference for electing me to preside over the Section of Indian Linguistics. I feel myself hardly worthy of this honour. But the confidence you have placed in me and the assurance I have of your constant help and goodwill prompts me to undertake this responsibility. I shall do my best to discharge my duties in this regard and nothing more than this can be expected of anyone.
2. I imagine that I am expected to point out the main events in the linguistic studies in the country during the last two years, to indicate the present position of the subject, and, if possible, to outline the ways along which the future development of linguistics may proceed. It so happens that I have followed with some amount of interest the growth of linguistic studies during the last two years, and naturally I have formed some ideas and views of my own on this matter. I may be allowed to express them here for what they are worth. As all of us are keenly interested in the promotion of linguistic studies, you will kindly pardon me, if I may sound a bit personal and intimate in some places.
3. All of you will readily agree with me when I say that linguistics in our country is slowly coming into its own, and is getting a respectable and honoured place in the academic life, both in the Universities and outside, and there are brighter prospects for it in the near future. The reasons for this revival of interest, long after its period of commanding position in ancient days in our country, are well-known and need not be recounted here in detail. Suffice it here to refer to the main events of the last two years to refresh our memories in this matter. As a result of a Conference held at the Deccan College in 1953, an extensive programme of linguistic studies was set on foot at the Deccan College, with the liberal help of the Rockefeller Foundation of New York. This programme carried on, over a period of five

Page #376

Indian Linguistics
367
years, under the inspiring guidance of Dr. Katre, an intense activity of making the elements of modern linguistics better known to the younger scholars with the help of an eminent faculty of linguists drawn from all parts of the country and from outside. It also helped formulate more precise objectives of these studies and infused new life in the Linguistic Society of India, which has shown a phenomenal expansion during the last few years. A concerted effort was made to put linguistic studies on a firm basis in the academic activities of Indian Universities, and in this it was greatly helped by the Universities themselves and the Universities Grants Commission. There is hardly any doubt that in all these undertakings, the Project has met with conspicuous success, mainly due to the unfailing help from the senior linguists of the country, along with encouraging enthusiasm shown by the younger scholars. The Linguistic Society of India and a host of other academic bodies have taken a leading part in these activities due to the realisation that linguistic studies. have a vital and important role to play in our academic and national life. I need not go into the details of all this intense activity, as the two welldocumented reports published by the Linguistic Society of India, of the Conference of Vice-Chancellors and Linguistis and of the Blue-Print Committee. of the Society, give a clear and comprehensive picture of all these events and are surely known to all of you. We can thus look into the future with far greater confidence than was possible a couple of years ago, and we all rejoice in this rapid progress of linguistic studies in our country, as these interests are dear to our hearts. The work of a number of Summer Schools and Autumn. Seminars of Linguistics held all over the country has made this bright picture possible and has assured us of an equally bright future.
4. Before I proceed with the problems which inevitably arise with at rapid expansion of a subject, let me briefly report on the work done on Indian Linguistics during the last two years. This may help us realise the present status of the subject and the task ahead of us and may throw into bold relief some of the urgent needs of these studies, and the lines along which they may progress with profit and advantage. For obvious reasons, I may be permitted to make this report as brief as possible by confining myself to book-size publications. I must add a word of apology by pointing out that any omission in this matter is wholly due to the limitations of my interests. and lack of sufficient energy and industry needed to do this work thoroughly. On no account should an omission be considered as either significant or wilful or in any way reflecting on the intrinsic merits of those particular studies.
5. It may be convenient to survey the field of Indian Linguistics by

Page #377

368
Amrita
following the usually accepted chronological divisions of Old, Middle and New Indo-Aryan and Dravidian Linguistics with a glance at work done in the Austric and Tibeto-Burman families of languages. As regards the Old IndoAryan period, we have now a revision of the monumental Altindische Grammatik of Wackernagel, begun more than half a century ago, brought upto-date by the joint labours of Debrunner and Renou, who have added their unrivalled scholarship in Vedic and Sanskrit linguistics to this work in the form of additions and emendations which summarise in a masterly fashion the results of scholarly work over more than fifty years for ready and convenient reference, besides their individual contributions in numberless details. It is extremely unfortunate that Debrunner should have been snatched away by the hand of death before the completion of this unique historical and comparative grammar of Sanskrit, to which he devoted the major part of his life, as did , his teacher Wackernagel before him. We earnestly hope that it may prove possible to complete this work with the help of the material left behind, by the devotion of some younger scholar as his successor. The historical grammar of Sanskrit, Thumb's Handbook, which served generations of young philologists as an introduction to the subject, is now available in a thoroughly revised form, due to the labours of Hauschild, and is likely to help the beginners as a good introduction, as did the two earlier editions of the work. While both these works stress the Indo-European origins of Sanskrit, and are in the best tradition of comparative grammar, the regular history of the Sanskrit language has found in Renou a competent scholar to do justice to it. His 'Histoire de la langue Sanskrite is a work of great importance and a masterpiece of scholarship and lucid presentation, the work of an acknowledged master of the subject. We owe to him also four volumes of his monographic studies, entitled Etudes Vedique et panineennes, touching upon a number of points in the history of Sanskrit language. Moreover, the senior Sanskritists of India and abroad have continued their contributions to the Old Indo-Aryan Linguistics in the form of numerous articles published in various Oriental and Linguistic journals, too many to be enumerated here.
6. The revival of interest in linguistics at the different University centres has again brought into prominence the healthy practice of inviting a competent linguist to deliver a course of lectures on the subject of his specialisation. As a result of this, we have in recent days a set of lectures delivered by Dr. Sukumar Sen of Calcutta at the University of Mysore, which are published under the title 'History and Pre-History of Sanskrit'. Dr. Sen here makes an interesting attempt to put together all the features which can be

Page #378

Indian Linguistics
369
ascertained of the spoken form of Old Indo-Aryan. While dealing with the pre-history of Sanskrit, he also gives a brief sketch of the Hittite language which is mostly neglected in Indian studies and together with Tocharian ought to be given a more prominent place in our Indo-European work. Dr. S. M. Katre of the Deccan College delivered two lectures in Linguistics at Agra, which are published by the Munshi Institute of Hindi Studies and Linguistics. The second lecture gives a fresh evaluation of historical method in linguistics for etymological investigations in which, as the Editor of the Sanskrit Dictionary on Historical Principles, he is a recognised authority. He has also delivered a course of lectures on 'The Introduction to Modern Indian Linguistics with special reference to Assamese' at the Gauhati University under the Pratibhadevi Memorial Lectureship and we are awaiting their publication. In October 1958, Sir Ralph L. Turner was invited by the University of Poona to deliver a course of lectures on Indian linguistics in the newly established series in honour of Dr. P. D. Gune, a distinguished linguist of the last generation. A course of five lectures was given by Dr. Turner, the Nestor of Indian Linguistics, who followed the best tradition in historical studies of masters like Meillet and Bloch, and elucidated a number of highly important and difficult points in the growth of Indo-Aryan languages. The publication of these lectures will make them available to a wider circle of scholars who were unable to hear them in person, and are bound to inspire many a young scholar to follow in his footsteps. We are also awaiting eagerly his longpromised comparative dictionary of Indo-Aryan languages, which is bound to mark an important stage in the history of Indian lexicography.
7. Though no major work in the field of Middle Indo-Aryan Linguistics published during the last two years is known to me, articles dealing with various aspects of the subject continue to be published by scholars both here and in foreign countries, many of which are valuable and thought-provoking. It may be worth mentioning that the famous Prākrit Grammar of R. Pischel is now available in an English and a Hindi translation and will certainly prove to be of immense use to many Indian scholars who were unable to use it so long. As a primary source of great value, mention may be made of the publication of the Minor and Major Rock Edicts of Asoka found at Erragudi long ago, by Dr. D. C. Sircar in the latest volume of Epigraphia Indica. The continued publication of Prākrit and Apabhramsa works should make it now necessary to have more comprehensive surveys of Prākrit dialects and their vocabularies.
8. In the new set-up of our national life, it is but natural to expect

Page #379

370
Amrita
that there would be greater emphasis on the analysis and study of modern Indian languages, both Indo-Aryan and Dravidian. This work is being carried on along two different lines of approach. The older historical approach continues to be cultivated with good and valuable results, while the more recent descriptive approach is being slowly but steadily developed in analysing Indian languages of all types, literary or spoken, cultivated or tribal, dialects or more wide-spread forms of speech, and socially marked different idioms. In this field, we owe to Dr. Vishwanath Prasad and Dr. Sudhakar Jha, a small linguistic survey of local and social forms of speech of the Sadar division of Manbhum and Dhalbhum areas of Bihar, which collects a good deal of interesting material including some tribal languages as well and supplies brief sketches of these dialects. An interesting attempt is made here to record the texts phonetically in a modified Devanāgarī script. We now possess, as a.. publication of the Central Government, a new basic grammar of Hindī, prepared under the supervision of a committee specially appointed for the purpose, which is in many ways an improvement on the earlier grammars of Hindi of even larger scope but, on the whole, it moves on more conservative lines and appears to neglect more recent methods of analysis. In this connection, mention may also be made of the scheme of the Ministry of Education, Government of India, to prepare a phonemic and morphemic count of major Indian languages, based on a count of one hundred thousand items selected on a random-sampling procedure. A few of these analyses are ready and others are making progress, and the information so collected will be not only useful for the purpose of inventing or improving rapid scripts or typewriters, but will also be of great value for linguistic studies and practical application of linguistics to the problems of language-teaching. A similar scheme of preparing common vocabularies of Indian languages, to judge from the two or three such bilingual lists available, appears to be not wellconceived and properly executed, and probably competent linguists will have to be consulted on this matter before acceptable results can be obtained. A number of these, on historical lines, tracing the growth of many Indian languages and dialects are prepared at different Universities, and it is highly desirable that they should be available in print for the use of interested scholars.
9. Analyses of some modern Indian languages on descriptive lines are. recently published. Kannada gets a brief and lucid description in W. Bright's work 'An outline of Colloquial Kannada' published by the Deccan College. Dr. Kelkar has published in a mimeographed form a detailed phonological and

Page #380

Indian
Linguistics
53
morphological descriptions of Marathi and Burling's description of Garo is in the press. We know that similar discription of Tamil, the Bāngru dialect of Hindi are ready and Assamese is being worked out at the Gauhati University. Dr. Krisnamurti of Andhra University has prepared an exhaustive historical and descriptive study of Telugu verbs, and other studies of Indo-Aryan, Dravidian and Austric languages are in preparation. Prof. S. Bhattacharya has given us a brief sketch of a new Dravidian language called Ollāri and we have an interesting description of the Toda language of the Nilgiris from the pen of Dr. Emeneau. The Comparative Dictionary of Dravidian Languages by Emeneau and Burrow is ready and nearing completion in print and will be soon available to all, thus giving a new impetus to the study of comparative Dravidian linguistics. I may also mention a brief but lucid exposition of comparative Dravidian phonology which Dr. Emeneau prepared for the Summer School of Linguistics at Coimbatore. The Annamalai University has also mimeographed the numerous scattered articles on Dravidian linguistics of the late L. V. Ramaswami Ayyar, which are valuable but difficult to get.
10. Recently a kind of survey of major Indian languages is published from Madras under the title 'The Languages of India - a kaleidoscopic survey edited by a group of three scholars with contributions from a large number of writers. While fully realising the importance of having brief sketches of different Indian languages, their scripts and their contribution to literature for a better appreciation of regional languages and literatures, the work is anything but satisfactory. Many of the chapters are written by persons whose competence in linguistics remains, at best, doubtful. The book by its numerous inaccuracies is likely to produce more harm than good. If anything, the work makes a more reliable survey of modern Indian languages a far greater necessity than ever before.
11. Of the numerous aspects of modern linguistics, we find Dr. Ganeshsundaram's work 'Distribution Characteristics of Speech elements in Tāmi! (1959) touching on a neglected side of linguistic studies, while quite a different aspect has been dealt with in great detail by Dr. Devasthali in his work 'Mimāṁsa, the Vakya-Šāstra of Ancient India.' The Dhvani-vijñāna of Prof. Dhall gives a detailed account of phonetic theory in Hindi and is a welcome addition to phonetic studies written in Indian languages.
12. Before I close this brief and inadequate survey of work done in this field, let me refer to two publications of recent dates. We are glad to have a new Oriental journal published from Holland called 'Indo-Iranian

Page #381

372
Amrita
Journal', which is in its third year of publication. In this periodical important contributions on early history of Sanskrit and Indo-Iranian period have appeared and continue to appear. Secondly, we have the publication of the four special volumes of Indian Linguistics by the Linguistic Society of India, to honour the names of Drs. Chatterji, Taraporewala, Bagchi and Turner, all well-known names in the field of Indian Linguistics. By their scope, variety of topics, new and more vigorous treatment of numerous aspects of Indian Linguistics, they vividly remind us of the vitality of these studies and assu their rapid growth. As remarked by Prof. Allen, along with the recent volumes of BSOAS, these are the volumes in which most of the important contributions to Indian Linguistics are to be found.
13. It may be appropriate to make a passing mention of the fact that the Linguistic Society of India has allowed affiliation of a number of Linguistic Circles, functioning in different places in the country and fostering interest in these studies at the seats of various regional Universities. The Linguistic Circle of Delhi follows its own independent course and publishes a small volume called "Transactions' once a year, in which valuable articles are found. One may seriously doubt whether the research output of this young science needs two organs of expression at this stage, and a close, co-operation among all concerned should not be found difficult, and may prove beneficial.
14. Let me now give expression to a few personal observation on linguistic studies in the country, which, at this moment, may be said to stand at cross-roads. Some definite decisions will have to be taken by the interested scholars for their further developments. The tradition of linguistic studies of India, of recent past, obviously derives from the 'London School which itself has drawn and synthesised the attitudes of the neo-grammarians of Germany and the sociological approach of French comparatives. More recently, the activities of the Language Project at the Deccan College, introduced the techniques of so-called descriptive linguistics from the United States, which, in turn, has drawn largely from the Prague school of phonologists, and from the Copenhagen school in Denmark, besides its own attitudes developed by Boas, Sapir and above all, by Bloomfield. We have also a few linguists who would like to follow the lead of Firth and the ideas developed by him. We are thus in a position of advantage in possessing followers of different linguistic traditions. With the characteristic Indian attitude of broad-mindedness, tolerance and love of truth, we can draw great profit from all these different approaches. But the younger student, who is newly attracted to linguistic studies is likely to be swayed by one attitude or the other, which he gets to

Page #382

Indian Linguistics
373
know 'first and is in the danger of developing some kind of apathy, if not hostility, to the other views. If I am not mistaken, now and then, one appears to notice a slight inclination on the part of the budding linguist to consider the descriptive or structural approach of a specific type to be the only proper linguistic approach and to feel a bit cold towards the historical studies of earlier tradition. An undue importance attached to the type of terminology one is accustomed to use, is likely to be a source of misunderstanding and may indicate a lop-sided development. If the elderly linguist may find it unusual or unnecessary to adopt the more recent and to some extent novel terminology, one fails to see why the younger scholar should not take more seriously to historical studies, along with descriptive work.
15. In spite of apparent differences in the terms used and some amount of basic differences in the attitudes to be adopted towards the study of language, which are ultimately traceable to what general theory of language. one holds and what one wants to get out of the language-study, a broad basis of common agreement is present among all these views, and it may pay us to emphasise these common features and try to appropriate them as the sound basis, of our own attitudes. Granting that the phonemic or phonological approach is essentially correct and renders great service in revealing the structure of the language analysed, one sometimes feels a bit embarrassed at the fact that other aspects of language, which is truly a manysided and complex object of study, are either getting neglected or do not receive proper emphasis. The studied brevity which is developed in describing the different languages, dialects and even idiolects, makes for difficult reading and understanding, and one may be pardoned for doubting that a brief sketch of a few pages can really describe a rich cultural language of long history. One may not find it also easy to agree with the application of the principle of economy in linguistic analysis, by which one feature of the data is called significant and all others are regarded as superfluous or redundant because predictable. The initial reaction against phonetic niceties has often led to complete neglect of the phonic material, which can be neither logically justified nor pragmatically likely to yield useful results. The constant demand to keep levels apart in language-analysis often leads to the setting up of more and more such levels and the organic unity of the language appears to be put to an undue strain, under which it may break into pieces. Nor can one go to the other extreme and decide to neglect the obvious differences between different layers of a language under the spacious plea of unity and the whole approach. No one doubts the basic distinction which de Saussure drew

Page #383

374
Amrita
between language and speech, but this in itself is no reason to turn linguistics into a kind of mathematical computation or algebraic notations and formulae, and on no account should we cut ourselves loose from the philological studies, which ought to form the sound basis of all linguistic analysis. Secondly, we ought not to limit the function of a language to its representative aspect only, setting aside both the expressive and appellative aspects, which are as important as the other one, and in a large context of philological studies and practical teaching, may even claim greater importance than the mere representative function. Some of the newer descriptive approaches appear to describe language on the basis of marginal cases as the real system of language in the interest of consistent theory and comprehensiveness, thus carrying the view that language is a perfect system beyond legitimate limits and neglect to take note of the fundamental fact that language is essentially full of redundancy. A less economical and broader approach will make our analysis fuller and more in agreement with the actual use of language in life and save us from a lot of mere formal analysis. What appears to be a more legitimate attitude on the part of an analyst is to take into consideration all the features which his material offers and then to make a sincere attempt to ascertain the different functions which they serve in the system of the given language. We may like to call this the principle of 'primacy of function as against the principle of primacy of form or expression.
16. The more formal the approach of linguistic analysis, the more chances there are of getting alternative systems, simply because, what the analyst has to do is to lay down some basic principles and to hold fast to them, whatever the consequences. With a different set of such operational rules, we can arrive at a different analysis of a language, and even a different emphasis may lead to divergent results. To argue that all these alternative analyses are equivalent because mechanically convertible, is simply to point out that the basic material for them all remains the same, and no one doubts it. This becomes an acute problem in morphology and we find that each analyst sets out to analyse and describe with his own set of rules. He may agree to admit a zero morpheme or not do it, he may set up discontinuous morphemes or refuse to do anything with them, he may believe in empty morphs or vehemently oppose them, he may think replacive morphemes allowable or consider them as processes in disguise and so on. Even a morphemic contrast may prove to be a matter of choice and the sweet will of the analyst. This brings linguistic analysis very close to subjectivism, though carried on under the name of objective approach. Those who believe that

Page #384

Indian Linguistics
375
non-uniqueness is not a virtue in itself, may like to bring in historical considerations to decide among different alternatives, at least to prefer one as against the other, if, otherwise, both are equally correct.
17. This brings us to the question of the basic relation between historical and descriptive linguistics. Not only is there no conflict between the two, but that each separated from the other is to a certain extent a falsification of facts. Just as no historical growth of a language is understandable without realising its structure at different periods and ascertaining the effects of a given change in the system as a whole, it is equally true that no systematic description of a language is possible with- out taking into consideration its history. In fact, there is nothing like a stable synchronic systems; but always a system in the making. Both the aspects are inseparable. Problems of the type, co-existent systems, multiple systems or status of borrowed words thus lose much of their point and importance.
18. Against this background of general linguistics, Indian Linguistics has two main tasks to carry on. It must continue to carry on the investigation of the historical development of Indian languages, old and new, with a more refined methodology obtained from recent advances in descriptive linguis and describe systematically the vast amount of language-material available in the country, which, for the major part, is not even poorly described. And at the present moment, the linguist cannot afford to neglect the practical application of his science to the teaching of languages, which alone is going to convince the man in the street of the usefulness of his discipline. Let me conclude by expressing the hope that Indian Linguistics will continue to progress with rapidity, but will also strike a line of its own and develop its own basic attitudes, which would be more suitable to our material, our mental make-up and our traditions.
OOO
Indian Linguistics
All-India Oriental Conference 1959

Page #385

Dictionary of Sanskrit on Historical Principles
1. INTRODUCTION
During the 19th century historical linguistics enjoyed great prestige and as a result of it, historical dictionaries loomed large both in linguistic studies and lexicographical work. In the study of the history of a language no sharp distinction was made between what may be called its history proper and its pre-historic stages. The spectacular results of the comparative method in historical linguistics gave greater prominence to etymological dictionaries which probed into the prehistoric stages of a language than to historical dictionaries proper which confined themselves to its attested periods of growth. The work of etymological dictionaries involved comparison of different but related languages which was far more attractive and offered more scope to imagination, while the task of tracing the growth of the vocabulary of a given language during its historical period involved greater labour and the results were not so striking. It did not allow much scope for imagination and did not reveal sweeping changes in semantic development, but involved a good deal of purely philological work. The result has been the preparation and publication of very few historical dictionaries which confine themselves to the attested periods of languages. Among them it is only the Oxford English Dictionary (OED) of the Philological Society, originally called the New English Dictionary, which stands alone in its majestic form to represent this type of lexicographical work, and that too because it was conceived before the rise of comparative linguistics into great prominence. All other attempts, started later and modelled after it, have remained either incomplete or are yet in preparation. These include the Latin Thesaurus, the Middle English Dictionary, the Greek Lexicon, the Irish Dictionary, the Dictionary of the Older Scottish Tongue, the Dictionary of Old Church Slavonic, the Pali Dictionary, the Sinhalese Dictionary, and the Dictionary of Classical Arabic. In comparison, the etymological dictionaries are many and

Page #386

Dictionary of Sanskrit on Historical Principles
377
are either completed or are in progress. These are often revised when need arises. Even for Sanskrit, there are a couple of etymological dictionaries which are complete and others in progress, but there is no historical dictionary worth the name.
Large-scale dictionaries of the usual analytical type include a fair amount of historical material. And the same is true of etymological dictionaries. When the lexicographer illustrates the different usages of a word, he naturally draws on the existing literature of different periods and when a large number of citations are to be included, it is but natural that they should be arranged in some kind of a chronological order. As the dictionary is meant to be used as a help to read older literature as also to ascertain the current usage, a distinction has to be drawn between the obsolete and the current meanings. If one has to trace the etymological connection of a word with those in other related languages in order to ascertain its original form and meaning, it is necessary to make use of the oldest and earliest attested forms and meanings in these languages for comparison and this implies following their history to some extent. Any large-scale dictionary conventionally also includes some indication of the derivative history of a word, its cognates, striking changes in its meaning and the range of its occurrence. It will thus be seen that the various types of dictionaries planned on a large scale tend to overlap and so cannot be strictly kept apart. So also a historical dictionary or more usually, a dictionary on historical principles. 2. THE PRINCIPLES OF A HISTORICAL DICTIONARY
. It is not easy to lay down the basic principles of a historical dictionary in the present state of lexicological studies due to the paucity of historical dictionaries. Many schemes for the typological classification of dictionaries have been suggested. Those of Malkiet and Shcherba are well-known. Their views on historical dictionaries differ from the practice of the OED, the only completed historical dictionary available today. No clear picture of the nature of a historical dictionary and the basic principles to be followed in it emerges from the existing literature. A start has to be made in this regard.
As the historical grammar deals with the structure of a language from an evolutive aspect, so should a historical dictionary deal with the vocabulary or lexicon of a language from a similar point of view. The various aspects of a language, like its phonology, its grammar and its vocabulary, are integrated to such an extent that a clean separation between them is not always possible except as a matter of convenience. In separating the material for a dictionary

Page #387

378
Amrita
from that of a grammar of a language, the concept of closed and open-ended sets is useful. Whatever aspect of a language, which cannot be delimited by means of a rule even of fair complexity and needs enumeration, should be the legitimate field of lexicology and hence find a place in the dictionary of a language. All the historical aspects of this material then should form the raw material of a historical dictionary. Obviously the lexical material will form the core of such a dictionary and other aspects of language like its sound system, its morphophonemic complexity and other grammatical and stylistic features will be of a marginal nature. The historical dictionary, then, takes for its study the formal and semantic aspects of the lexical material of the language from an evolutive point of view. This is not much different from the view expressed by the OED in the words, "the aim of this dictionary is to present in alphabetical series the words that have formed the English Vocabulary from the time of the earliest records down to the present day, with all the relevant facts concerning their form, sense-history, pronunciation and etymology" (Preface). A more recent description of a historical dictionary is given by Prof. A. J. Aitken, the editor of the Dictionary of the Older Scottish Tongue, in the following words : "A historical dictionary such as the Oxford English Dictionary (OED) and the many other large academic dictionaries of national languages, has the following defining characteristics. Each of its word-entries is accompanied by a comparatively copious selection of illustrative quotations and references; these are set out so as to display, as fully as may be, the distributions of the word in question in various dimensions—the linguistic dimensions of form, sense or meaning and habitual collocation and the syntactical situation and the extra-linguistic dimensions of time, region and genre. Thus by its arrangement of illustrative quotations a historical dictionary purports to display all the available distributional information about each word, except the purely statistical; and even this can be displayed or at least implied in a rough and ready way, for example by the proportions of references given under the different heads or by means of brief explicit statements. The definitions and descriptive notes, which are also a normal feature of such dictionaries, may be reg-arded as fulfilling a somewhat secondary purpose, that of sign posts or labels to the particular subset of quotations which follows!.”
Malkiel, is more sceptical about the nature of a historical dictionary.. He remarks, "the logical avenue of approach to the unadulterated diachronic view is the historical dictionary, provided its materials are so ordered as to bring out plastically the dynamics of lexical development, with heightened

Page #388

Dictionary of Sanskrit on Historical Principles
379
attention to the succession and mutual compatibility of meanings. For not a few languages, even though subject to philological inspection, no dictionary aiming uncompromisingly at this particular goal exists or is clearly in sight at present. In the existing historical dictionaries all too frequently the meaning listed first is not the oldest on record nor indeed the oldest by the standards of reconstruction, but the one most familiar to moderns or held 'fittest' to have acted as a semantic fountainhead."
Rey-debove makes an important distinction between two types of dictionaries in relation to the history of a language. He points out that in one type the lemmata include the vocables current in the different periods of the language and all of them are never in use at a given time. The main purpose of such a historical dictionary is to supply the reader with an instrument of decoding and such an attempt, illustrated to some extent by the French dictionaries of Littre and Robert, is far from the structural concept of a language. An integral historical dictionary of a panchronic nature does not exist for any language today. The other type is historical in the sense that the earlier and obsolete senses of current words are also included and the historical information is considered useful for a proper understanding of the words of a language at a given period and is found necessary for their proper use. Whatever the nomenclature, a Sanskrit dictionary on historical principles can only be of the first type because its primary purpose is certainly to supply an instrument of decoding and research.
To sharpen the profile of a historical dictionary still further it may be useful to see what goes into it and what is kept out of it from the facts pertaining to the history of a language. The whole field of sound changes and regular phonemic changes may not find a place in the dictionary of such a type, but will constitute what is usually called diachronic phonology. There is even less reason for including changes in the graphic forms of letters and different ways of spelling words at different times as does the OED. The various ways of writing words, like the gemination of consonants in Sanskrit clusters or the indications of hidden length in Latin closed syllables should be included in its scope to the extent that these cannot be predicted. Individual changes of sounds of lexical items which cannot be stated by rules should also be included, being historical facts of the language not finding a proper place in its grammar. But sandhi rules of Sanskrit, a morphophonemic rule of devoicing final consonants in German or Russian or the addition of an unrounded high back vowel at the end of borrowed word in Tamil are not facts of a historical dictionary, though exceptions to them have to be

Page #389

380
Amrita
onary.
included. The shape of the negative particle as a- or an- in Sanskrit before consonants and vowels respectively has no place in the dictionary but the origin and scope of words where na- occurs before consonants like na-ga, nacira, na-pumsaka etc. must be indicated. Shift of accent in course of time like pañcan and pañcan ‘five', paśu and paśu 'cattle', dāśan and daśān 'ten', mati and māti 'devotion' which cannot be stated by a rule and shift of accent like brāhman devotion and brahmān 'priest, varā 'suitor and vāra 'choice which is regular but limited to a few words must be included. All such historical facts form what can be called the history of the shape-changes of lexemes and constitute one aspect of the historical dictionary.
The same approach holds good in the case of grammatical facts. Regular facts of grammar, whether synchronic or diachronic, have no place in a historical lexicon. But the change of a noun like 'cow into 'kine' in the plural at an early stage and its change to 'cows' at a later stage has a place in historical dictionaries, as do the shift of a German verb from a strong to a weak conjugation in its history. Similar changes in Sanskrit like those of gender in saru 'arrow' (from Fem. to Mas.) in the conjugational class (from krnoti to karoti) or in the declensional form (from Madhvā to madhunā) or the loss of various types of infinitives etc. must find a place in a historical dictionary of Sanskrit. So also changes in the syntactic structure including the formation of compounds provided they are of an individual nature.
But all such facts are of a marginal nature for a historical dictionary, the core of which is formed by a systematic study of the changes which affect its stock of words during its life, both in their formal and semantic aspects. We can consider them best as lexical changes and semantic changes and a historical dictionary has to deal with them for the whole stock of the words of the language during the period of time covered by it. This can only be done against the background of a general conceptual scheme which underlies such changes. These concepts can be called the basic principles of a historical dictionary.
The changes which lexemes undergo in course of time can be classified as follows:
a) A lexeme may be lost in course of time. This is mostly due to its disuse, in turn, caused by the lack of necessity to use it on the part of the speakers, ultimately due to changes in the cultural situation. Words about falconry disappeared from English because the pastime went out of use, and the complicated terminology of Vedic sacrifice disappeared because the ritual

Page #390

Dictionary of Sanskrit on Historical Principles
381
was no longer performed on an extensive scale. This loss can be either complete or partial. In the second alternative the words, confined to only a few limited situations, survive under special conditions, 'Kin' remains in English only in the set phrase “kith and kin' and ceases to be used alone. Abhidhānī in the sense of a halter is lost in post-Vedic days but survives only in the set expression aśvābhidhānī and a nyāya based on it in the Mimāṁsā literature. The root dagh - 'to reach a level is no more used in classical Sanskrit but only its derivative daghna continues as a suffix. Only the causal form kāmayati of kam- 'to wish or love' survives, and in case of kan- 'to be glad', a few stray forms alone are found.
b) As lexemes are lost, languages also acquire new ones. The usual process is that of borrowing from other languages with which they come into contact. One important reason why the vocabulary of Classical and late Sanskrit is different from the Vedic is the extensive borrowing which has occurred in the language after the Vedic period. The main sources of borrowings are the Dravidian and the Munda languages. Such words become more frequent in the post-Vedic period. The major part of the borrowings belongs to the sphere of rural life, such as words relating to natural objects animals and birds, trees and plants, dwelling places, utensils, ornaments, implements and instruments, food and drink, diseases etc. There are however some words, mostly of Dravidian origin, which belong to urban life and a developed civilisation. These words are simply incorporated or adapted to the phonemic system of Sanskrit. Absolute new creations are usually rare and production of artificial words in Sanskrit is confined to the technical terms of grammar and the Tantras. Their status as words of a language is difficult to decide. Such words in grammar are ghi, ghu, ti, bha, ru, la, etc. Care has been taken by the grammarians to see that these are kept as much apart as possible from the usual words of the language by giving them a phonemic shape which is not normally possible in Sanskrit, a process opposed to adaptation in case of the loan words.
c) The more usual lexical changes, however, fall under the category of replacements. In the history of a language a given lexeme may be completely replaced by another, borrowed or inherited, so that the one becomes obsolete, the other the current word. Thus Sanskrit aśva was replaced by ghotaka in Indo-Aryan history, and the older svasr by bhagini and āpah by jala or pānīya or udaka in different Indo-Aryan areas. In Sanskrit itself there are such replacements like saru by sarā, āre by ārāt, gātha by gāthā and many others. The replacements can be also partial as when English meat replaces flesh in

Page #391

382
Amrita
some of its environments. In some dialects of Greek even a particle like an is replaced by ke(n). The Sanskrit word abhidhānī in the sense of a halter is replaced by nidāna in the kośas and rarely occurs in usage. In the classical languages, however, due to the constant use of older literature as a model for later works, very often old and new words are used side by side and obsolete words get a new life in some archaizing works. The Bhāgavata-purāna appears to make a conscious effort to use Vedic words which were no more in use in the Classical language. The patterns of lexical replacements can become very complicated indeed and the historical dictionary has as one of its important tasks the disentanglement of such cases.
d) In the course of time the internal boundaries of lexemes may undergo change, one word giving rise to two. This can happen because of metanalysis as when an original a norange and a napron became an orange and an apron. In Sanskrit we have such cases as asura, asita, vidhava etc. where the setting up of a negative a or a privative vi has given rise to new words like sura, sita and dhava. A sound change may also split a word into two and these may develop different meanings. This is the well-known case of differentiation and Sanskrit examples are rocana, locana; pārayati, pālayati; carati, calati; rajyate, lajyate; śarya, śalya, etc.
e) The reverse case of two lexemes becoming one is less frequent. This can happen by the process of suppletion when two independent morphemes so arrange themselves as to make up the pattern of a single unit. Látin fero and tuli, Sanskrit ad-and ghas-, as- and bhū-, Hindi he and thā, Marathi jāand ge-, English good and better, Greek esthio and ephagon are well-known cases of this process. Sometimes two morphemes become one because of phonemic change or abbreviation and give rise to new vocables. Sanskrit śvāvidh has given rise to sedhā and a motor car has become simply car in English. Sanskrit sabhā gives no indication of its compound formation nor is the word sukha felt as made up of su and kha. No negative is felt in anala and no pronominal base like a- is usually seen in words like atra and adya.
The other main problem of a historical dictionary is to trace the semantic development of the words in a language. Here the difficulties are two-fold. It is easy to show changes of meanings and their causes and mechanisms in the case of a few words about which we have good information and where the changes are striking. But to do this for all the words of a language is a stupendous and well-nigh impossible task. Secondly, no well-established universals of semantic change or development are known

Page #392

Dictionary of Sanskrit on Historical Principles
383
which could help us solve the conflict between logical priority of a meaning and earlier attestation of a later sense. The problem is made more difficult by the fact that the associations underlying semantic development are too complex to admit of any rigid classification with clearly demarcated pigeonholes for them. Like the linguistic geographer, a semanticist has to say that every word has its individual semantic history and the task of a historical lexicographer is to recount it.
The traditional classification of semantic changes such as generalisation (Latin molina 'corn mill', Eng. mill), specialisation (mrga wild animal', then deer) pejoration (knave), amelioration (knight) and transference of meaning, based on various relations like part and whole, from space to time, quantity to number, cause and effect, material and product, synaesthesia, physical and mental, general and specific, contiguity and a host of other types, is wellknown and needs no comment. To this we may add cases where old meanings are lost and new meanings are added in the course of time. In fact in the history of a given word many such changes may occur one after another in a complex manner and different lines of development may run parallel to one another. In this regard each word needs a separate treatment based on the evidence for it, and hence the importance of a historical dictionary
far more important for the task of a historical dictionary are the clarification and implications of the concepts of homonymy, synonymy, polyseme and hyponymy among words and the procedures to handle them in a historical perspective. This is likely to differ from the one for a synchronic approach. The field of homonymy is likely to be greater if the history of the item is taken into consideration. Elements of a language which are otherwise identical but have different origins must be set up as homophonous and hence distinct entities. In Sanskrit a- as a pronominal base, a negative particle, the verbal augment, and as interjection were certainly distinct entities from the IE period with different origins. The suffix a of perfect first pers. sg. must be separated from that of the third pers. sg. because of the different original forms. The situation in the case of polysemy is the exact reverse of this. With an item which can be shown to be historically identical in spite of its varying shapes due to changes affecting its form, the lexicographer will have to trace and follow the various factors changing the meaning from one to the other and forming a long chain of evolution producing a case of polysemy across time. Only when he fails to connect two meanings of an item with the usual procedures of semantic development, will he take recourse to set up a new

Page #393

384
Amrita
item and consider them as homonyms as a tentative solution. In Sanskrit the meanings of the word śara like the reed of a plant, the shaft of an arrow, a thorn, a cauterizing instrument, the number five, the versed sine of an arc, one of the nābhasayogas, injury, a poisonous drink, a pose of the hand, a kind of animal, the autumnal wind, a type of poetic composition, a variety of metre, and names of a sage, a king and a mantra are all seen to be connected with one another and hence form the semantic spectrum of the same basic word. But the development is not single and linear, and falls into four different lines running parallel to each other to some extent. In the case of nidāna, the meanings rope for tying cows, for tying the calf, the cause, the original cause, the basic text, the occasion for the use of a mantra, the twelve causes of existence, a kind of penance, the cause and diagnosis of a disease, and the name of different types of works all form a single line of semantic development. But meanings like end or cessation or cut or purification must be assigned to other homophonous items at least tentatively. After the set of meanings has been thus identified the next task of the historical lexicographer will obviously be to arrange them in the chronological order of their occurrences.
The implications of synonymy and hyponymy are of a different type. As a rule a lexicographer has to begin with the Bloomfieldian hypothesis that there are no real synonyms in a language and has to use all the material available to him to define the meanings of items as precisely as possible. The end result of such an attempt will be to distinguish them from all their near synonyms so that some component of the meaning will be the propria of each item in spite a large amount of communis between them. In other words the picture will resemble a series of circles overlapping to a great extent but not coinciding with another. Logically they cannot be also concentric and hence their 'Grundbedeutung' cannot be identical. The methods of structural linguistics based on various types of distribution like complementation, contrast and free variation will be the most useful in this regard. The attempt would be to divide and subdivide the environments to such a degree of delicacy that cases of apparent complementation and free variation will emerge as those of contrast of a refined type. While the contrasting features can be called the semantic differentials the other meanings will be distributed along the axis of hyponymous hierarchy. Other tools to achieve this end would be to make use of the concept of a “lexical field developed by Trier and others and componential analysis of the anthropologists and semanticists.
In this connection it is of some interest to look at the collection of

Page #394

Dictionary of Sanskrit on Historical Principles
385
Vedic words which go by the name of the Nighantavah, which is the oldest Sanskrit lexicon in some sense and on which Yāska wrote his Nirukta forming one of the important Vedāngas. The five chapters of the Nighantu fall into two groups, the first three which are usually taken to put together synonyms current in Veda after the manner of the later kośas form one group and the last two which enumerate individual words the other. The last chapter is known as the daivatakānda and gives the list of Vedic deities including objects which are either deified or to whom hymns are addressed or which are closely connected with regular Vedic gods. The fourth chapter is called the aikapadika and is believed to contain words which are generally obscure and need the help of etymology to understand them. The first three chapters enumerate a large number of words grouped under a common designation given at the end and are taken to be synonyms of that concept or word. There are 122 verbal forms going under the name gatikarmāṇaḥ (2.14), 100 words said to mean water (udakanāmāni) (1.12), 46 words standing for war or battle (sangrāma) (2.17), 44 verbal forms to mean to worship (arcatikarmanah) (3.14) etc. Both nouns and verbal forms are put together in the list of 25 names meaning great (mahannāmāni) (3.3). It is not possible to consider them as synonyms even in the loosest sense of the word, and the best way to understand the mind of the compiler is to think of 'a semantic field' which he intended to give as a help for a better understanding of the vocables. That there can be differences of opinion in this regard can be seen by comparing the list of words which Hillebrandt“ gives for the Vedic poet, singer, bard viz. rsi, kavi, kāru, jaritr, rebha, vipra, stotr, etc. which fall under two lists of the Nighantu viz. 24 medhānāmāni and 13 stotrnāmāni. That other such fields can be formed is seen from the one set up by him for ‘song, poem, thought' consisting of words like ukthā, gir, dhi, dhīti, mati, mănman, stutī, stūbh, stötra, stoma and others. In any case such fields have to be set up and the precise meaning of a given word has to be defined in comparison with other words in the field.
From the above theoretical considerations. we can now draw some practical conclusions as prerequisites of a historical dictionary. The lexicographer has to record all the vocables of the language existing in the period to be covered by his dictionary with as much evenness as possible for all the period or periods. If the literature has marked varieties due to different geographical locations, types of literature, differences in the subject matter etc., he will have to cover them all in an orderly fashion giving due consideration to the importance and volume of each branch or type. All the

Page #395

386
Amrita
meanings recorded for the vocables must be collected with illustrative examples sufficient to make it possible for him to set up all the the relevant. environments of their use. The whole of the material has to be basically. analysed both analytically and historically before a scheme is drawn up to represent the semantic history of a vocable. The nature and limitations of the material will alone decide whether the analytical or the historical scheme will be the primary or the secondary principle of classification. It is likely that the lexicographer will have to change from one to the other as cases arise in the course of his work.
3. A SURVEY OF LEXICAL MATERIAL FOR SANSKRIT
Before deciding upon the exact nature of a Sanskrit Dictionary on Historical Principles, indicating its scope, its perspective and its presentation, it is necessary to survey the nature of lexicographical material of Sanskrit so far available. This material falls into three distinct groups, the traditional lexicons of all types, the early attempts of modern scholars to collect vocables and write dictionaries of Sanskrit during the last century and a half and the raw material which may be collected for the work itself. To this survey has to be added the salient features of historical dictionaries of other languages to fix up the methodology to be followed and if possible, to improve upon them. In this connection two things must be borne in mind. The survey of the earlier works is not intended to indicate their shortcomings but to assess the amount of material in them which can be useful for a historical dictionary. Again, for a dictionary which intends to include nearly all the vocables of a language, it is believed that even the earliest and the most modest attempts have something to contribute.
The ancient lexical material, from the Nighantu to the medieval and late kośas of various types, will have to be handled like any other source material of Sanskrit, pertaining to other branches of scientific writings and belles lettres and should really form part of the corpus on which the historical dictionary is to be based. As in the case of the Dhätupäthas, Ganapathas and other similar works, the kośas need a critical assessment before their utilization for historical purposes, and hence they are kept out of this survey.
Material of the second type can be classified for convenience into three groups. There are glossaries of individual books, like the dictionary of R.V. by GRASSMANN, primarily intended for their study which in some cases assume the form of small dictionaries, vocabularies meant as help to readers containing passages drawn from various types of works and the

Page #396

Dictionary of Sanskrit on Historical Principles
387
chrestomathies which are naturally more extensive than the earlier ones and finally general dictionaries meant to be used to help understand Sanskrit literature as a whole. In between fall vocabularies dealing with texts of a particular type. Closely allied to the first type are the indices of individual books, either arranged alphabetically with the use of full forms as the basis, like MAX MULLER'S index of R.V., ROTH and WHITNEY'S index of A.V. or arranged on the basis of the grammatical analysis as do the various indices of Dr. Vishvabandhu SHASTRI. Here should find place also special indices of proper names like SORRENSEN'S index of Mahābhārata. Glossaries of groups of related works like that of STENZLER for the Grhyasūtras and of L. RENOU for sacrificial terms naturally gave rise to specialised dictionaries of particular sciences like the Dharmakośa, the Mīmāṁsākośa, the Ayurvedakośa, the Nyāyakośa, the Bharatakośa, the Dictionary of Grammar and others which specialise in the vocabulary of a special subject and therefore can afford to go into greater details and come closer to an encyclopaedic treatment of that subject. Glossaries of readers by LASSEN, BENFEY, BOPP, DELBRUCK, LANMAN, 'LIEBICH and others were miniature dictionaries and some of them did develop into full-scale lexicons of Sanskrit. Some of those important dictionaries are characterised below.
One of the earliest lexicons of Sanskrit is Franz Bopp's Glossarium Sanscritum in quo omnes radices et vocabula usitatissima explicantur et cum vocabulis Graecis, Latinis, Germanicis, Lithuanicis, Slavicis, Celticis comparantur, written in Latin and containing about 400 pages, published in Berlin in 1847. The material includes both Vedic and Classical texts on a small scale, selections from the RV including the whole of the first Mandala, selections from Mahābhārata and Rāmāyana, the Bhagavadgītā, Manusmrti and Devīmāhātmya, the grammars of Pānini and Vopadeva, Amarakośa and the poetic works Bhattikāvya, Bharthari's Satakas, Ghatakarpara, Gītagovinda, Hitopadeśa, Raghuvamsa, Kathāsaritsāgara, the dramas Mrcchakatika and Sakuntala and LASSEN'S anthology. More than its contents, including extensive cognates from the IE languages, the glossary even now has some value for the methodology followed by this pioneer of IE linguistics. In the preface Bopp points out how the root stigh- in the Dhātupātha, though not attested in the Sanskrit literature (now found in two doubtful forms of the Mais.) is confirmed by the IE cognates German steigen, Gr. steixo etc. Setting up the root chyu- on the evidence of Vopadeva's Kavikalpadruma (st. 58) in the sense gatyām and confirming it with the evidence of old Persian siyu is doubtful. Bopp has also hinted at the fact that Sanskrit roots are given by the

Page #397

388
Amrita
tradition in their week grades but full-grade forms like kar- for kr- jar- for jretc., are original; and this procedure was fully exploited by BR in their lexicons later. Bopp also used the method of putting all the verbal forms with preverbs under the same article with their meanings, a method which also is followed by BR. Thus under kar- to do (p. 80) we find the prefixes adhi, anu, apa, abhi, alam, ā with apa, à with upa, ā with nir, upa, ni, nis, pra, pra with ni, pra with sam, prati, vi, sam, sam with upa. With very limited Vedic literature available to Bopp, the roots and words which are peculiar to the Veda alone are illustrated with Vedic quotations while all others are explained with classical citations. Some of them, mainly Vedic, he also translates in Latin. With the very few citations used, the question of arranging them in any order did not arise. At the end of the main article, the IE cognates are discussed in full but now, their value is naturally very little. A couple of articles translated into English should give a fair idea of Bopp's dictionary.
A purely Vedic entry is (p. 208) pan- P. A. (Panāyāmi, panāye Pan. III. 1. 28 referred to the first class; sometimes to the 10th class, lengthening of a characterises the aya, but it is omitted in the Vedic dialect) praise, celebrate, RV. 20.6. tat panayad vaco vah 'it approves of your words'; 87.3. svayam mahitvaṁ panayanta dhūtayah 'the shakings declare their power themselves' (p. 284). rac- 10 P. racayāmi (pratiyatne K. krtyāṁ V.) to make, to prepare, R. Schl. II. 13. 12, maya'yam racito' ñjaliḥ P. 33; nagaram racitam Gita-Gov. 5.10; śayanam racayati 4.23; racayati cikure...kurubaka-kusumas; with vi id. Ragh. 5.75; viracitavāgbhir bandiputraih kumāraḥ sapadi vigatanidrah 5.76, kušalaviracitānukūlavesah (Schol. kušalaih viracitah krtah 15.19) Śivadāsaviracitāyāṁ.....kathāyām. Counting the prefixed verb as a separate entry the glossary contains an estimated 10,000 words as its extent. A third edition was issued in 1867 (pp. 492) in which accents were added to the Sanskrit words, all the citations were given in a transliterated form and indices of words of other IE languages used for comparison were also supplied.
An early dictionary of Sanskrit for general use is the one prepared by H. H. WILSON with the help of the Pandits of the College of Fort William, first published in 1819 and revised and extended in the second edition of 1832, covering 982 pages of large size. As explained by the author in his preface, 'the chief object of the original compilation was to embody in one digest, the whole mass of indigenous lexicography and the work was primarily based on the kośas called Amara, Vaijayantī, Utpalini, śāśvata, Keśava, and Siva. The second edition used also Hārāvali and the published

Page #398

Dictionary of Sanskrit on Historical Principles
389
parts of the Sabdakalpadruma of Raja Radhakant DEB. This reliance on the kośas has placed a permanent mark on the Sanskrit dictionaries which followed, in the fact that a large number of meanings given to words in the traditional kośas are always included in them and later scholars had to make a studied attempt to demarcate them as only lexically attested with the addition of a symbol like L. In the second edition, WILSON also used indices of the following Sanskrit works : Siśupālavadha, Kirātārjunīya, Hitopadesa, Manu, Bhagavadgītā, Gītagovinda, Durgāpātha, Amaruśataka, Siddhāntakaumudī, part of Bhatti and part of Mitākṣarā. The number of words recorded is calculated as between fifty and sixty thousand and the author was fully justified in feeling that the dictionary will serve the purpose of all the students and teachers of Sanskrit and will help to spread the knowledge of this language. As regards the way in which the meanings are arranged, WILSON explains that instead of following the order of meanings as given in the oldest kośas like Amara and others, 'the sense that seems most naturally to result from the etymology of the word takes in general precedence of the rest; others' follow according to their more or less direct relation to the original purport, as far as this can be traced.' He illustrates the difficulty of indicating the exact shade of meaning of a Sanskrit word in a given context by an English equivalent which is bound to differ from case to case, with the forms of the verb asri - : svabāhubalam āśritaḥ 'confiding in the strength of his own arm vātāyanāśritā, 'standing at a window'; asanam āśritaḥ 'seated on a seat'; mayüraprsthāśrayinā guhena 'by Skanda, riding on the back of a peacock'; vīryam upāśritya 'displaying valour'; kāmam āśritya ‘affected by · desire'; ūrjāśrayaṁ vacanaṁ 'high spirited speech'; samskstam āśritya 'speaking Sanskrit; prakrtyāśrayah pratyayāśrayo và niședhah'a prohibition regarding either a crude noun or an affix.' Another example given by him is prasarga with its various translations. The entry in the dictionary runs : prasanga m. (gah), 1. introduction, insertion, 2. association, connexion, 3. addiction or attachment to, 4. connected language or style, 5. connected reasoning or argument, 6. double or conjoint result, producing two effects, attaining two objects by one act, 7. revealing a secret, 8. (in dramatic language) a second or subsidiary incident or plot.' E. pra preceding sañj - to join, aff. ac. (p. 581). No. 3 is based on an expression like indriyānāṁ prasangena. It is easy to see that the meanings given by him are based on the best translations of passages in which they occur and represent what may be called the sense bound to a specific context. No attempt is made to separate the meaning of the word and the meaning implied by the specific context. Without recourse

Page #399

390
Amrita
to the passages used by the author, and the author has given neither quotations nor references, it is not possible now to ascertain the accuracy of the meanings given to a particular vocable. That is why later dictionaries have often quoted the meanings of WILSON with reference to his work without any attempt to evaluate them. Thus the meaning 4 mischief, injury, hurt for sara is carried over by later dictionaries and only a chance discovery of its use in an obscure and late work like Raghavanaiṣadhīya (1.15), which śleṣa running throughout, makes it possible to ascertain the value and status of this particular meaning. As is to be expected the words are mostly from classical language, and there is a total absence of purely Vedic vocables, marks of accentuation, references and citations. The striking features of the work are its comprehensiveness and its explicitly stas d analysis of all words. After the publication of the second edition of WILSON'S dictionary, Rev. W. YATES published in 1846 at Calcutta an abridgement of the same under the name 'A dictionary in Sanskrit and English designed for the use of private students and Indian Colleges and Schools' (pp. 928 of small size). The author states the purpose of abridging the text to be the fact that the unabridged work was beyond the means of most students and they were in need of a Sanskrit dictionary within their means. He did this abridgement without, as he says, reducing the number of vocables but by dropping derivations and unusual meanings. Very few abbreviations are used. For the verbs only, a single form of the 3rd p. sg. is given and usually a single English translation of a word is supplied. The number of words is about 45000. After the root are given the Dhatupatha symbols like bhās (n,r) kṛ (d, ñ, du), pat (1), bhṛ (ñ),. kr - bharati, bharate, (li, ñ, d) bibharti, bhṛte but no explanations are added and the reader is expected to know them from his study of the Sanskrit grammar. Compounds consisting of only two members are given and other combinations are expected to be analysed and understood by the students. The work possesses no merit of its own but may have been useful as a handy volume. in those early days of Sanskrit studies.
-
Practical dictionaries on a moderate scale developed out of the glossaries for the Readers. BENFEY'S glossary, prepared for his handbook, was soon rewritten as a regular dictionary called 'A Sanskrit-English Dictionary with reference to the best editions of Sanskrit Authors' (London 1866, pp. 1127 of small size). Its scope includes all the words occurring in the different chrestomathies and selections generally in use (as LASSEN'S anthology, his own chrestomathy, BOPP'S Nala, JOHNSON'S Mahabharata selections etc.) and the texts usually read by the beginner like the Hitopadeśa, Pañcatantra,

Page #400

Dictionary of Sanskrit on Historical Principles
391
Man's Laws, sakuntalā, Vikramorvašī, Uttararāmacarita, Mālatīmādhava, and Meghadūta. It excludes purely grammatical terms and technical terms of philosophy, as also Vedic words not included in the selections, except those in Vedāntasara, Bhāsāpariccheda and Rājatarangini (5). A peculiarity of the arrangement of this lexicon is that the compound words are arranged under the second member and not the first as is usually done. Thus under jāti are given, ajāti, ekajäti, durjāti, dvijāti, pūrvajāti, mrgajāti, sajāti, and hinajāti. Under yantra are given kodanda-, kūţa-, grha, ghatikä-, ghati-, jala-, tāla-, taila-, toya-, dhāra-, dhvaja-, nādi-, su-, sūtra-, stri-, and svamāyā-. This procedure has given rise to a number of problems for the user as well as the compiler. A word like trisüla has to be sought under śūla while trīšarku is found in its natural alphabetical order. If the second member of a compound does not occur in that form as a sep-arate word the procedure cannot be easily followed. BENFEY sets up a form like rāja stating the form occurs only as a second member and then includes under it such compounds as amara, adi-, rsi-, trna-, deva-, dvija-, dharma-, nakṣatra-, etc.
In the case of the verbs, the prefixes are all given below them and are arranged in the alphabetical order of the first prefix. But along with this, BENFEY also gives the various words and forms which form a compound with the verb, which are known to Sanskrit grammarians as gatis. Both the preverbs and the gatis are then merged into a single alphabetical order. This can best be seen in the case of the very common verb kr - 'to do': agnisāt, angi, ati, adharī, adhi, anu, apa, abhi, abhyantarī, alam, abhyalam, upālam, samalam, svalam, ava, ā, apā, upā, samupā, nirā, vyā, ākulī, ātmasāt, āvis, āharani, unmanī, upa, pratyupa, upahāri, urī, ușni, ūrī, eki, ausadhi, kadarthi, kapili, karadī, karī, kaluşī, kasāyī, kāt, kundalī, krtārthi, k.śī, kosthaki, khandi, khalī, khilī, gandūși, gospadi, grāsī, cakrī, candi, camat, cittī, citrī, cihnī, cūrņi, jatili, jatharī, jadi, jarjari, tanū, tāmrī, tiras, tīrthi, tucchi, trni, tvam, daksini, dāsi, divasi, dirghi, duḥkhā, dūrī, drdhi, dviguni, dvijātisāt, dhik, dhvaji, nagni, namas, navi, ni, vini, nimitti, nirdhani, nis, nispatrā, nispandi, nyak, nyāsi, pañcaśikhī, pañcī, parasāt, parāňmukhī, pari, paruṣī, paropakarani, parvati, pavitrī, pasū, pātrī, puras, pulaki, pūt, peśī, peși, pra, vipra, sampra, prakati, pragunī, prati, prativacani, pradaksini, pramāni, prädus, phut, phūt, bhasmī, mandi, malini, mekhali, raktapati, riktī, lakṣī, vandī, vašī, vase, vaşat, vi, vikacı, vinigadi, vimanī, vivarṇamayi, viraji, visayi, vyākulī, śithilī, śirşāvašeși, śūnyi, saṁkramī, sajji, sat, sanāthī, sapatrā, saṁ, abhisam, upasam, pratisam, samrddhi, sācī, sārī, sugupti, sthirī, svī, a total of 157 words. The problem involved in the arrangement of Sanskrit preverbs and gatis has not been yet

Page #401

392
Amrita
satisfactorily solved and needs a fresh look.
Other features of this dictionary can be briefly indicated. The words are given in Devanāgarī and in transliteration. Accents are not marked. Only a few Vedic words and meanings are given. Grammatical analysis is indicated by + between elements. Unauthenticated verbs and nouns are indicated by prefixing a dagger. References to passages are given but no citations are included. The Latin parts of speech are used. Compactness is achieved with remarkable success and cross-references are completely avoided. Cognates from the IE languages are added at the end of the articles. Meanings are classified into major groups indicated by the Roman numerals and minor groups indicated by the Arabic numerals. No very consistant principle can be observed in both these classifications.
In the same year 1866 was published another dictionary of Sanskrit and French by E. BURNOUF in Paris (pp. 781), of comparable size and scope, also intended for the use of students and beginners in Sanskrit. As the author points out on the title page, it was based on the works of WILSON, BOPP, WESTERGAARD, JOHNSON and others. The words are given in Devanāgarī and Roman transliteration and it contains all the roots of the language. Accents are not marked. Compounds with the given word as the first member are arranged below the word itself but only in a transliterated form, which saves a good deal of space. Under artha are given arthakrt, arthana, arthanibandhana, arthapati, arthaprayoga, arthabhrt, arthavat, arthavāda, arthavijñāna, arthavyayajña, arthaśāstra, arthārthin. Under dala we find dalakośa, dalana, dalapa, dalapuspī, dalayāmi, dālayāmi, dalasuci, dalasnasā, dalādhaka, dalāmla. Following the model of the classical languages like Greek and Latin, the verbs are given in the 1st person sg. like pariharāmi, paribādhe if they are joined with prefixes or if they belong to the secondary conjugations. Roots and the forms of the primary conjugations are given separately under the roots themselves : lup - lumpāmi, lumpe, lulopa, lulupe etc., but vilumpāmi is given as a separate entry.
BURNOUF states in his preface that 'in each article one can present the meanings according to the order of their historical sequence and thus prepare a sort of a historical dictionary of the language. But, in the present state of Oriental studies, we believe that a work of this type is well-nigh impossible for Sanskrit. But as most of the Sanskrit words have their roots present in the language itself, one can nearly always, beginning with the etymology, classify the different meanings of a word in their logically

Page #402

Dictionary of Sanskrit on Historical Principles
393
derivable order. This is what we have attempted to do each time it was possible to do it.' (p. vi.) In the two appendices to the dictionary he gives the list of Sanskrit suffixes and the roots classified according to their final sounds.
A dictionary of Sanskrit which has remained a fragment, but a magnificent fragment nevertheless, is the revised version of WILSON'S work undertaken by Th. GOLDSTUCKER, the first volume of which was published in 1856 in London. The copy which I am using has 320 pages and breaks off with the word abhyāhata. It appears that the total number of pages printed was 480 and that the work had reached the word arindama, thus not completing even the first letter of the alphabet. Though begun as a revision of WILSON, it soon grew into an independent work. As one follows the work one gets the impression that GOLDSTÜCKER wanted to include in it as much information as possible wherever he could lay his hands upon it. The earlier part looks like a dictionary but the later parts assume the form of a vast encyclopaedia with all kinds of information included in it. If one compares the entry aja in this dictionary with that of the original work, one is certainly impressed by the improvements effected and the final result is a good specimen of an article in a large-scale dictionary. The seven lines of WILSON are expanded into thirty. The three grammatically distinct forms m., f. and adj. are clearly separated. Derivations from the root aj. are kept apart from the compound form a-ja and numerous meanings, absent in WILSON, are added. The relation of the sense 'goat' with the root is brought out and the use of the word to refer to various Indian gods is traced to its use in the expression aja ekapad. Doubtful meanings like a kind of rice or moon are indicated with a question mark. The word in the plural, meaning a name of a people and a group of sages, is duly separated. He has left the question, whether the meanings Brahmā, Visnu, Śiva, Kāma etc. in the masculine and prakrti in the feminine belong really to the verb aj- or to the word a-ja undecided but he has indicated that this may be a later growth, as can be guessed from his words 'this meaning has been applied to' (p. 40). No quotations are given in these early entries. Very soon, however, GOLDSTÜCKER begins to enlarge his articles so as to include more and more information, not strictly lexicographical, as can be seen in the case of the word angiras about whom 38 lines are written, and this tendency is on the increase. cf. atri (p. 44); Appayyadīksita (p. 206); apsaras (pp. 222-224). Other topics of historical, philosophical and scientific nature are turned into long essays covering many pages and while dealing with them, long passages are quoted and discussed, cf. abhiseka (pp. 274-287), abhyārohaniya (pp. 315

Page #403

394
Amrita
317), abhyasa (pp. 318-320) etc. By the end of the first volume the work has ceased to be a dictionary and has assumed the form of an encyclopaedia containing discussions on diverse aspects of ancient Indian literature, culture, philosophy and what not. GOLDSTÜCKER appears to have overlooked the fact that the prime duty of a lexicographer is to effect maximum economy of space. The usual grammatical information is given by him in an elaborate manner. The root ad- is given as ada, adādi, anudātta, anudāttet, which comes to mean II P.; anu is described as nipāta, upasarga, gati, karmapravacaniya where only the last word would have been sufficient, as it implies all the rest. Quotations are long and given with the names of authors, works or even systems of philosophy with no references which can be traced, probably because they were mostly taken from Mss. sources. Often no indication of their source is given. Thus under anusnaguh 'moon' we have the citation viksipyate svapātena svakräntyantad anusnaguh with no further indication regarding its source. cf. quotations under words like aparāga, aparāc, anuvāsana, etc. But inspite of all these shortcomings, mostly in the method of presentation, one cannot but admire the scholarship, industry, accuracy and the insight of the author.
It stands to the credit of BÖHTLINGK and ROTH to have begun and completed a dictionary of Sanskrit so as to cover the whole range of Sanskrit literature known to them at that time and present the vocabulary of the language in a systematic and scholarly fashion so as to be useful both to the learner and the scholar of the language. Theirs constitutes the only dictionary of the language which can be considered as a tool of research and worthy of being called an exhaustive work for the time of its publication. The dictionary was published by the Imperial Academy of Sciences at St. Petersburg in seven parts, each newer part bringing the work as close to the ideal as was then possible, with the additions in the form of supplements and corrections all along. The steady progress made by the work, due to extreme devotion shown by the two scholars, can be seen by following the dates of the publications of its various parts : 1st 1852-1855; 2nd 1856-1858; 3rd 18591861; 4th 1862-1865; 5th 1865-1868; 6th 1868-1871; 7th 1872-1875, a total 9478 large-size half-pages published in the course of 23 years. Of course this does not give a very accurate idea of the time actually required for the preparation of this work of stupendous nature, because many more years of preparation on the part of both the scholars must have gone into its making. The labours of the two authors were quite disparate. ROTH'S work. was confined to the RV., the very basis of all historical lexicography in Sanskrit,

Page #404

Dictionary of Sanskrit on Historical Principles
395
which he did with thoroughness and the medical treatise of Suśruta. BÖHTLINGK had to handle the whole of the classical literature and many branches of sciences and DELBRÜCK estimates that B. contributed as much as 9/10th of the total volume. Valuable additions were made by WHITNEY for the AV., WEBER for the ŠataBrā, and other ritual works, STENZLER for the Manusmrti and KERN for Varähamihira's Brhatsamhitā on which these scholars were engaged at that time. Both the scholars utilised to the full the lexicographical material available to them in the original works as well as earlier lexicons. The monumental nature of the work is primarily due to the fact that they collected, arranged and added an enormous number of quotations from Sanskrit literature to illustrate and explain the meanings of the words of the language. As WINDISCH remarks, it was a lucky event for Sanskrit lexicography that at the time the work was conceived and begun, the range of Sanskrit literature available in a usable form was limited. For even the boldest mind would have lost heart to do a thing like this if he had known the range of Sanskrit literature and its complexity as it became known by the beginning of the 20th century. The matter has become still more hazardous with the progress made by Sanskrit studies during the last 70 years of the present century with the publication of many more important works newly discovered. To handle now the whole of Sanskrit literature known today, in the manner in which BR handled their material, is nearly an impossible task. The limitations of their work were dictated not by the will of the authors but by the nature of Sanskrit studies at that time, and hence additions on an extensive scale were made from part five onwards. The older uncritical work, the Śabdakalpadruma, was used for words of botany and other Šāstras as also for the less known kośas in Sanskrit itself.
It is not easy to characterise the merits and defects of this work. Though not consciously planned as a historical lexicon, the keen sense of the authors has made them follow the sequence of meanings in a historical order and this is accentuated by the fact that the RV. was treated by ROTH and the material from later literature by BOHTLINGK, thus keeping them separate from each other. The limited scope of ROTH's work made it possible for him to deal with it exhaustively and his work gives the impression of greater thoroughness than that of BÖHTLINGK. All the later works on the history of Sanskrit are primarily based on this lexicon and even today one can follow with its help the historical growth of Sanskrit from the Vedic to the classical "language, as does the historical grammar of Sanskrit by WACKERNAGEL and DEBRUNNER in four volumes without the part dealing with verb which is not

Page #405

396
Amrita
yet published. BR have kept distinct with the use of an asterisk the attested vocables and forms of Sanskrit from those which are not attested. The process of attestation is one of diminishing returns and even today with a corpus ten times larger than that of BR it will not be possible to increase the attested material two-fold. A new attempt at a Historical Sanskrit Dictionary can differ from the work of BR not so much in its harvest of a larger number of vocables, but in a much more detailed account of stage-wise development, a more accurate analysis of meaning accepted and even the rejection of some of the senses, a more balanced choice of illustrations, a more comprehensive treatment of technical literature and a more refined methodology to be followed, besides naturally, the use of new material which was not available to them at that time.
BR give the Sanskrit vocables with accentuation marked by writing a superscript over the accented syllable and attempt a classification of the senses both on a historical and logical basis with exhaustive citations from the early literature, particularly the RV. and the AV. and a wide selection from other works. Grammatical information is supplied fully and references to traditional lexicons are amply given. Emendations and corrections to the material used are suggested all along and information about proper names, names of books and authors is included. Less attention appears to have been paid to geographical names and other information of an antiquarian interest, in which neither of the scholars was greatly interested. Inscriptional material is only imperfectly exploited and the same is the case with the books on such sciences as music, medicine, erotics, astrology etc. The citations are given in full and many others are referred to by numbers only. The compounds of words are treated as full entries but all the prefixes of verbs are collected together and treated under the basic verbal root, which is given in its fullgrade form which differs from the usual Sanskrit tradition. This is practically the only dictionary available for Sanskrit, which a scholar can use for research work in the language. A perusal of the major entries in this lexicon has always proved to be a rewarding study.
The different supplements to the various parts made it necessary for BÖHTLINGK to think of a new shorter version of the dictionary, which he did between 1879 to 1889, also in seven parts, with a total of 2107 pages with three columns on each. The bulk was reduced by dropping all the citations but new words were added and only new references were included so that, taken together, the two works would include the maximum material. Owing to the attitude of both the scholars, the indigenous commentarial literature

Page #406

Dictionary of Sanskrit on Historical Principles
397
was not fully exploited and probably not given its due weight. Much later, SCHMIDT added a volume of additions to the whole work by putting together appendices and new material collected by him. All these lexicons together make use of about 500 books in the Sanskrit language and contain an estimated number of more than 2 lakhs and 20 thousand vocables, counting the compounds as separate words. Incidentally, it may be observed that BR include in their entries long compounds, particularly those ending in a possessive suffix, than is really necessary, as for example : khadgakhetadhanurbanakamandalvaksasutrin, pitrmặtrg-urusuorusadhyanavant, samādhiyogarddhitapovijñānaviraktimant, avyāpakavisayatāsūn- yatvapatra, or the names of works like Tathāgata-gunajñānācintya- visayāvatāranirdesah, Tripadvibhūtimahānārāyanopanisad. Later dictionaries were naturally based on these comprehensive reference books and were prepared either to help the student who could not procure the larger works or were intended to be more handy digests of the same. Thus CARL CAPELLER published his 'Sanskrit Wörterbüch nach der Petersburger Wörterbüchern bearbeitet' at Strassburg in 1887 and an English version of the same in 1891 (pp. 672). The work contains about 50,000 entries in Devanāgarī with the accent of udātta and independent svarita marked. Vedic words are retained to a limited extent and words peculiar to the Brāhmanas and the Sūtras are excluded as being not useful to the beginners. The author claims to have included all the primary words with ascertained meanings. Formal analysis is not given as being fairly apparent. Roots are given in the traditional form as kr- or kr- and prefixes are enumerated under each verb. Explanations are reduced to the minimum and technical terms thus get a colourless explanation like a kind of comparison for as many as 30 different varieties of upamā given by Dandin which is not likely to help any one. cf. saṁsayopamā, nirnayopamā, vastūpamā, dharmopamā etc. Of a comparable size is 'A Practical Sanskrit Dictionary of A. A. MACDONELL (1924 pp. 382). The Vedic language is represented by 120 hymns of the RV., the Aitareya Brāhmana, the Grhyasūtras of Ašvalāyana and Paraskara and the Dharmasūtras of Gautama. The classical literature is represented by the books which are usually read, a total of about 40 works. Of the technical vocabulary, a considerable number of words from grammar and rhetorics is included as being needed for reading the commentaries on the kāvyas. The Vedic words are given with accents in the Roman transliteration. No citations are found and references are also excluded except to mark the word as coming from RV., AV., YV. etc. Attempt is made to reduce the bulk as much as possible by including the compounds under the initial member and the prefixes under the roots and it is claimed that all the

Page #407

398
Amrita
roots of the language find a place in it. MACDONELL writes in his preface (p. ix) "though not comparative this dictionary is historical and etymological in its character. It is historical inasmuch as the meanings are not given in an arbitrary order, as in the Indian dictionaries, while the literary period to which both words and their senses belong, is plentifully indicated. It is etymological from the Sanskrit point of view inasmuch as all words...are broken up into their component parts in the transliteration by means of hyphens." Obviously the author has a very narrow concept of what a historical or an etymological dictionary should be like. All the same, a remarkable compression of the material is achieved by all kinds of technical devices, with which its mere 382 pages contain as many as 50,000 vocables.
In 1932 N. STCHOUPAK, L. NITTI and L. RENOU published Dictionnaire Sanskrit Francais (pp. 897) as a student's dictionary with words in transliteration, not using Vedic works and Buddhist literature though for different reasons. About 30 authors of the classical period and the 13 principal Upanisads are covered by its scope. Naturally accents are not marked; compound words are given below the first member but in a different type making it easy to locate them. Verbs with prefixes are given as separate entries and again the typographical device makes it easy to spot them out clearly. Under each root are included the basic forms which are therefore not repeated in its prefixed form. No references or citations are included and the work is strictly meant for the student of Sanskrit who wants to read classical literature with the help of a dictionary. Different meanings are separated by means of a semi-colon and numerous synonyms are given to convey the sense or senses of a word. The technical make-up is very good which helps its user a good deal.
Two more dictionaries which were intended to be exhaustive lexicons of the language and not merely a help to students were published after the St. Petersburg volumes. One of them is A Sanskrit-English Dictionary by MONIER-WILLIAMS, first published in 1872 and based mostly on the BR, with an attempt to compress as much of the material as possible by leaving out the citations but keeping the references in the form of names of works like RV. MBh. R. etc., and now and then giving the exact source when found important. The work was thoroughly revised with the help of LEUMANN and CAPPELLER and brought out in its definitive new edition in 1899 comprising 1333 large pages printed in a very small compact type and giving much of the material of BR. Whatever new material was added was due to the two scholars who collaborated in the new edition. This work has now become the

Page #408

Dictionary of Sanskrit on Historical Principles
399
standard dictionary of Sanskrit for English-knowing scholars and is often reprinted. The main entries are given both in Devanāgarī and in transliteration with the accents marked. Compounds are given under the entry of the first member to save space but verbs with different prefixes are given as separate entries. The strictly alphabetical order is set aside when words derived from the same root are brought together and cross references help the user when the words are very much separated from their alphabetical place. As a handy lexicon to be used for all purposes, it is useful but possesses no independent merit of its own.
The Practical Sanskrit-English Dictionary of V. S. APTE was intended to make available a lexicon in one volume of more than 1000 pages, compact printed, for, general use with words without transliteration and with quotations from well-known Sanskrit works to illustrate the meanings. More recently this work was revised and published in three volumes comprising 1768 pages and 112 pages of appendices containing rules of Sanskrit prosody, names of Sanskrit authors and their works, geographical names, list of Sanskrit kośas, maxims and grammatical terms and works. Vedic words and citations are given only sporadically and the arrangement of meanings is not based on any discernible principle. Traditional and modern interpretations are often mixed up which lowers its scientific value, and some of the shortcomings of the original work continue to mar the enlarged edition. Even then its usefulness for practical purposes cannot be denied and Indian scholars find it convenient to use it, particularly on account of its apt quotations and other sundry information about Sanskrit language and literature. It can hardly be considered a tool for research. It makes use of about 225 books as against 515 of MONIER-WILLIAMS. In spite of its larger number of pages it contains less information than the other work.
4. NATURE, SCOPE AND AIMS OF A HISTORICAL SANSKRIT DICTIONARY
Given the basic concepts of a historical dictionary, the lexical material available for it and the nature of the methodology used in the best attempts of this type, one may naturally raise the question 'What kind of a historical dictionary of Sanskrit can be attempted with the use of available resources to be completed within a reasonable period of time ?' Naturally the term *reasonable' is undefined and allows a large amount of latitude, making the nature of such a work flexible enough. But some basic facts and limitations are obvious and need a careful consideration, if only to avoid the charge of being impracticable. What is indicated in the discussion that follows are the

Page #409

400
Amrita
minimum requirements for such a work so that the results of such labour are worth it and should prove useful to Sanskrit philology as it stands today. Sanskrit is the classical language of India with a remarkably long history of over three thousand years, and contains a vast technical and scientific literature along with purely literary works. It forms the bed-rock of Indian languages. This fact has some bearing on this problem.
1. Such a dictionary should record all the available vocables in the language from all its periods, beginning with the RV., the oldest available text, to the end of the 18th century. It is true that the later limit can be drawn differently, either at the end of the Vedic period, at the end of the classical period, on the eve of the Mohammedan conquest, the rise of modern Indian languages or at the very end of Sanskrit writings including the most recent attempts, which continue to the very present day. In the case of large dictionaries a decision to limit the period is always attractive and has been often followed. The Latin Thesaurus confines itself to the end of the classical period of that language in the 6th century A. D. The work of Mittellateinisches Wörterbuch begins with the 6th century and ends with the 13th century A. D. A further limitation is put by confining the work to a particular area. No one has thought of handling in the same manner later Latin works of more modern period for obvious reasons. Even after the publication of the Oxford English Dictionary which covers the whole range of English from the very beginning but confines itself to such words as are present in the language in the 11th century, it was found necessary to prepare a similar dictionary of Middle English so as to cover the earlier period of the language thoroughly. DIELS warned against an attempt to cover the whole range of Greek literature beginning with Homer and ending with the present time in one comprehensive work, on the ground that such an attempt was likely to go beyond the abilities of the whole community of Greek scholars. The general editor of such an attempt, he said, will very soon cease to be an editor and remain merely a general commanding a large army of editors working under him. That is why the attempt at a Greek dictionary by SNELL is confined to the early epics and epic poetry. Even this is likely to take a very long time to complete. Dictionaries of modern languages have certainly a shorter span to cover and their problems are comparatively simpler.
There are, however, reasons why Sanskrit cannot be handled in this piecemeal manner. The bulk of Sanskrit literature is concentrated in the middle period and it is so closely connected with the earlier Vedic literature

Page #410

Dictionary of Sanskrit on Historical Principles
that no historical dictionary can really be attempted by dealing with them separately. Much of the value of such a work is likely to be lost because the most fascinating and interesting part of the work is to bridge the gap which separates the post-Vedic literature form the Vedic, both in language and thought. It is easier to argue that a dictionary of this type should stop at the end of the classical period and this can certainly be done without losing the historical perspective. But there are good reasons against taking such a decision. More than the belles lettres which form only about 1/15th of Sanskrit literature, the works on various scientific and religious subjects form the bulk of the language and this literature has a greater cultural value for India today than the merely literary works. And there is far greater continuity in this technical literature from the classical period to the modern times, some branches of which like medicine, astronomy and law, continue to be living subjects of even practical use. Moreover some of the valuable contributors in sciences and subjects like Grammar, Law, Poetics, Kāvya and Nyāyaphilosophy belong to the post-classical period. As closing the work at the end. of either the 16th, 17th or 18th century makes no real difference in its scope, it is certainly advisable to cover as much period as possible and the end of the 18th century is a good closing point. It is obvious that selection of works from this later period has to be done with great care and with an eye on its pragmatic value.
401
2. To cover the whole range of Sanskrit literature for a dictionary from the Veda to the end of the 18th century, one can roughly estimate the number of works which will have to be utilised for this purpose. It is difficult to say how many books are found in Sanskrit, as many of them are not published and the bulk of the literature is in the nature of commentarial works making it difficult to estimate their actual number. Five thousand may be a little liberal estimate. Out of these a well-chosen collection of 1500 books should suffice for any lexicographical attempt to capture almost all the vocables in the language. If anything remains left out it will be only compound expressions with no special meaning other than that of their components and as such of no real lexical value. Any way, with the freedom of using very long compounds which Sanskrit boasts of, particularly in its prose works, it is impossible to record all of them and some kind of limitation must be imposed on them before they are admitted as the lemmata of the dictionary. The number of books used by the existing dictionaries may give a correct picture of the scope and value of a collection of 1500 books for this purpose. BR. use 450 books. MONIER-WILLIAMS uses about 500, and APTE'S

Page #411

402
Amrita
revised version uses 250 books on the whole. The new dictionary will have used three times the books used by the earlier dictionaries and this is certainly a very large corpus. This is however very small if compared to the 16000 books used for the OED, but is comparable to the Latin Thesaurus and is slightly higher than the number of books used for the Middle Latin Dictionary. The larger number which is seen in OED is mainly due to the fact that it deals with a modern living language of international importance and can tap a vast number of books produced in it during the last three centuries. But the density of extraction is proportionately low. Out of the 16000 works, only 6 million slips were extracted while in the case of the classical languages like Latin or Sanskrit there will be as many as 10 to 12 million excerpts out of a stock of 1500 works. Thus the density of extraction becomes 20 times more.
It may further be pointed out that the number of some 1500 works used does not correctly indicate the real extent of the corpus. It leaves out of account the vast commentarial literature based on these works which the dictionary has to utiilise in full. Such are the voluminous commentaries of Sāyana, Mahīdhara, Bhatta-Bhāskara and others on the Vedic texts and of Durga, Mallinatha and many others on the later classical works. These commentaries have their own names like Vedārthaprakāśa or Ghantāpatha and one can include them as separate works among those quoted with suitable abbreviations and double the number of works used. But their most profitable and economical use will be along with the text on which they are based and are best cited with them. The names of such commentaries are less known and hence the author's name is more conveniently used as abbreviation. Some works though based on others are given an independent status like the Šābarabhāsya or even a sub-commentary like Bhāmati, because of their importance.
3. In an extraction of this type every word will have to be included and all meanings of each word will have to be separately noted. But no attempt can be made to obtain a statistically significant picture of either the frequency of the word or of the different meanings in which it has been used. To prepare complete indices verborum of 1500 books is an impossible task and the nature of the material is such that such a collection will give a wrong picture of the distribution of the vocabulary due to an uneven spread of the. material. All that can be done in this regard is to effect a complete extraction of some selected books and such books would naturally be more for the Vedic period than for the later phases of the language. Whenever a full index for

Page #412

Dictionary of Sanskrit on Historical Principles
403
a book is available, it can be used to get the general outline of the history of a word and it can then be clothed into a more life-like picture with the help of the selectional extraction of other books. While a complete index of the RV. is needed with an estim-ated number of 175000 slips, no such attempt can be made for a work like the Mahābhārata with an estimated number of two million entries and works like the Skandapurāņa or a medieval treatise like Viramitrodaya or Krtyakalpataru will have little justification for tremendous labour being spent on them. Only in case of selected works and for selected literature can any statistical statement be attempted.
4. All the available meanings of the words recorded will be given with illustrative examples. The real value of a historical dictionary lies in this collection of well-chosen examples on which ultimately is based the analysis of meanings and which any user of the dictionary should be in a position to do for himself if he so desires. This corresponds to the verifiability or reproduction of an experiment in natural sciences and is equally important in lexicography. Any one who has used dictionaries where meanings are given but no examples or even mere references by way of justification, knows how often he is at a loss to estimate the value of a given meaning. In the vast literature of a given language it is impossible to trace the source of meaning if no reference is given by the lexicographer who sets it up. In fact only authenticated material should be given and no form or meaning should be included in a work of this type which is not supported by an exact reference and often with an illustration. This may not be an invariable rule in the preparation of the dictionary of a living language where an appeal can be made to the 'Sprachgefühl of a native speaker and where meanings and uses are often illustrated with coined examples. But in the dictionary of a classical language like Sanskrit the above rule has to be observed strictly and no exception can be made to it. This principle has some practical consequences for Sanskrit. We know that Panini taught the structure of Sanskrit, both Vedic and Bhāṣā, as an accented language and it is possible to accent Sanskrit words, phrases and even stanzas according to the rules of his grammar. In fact Indurāja in the Laghuvṛtti on Udbhata's Kāvyālaňkārasārasaṁgraha 4.9 and the stanzas 15-17 illustrating slesa successfully accents these verses in order to distinguish between arthaślesa where no accents are different and sabdaślesa where accentual difference is noted, a view held by Udbhata in this context. But a historical lexicographer of Sanskrit will not reconstruct accents of words and formations according to such rules and insert them in his

Page #413

404
Amrita
lexicon. On the contrary the accepted practice is to record the accent of a word or form only when it is actually found in an accented text, and there also in many cases words may not show accent due to reasons like their occurrence as a vocative form of a noun or the finite form of a verb. Even here no accent will be shown if not actually attested. The same holds good for the forms which can be produced by the rules of grammar for nouns and verbs. If a writer on a Dhatupatha produces a large number of forms from roots and gives them as does Madhava in his Dhätuvṛtti, these will be recorded but will be indicated as such, so that they are kept distinct from forms which are found actually used in literautre. The same rule will have to be applied to the various senses given to words by the traditional lexicographer. It is now an accepted practice to indicate such meanings as are not found exemplified in literature by the addition of the siglum L. A similar treatment is given to the large number of roots and their meanings given in the Dhatupāṭhas but which are not exemplified in literature. On the other hand, the roots not recorded in the indigenous lists are few and generally restricted to Vedic works and very late Prakritic writings including the Buddhist Hybrid Sanskrit.
5. As regards the indication of meanings of items, Prof. Aitken goes to the extent of suggesting that the various meanings and senses are mere labels for the carefully collected and insightfully classified sub-sets of illustrations and are of secondary importance. But an important distinction has to be drawn here between a dictionary proper, however detailed it be, and an encyclopaedia or a real lexicon of a literature or culture. The meaning in a strictly linguistic sense is the concern of a dictionary, while the things themselves are the subject-matter of an encyclopaedia, though naturally elaborated with the help of the language used. As The Concise Oxford French Dictionary expresses it, this book is not meant to give information about the things represented by French words, but only about the nearest equivalence between French and English of the words and phrases which represent these things (p. iii)'. But we need not go further with it and argue that 'we aim at translations and not explanations.' This may be true of a dictionary which aims at helping translations of French works into English but cannot be true for a dictionary intended as a research-tool of a classical language like Sanskrit. Explanations are in order here, and probably we will have to err on the side of greater length in the case of technical and special vocabulary than for common words where a neat translation may do the job equally well.
To give the meaning of words is not the same thing as describing or

Page #414

405
Dictionary of Sanskrit on Historical Principles
explaining things, objects etc., nor is the reverse true. Both approaches go together to some extent but they soon diverge from each other. To give the meaning of saru or sara as arrow or reed is not the same thing as describing what a Vedic arrow was like or to what use it was put. Nor is the identity or difference between various kinds of reeds the topic of a dictionary event though it may note that sara is sometimes identified with muñja, bhadramuñja, gundra or is distinguished from them. The various parts of an arrow like anika, apaskambha. apāstha, kulmala, tejana, parṇadhi, parnäni, salya, śrniga, and their mutual relations do not form the meaning of the word sara. They have their own meanings and the user of the dictionary will certainly find them in their proper places. Even a historical dictionary need not reconstruct the history of the arrow as such.
6. The bulk of the dictionary mainly depends upon the number of examples included in it. Nearly 9/10th of the OED is made up of the citations drawn from as many as 16000 books. With a larger selection of examples, the extent of the dictionary can be easily increased. With as many as 1500 books. extracted and with an average of three to four meanings for each word in Sanskrit, one may come across as many as 5000 citations for a common word and in some cases they may even be more. To give them all is out of question and even a single quotation per book may prove impracticable. The real problem for such a dictionary is to keep its bulk as small as possible consistent with the purposes of a historical approach. The practice of PW is to give full references to the RV. and some other books, and only a random choice from classical Sanskrit, which has produced an unevenness in its presentation. The Latin Thesaurus quotes nearly all the passages where a word occurs, which has increased its bulk to such an extent that its completion has become a matter of centuries. The Greek Lexicon of the early epics cites all references which it can afford to do because of its very limited scope. The Mittellateinisches Wörterbuch gives a fairly large selection of examples and following it a Sanskrit lexicon of similar dimensions should give a fairly numerous but carefully selected examples from all the branches of learning. represented in it. Only in the case of obscure words can all or the majority. of the citations should be given because this part of the work is a starting point for later scholarship to work on it. To do this properly, the examples. will have to be spaced at a distance of centuries if the word occurs over the whole period of the language or a large part of it, unless there are special reasons to do otherwise.
7. We have seen that a historical dictionary differs from an analytical

Page #415

406
Amrita
or a descriptive dictionary by using a historical approach. What is expected of it is a picture of the various semantic and formal changes which a word undergoes in the language during its life-period and the ramifications it gives rise to. It is like describing the history and growth of a large tree which has grown over decades with a large trunk supporting it and with various branches it has given rise to, with twigs growing out of them and ending in a thick foliage. It is obvious that the various branches have grown out of the trunk at different times and in turn have given rise to smaller branches in some historical sequence. Many of them live side by side at any given time while a few others have either stopped growing or have withered away. One can easily imagine the difficulty of giving an accurate picture of the growth of such a tree with a further complication that some of the earlier branches have become visible only at a later stage and in some cases the relation of these branches and the twigs to each other and to the main trunk remains out of sight. An ideal approach for a historical dictionary would be to take the cross section of the life-history of a lexical item at stated intervals, say a period of 100 years, and then try to describe its meanings, their distribution and interrelation in a synchronic manner. Then on comparing each such picture with the one that precedes and follows, one can capture the changes which have occurred in the interval. But this is easily said than done. A near approach to it but in the field of orthography can be seen in the way in which the OED tabulates the various spellings of words century by century before its final stabilization. No dictionary known to me has done anything approaching such an ideal.
In actual practice, the Sanskrit material has to be classified both analytically and historically. The primary division has to be analytical in which each distinct meaning of a word is separated so that the usefulness of the dictionary is in no way compromised. Then each meaning is followed historically with the help of examples chronologically arranged. The historical approach is seen in two features of such a dictionary. The reader of the articles will be able to form the rough chronology in which the various meanings of the vocable may have developed with the help of the earliest citations under these meanings and ascertain the oldest and latest use of a vocable in Sanskrit literature. In this context, it is necessary to keep in view the remark of the Oxford Latin Dictionary that 'not too much reliance should be placed on chronology in the arrangement of senses' (p. vi). Secondly the material under each meaning will be arranged in strict chronological order except for specific reasons like imitative and secondary passages or remarks of

Page #416

Dictionary of Sanskrit on Historical Principles
commentators or the nature of the work as being an anthology or a lexicon etc., and the growth in the idea or the concept under study will be historically followed. This will distinguish the work from purely analytical lexicons in which frequency of usage is the deciding factor, the more common or usual meaning being given precedence over others. No apparent principle of arranging meanings has been observed in the existing Sanskrit dictionaries. in spite of their specific claims to the contrary. It is only the PW which first giving citations from the Vedic and then from the classical literature, gives the impression of following a rough historical growth but nothing more.
407
8. It is expected that different meanings of Sanskrit words are likely to show a close correlation with the types of literature in which they occur. Hence the method suggested above is likely to result into a historical study. of the growth of meanings of words in a particular branch of learning or in a particular subject and the way in which a word has undergone changes of meaning in it. The validity of such an approach depends upon the correctness of the chronology of the works on which it is based. Now it is well known that this is one of the basic drawbacks of Sanskrit literature. A strict chronology which can be seen in the OED with definite dates for books used in it, or even the approximate one of classical and medieval Latin authors. whose dates are fairly certain in the majority of the cases, is lacking for Sanskrit literature. With the maximum care one can take, one can be sure only of a relative chronolgy of the books in a given branch or subject because of the inter-dependence of the books on each other, while the sequence across the various branches remains a hopeless task with no chance of any success whatsoever. One can only recall the controversy among Sanskrit scholars about the relative dates of such major works as the RV. and the AV. and the conclusions to be drawn about the language, religion and other ideas based on them. There is another reason for such a limitation. Both linguistically and culturally the deciding factor has not always been the age of a book but its school affiliation and the tradition to which it belongs and these are best seen in the types of branches of studies in the language. The history of Sanskrit words will have to be followed subjectwise, at least for the majority of technical and semi-technical words, and any attempt to do it with the whole literature as one unit is likely to fail and bog down the work indefinitely.
9. The scope of a historical Sanskrit dictionary can best be seen by looking at the information it is expected to supply for a specific word: (i) It will indicate its earliest occurrence in the Sanskrit literature and its latest

Page #417

408
Amrita
occurrence if it has become obsolete in the course of its history. (ii) It will indicate whether a word existed throughout the history of the language, or was confined to definite periods, by showing its presence in a given meaning with examples coming from different periods. (iii) It will show its provenance over the various branches of Sanskrit, whether confined to a single science or current in more than one, or a word of common use occurring in all types of works. (iv) It will record all its available meanings, both common and technical, and by their arrangement show their historical relationship if traceable. (v) For a given meaning, changes in its connotation and development will become apparent by following the examples arranged chronologically in a given branch. (vi) Relation between words and their mutual replacements of different types would be apparent by comparing corresponding periods and branches of any two items chosen from the dictionary. (vii) Doubtful meanings and artificial senses will be shown separately. The dictionary can be thus described as a well-classified stock of meanings of words and their occurrences in the whole range of Sanskrit literature with suggestions about their historical growth and interrelations. Between a speculative and a conservative, a conservative attitude will be adopted as being more useful to scholarship in the long run. 5. PROBLEMS OF A SANSKRIT HISTORICAL DICTIONARY
The problems involved in a historical dictionary of Sanskrit can best be visualised in a concrete manner by trying to sketch the general plan of a work of this type. There are both advantages and disadvantages in the choice of Sanskrit. Such an attempt is worth-while both for the theory of such a dictionary and its practice in this particular case. Compared to the European classical languages like Greek or Latin or those of the Middle East like Arabic or even Persian, Sanskrit has a much longer history and hence it is expected to give greater scope for the historical development of its lexicon. The literature available in it is of vast proportions and of great diversity and hence lack of material will not come in the way of building up a really useful historical work. But while most of the ancient works in Greek and Latin are carefully edited and supplied with indices, much of important Sanskrit literature either lies buried in MSS. or is indifferently edited and lexicographical help is known only for a small portion of it. The importance of scientific and technical terms in Sanskrit is far greater than in many other languages and Sanskrit poses problems of its own in the creation of a purely artificial terminology in the fields of grammar, philosophy and occult sciences. But the most serious difficulty for a historical Saskrit dictionary lies in the

Page #418

Dictionary of Sanskrit on Historical Principles
409
uncertainty of the chronology of its basic works, which forms the very backbone of such a venture and makes the task of the lexicographer difficult. It is obvious that while some of the problems are peculiar to Sanskrit, others are inherent in the concept of a historical dictionary as such and hence of wider interest.
A typical article on a vocable in a historical dictionary of Sanskrit will have the following form with variations in individual cases. This is a kind of a generalised scheme and is bound to differ from article to article. But it has the merit of being fairly comprehensive. In the case of many words there will be some gaps, either due to lack of material or the inapplicability of a particular aspect to it. The scheme can be outlined as follows:
1. Lemmatization, giving the form of the vocable as a guide to finding out the word in the alphabetical order.
2. Its grammatical status, giving the part of speech to which it belongs and further information regarding its syntactic features.
3. Orthography, accent, canonical shape and other formal features.
4. Etymology, IE origin, derivation and the development in Indo-Aryan.
5. The recording of important and irregular bases and forms.
6. Scholia; traditional information from Sanskrit itself; grammatical and etymological explanations and other information in Sanskrit literature.
7. Summary of meanings.
8. Analysis of meaning, different meanings arranged historically and under each meaning a chronological arrangement of the citations to clarify the meaning, along with an indication of the type of works from which they are taken and a precise mode of reference.
9. Labels for the status and nature of the meanings recorded according to the subject matter, literary type and their functions in communication.
10. Peculiarities of usage and idiomatic expressions.
11. Cultural, historical, archaeological, mythological, geographical and other relevant information, if called for.
12. Indication of compound formations of nouns and of verbs with preverbs with cross references.
13. Statistical information.

Page #419

410
Amrita
To give a precise idea of a dictionary of such a nature it seems best to deal with each one of these aspects in detail, note the problems which are involved therein and arrive at some practical decisions.
1. Setting up of the lemmata. The main problem involved here is to fix up the form of the vocable for its entry in the alphabetical order. This is likely to be different for the two main types of words in Sanskrit viz. ro and nominal forms and they need separate treatment. The more important of these is the fixation of the form of roots in Sanskrit. The practice followed in the existing dictionaries differs from one to the other. Often no consistent principle appears to have been followed. Even in the Sanskrit tradition the mode of stating the verb was not uniform. The earliest practice, as revealed in the Brāhmanas, was to use the past participle of the verb to stand for it. Later on, the form of the third person singular was used for stating the root which is the usual practice of the Nirukta, though the Nighantu often gives other forms as well, if no third person singular was found used in the Veda. Thus under attikarmānah (Nigh. 2. 8) is included bhasathah RV. vi. 59. 4. Sometimes the most frequent form is used to represent the root as in Vidhema Nigh. 3. 5, manāmahe Nigh 3.19. Whenever Yāska refers to a root he does so by using its third person sg. form cf. bhavatīti bhāvasya, áste sete vrajati tisthatīti Nir. 1. 1. This form is often inflected for either the ablative or genitive case to indicate its derivatives : dasyateh ksayārthāt Nir. 7. 23; bravīteh sabdakarmanah 2. 22; 5. 19. In the case of Ātmanepada roots, the Parsmaipada form is often used as it is easy to inflect, dayatir anekakarmā 4. 17 : rūpam rocateh 3. 13. Rarely he also uses a form ending in i; grbher grnātyarthe 10. 23; tviser dīptikarmanah viii. 3. He thus supplies all the evidence for Kātyāyana's Vārttika 2 on P. iii. 3. 108 ikśipau dhātunirdese. This tradition may be compared with the one current in Greek and Latin lexicography where the root is given in the first person singular and which has influenced BURNOUF to the extent of using the same form in his Sanskrit lexicon
Pāṇini refers to the roots in his grammar in a number of ways and his practice is not uniform. This has given rise to the idea that there may be interpolations in his work. He uses the bare root in many places, with or without the euphonic vowel a if it ends in a consonant. The root is generally used without a vowel if it occurs at the end of a compound and is easy to inflect but if it is found in the middle of a long list it is usually supplied with a vowel a to help avoid clusters difficult to pronounce and thus obscure the real endings of the roots. In some cases i is added. In some roots which have

Page #420

Dictionary of Sanskrit on Historical Principles
411
a nasal in their make up, like kamp-, krnv., dhinv., etc. they are given with the use of the indicatory vowel i and also with the nasal, a practice which is at variance with that of the later Dhātupāthas where the nasal is not shown in the root. Some roots are given with the indicatory letter u and in all these cases Pāṇini's usage is in agreement with the practice of the Dhātupātha. Verbal forms of the third person singular are also used with or without guna, both for active and middle roots. Attempts to find some significance for this usage have not been successful as in the criticism of the kārikā :
śtipā šapānubandhena nirdistam yad gañena ca|
yatrekājgrahanam caiva pañcaitāni na yan-luki ||| In a few cases Panini uses the indicatory letters n, ñ, k and n to distinguish between otherwise similar roots and this practice is continued by the later tradition. As against the later use of vowels like e, ai, and o Panini only uses the a ending for such roots in his grammar. He quotes about 525 roots in his Astādhyāyi. Kātyāyana in his Vārttikas often gives the form of the root which is different from that of Pāṇini and in all such cases he adds the indicatory vowel i to the root which is absent in Pāṇini. The use of i as the indicator of the roots is the usual practice of the Mahābhāsya whenever Patañjali uses it on his own, but not while repeating the forms in the Sūtras or the Vārttikas. By his time the standard form of citing the roots in Sanskrit has become the one with the addition of an i at the end. With the further development of grammar, the roots were given a definite form with indicatory letters in which we know them in the various Dhātupāthas, which were needed to explain the peculiarities of the pada and other changes in formation.
The mechanisms used by the Dhātupātha consist of the following devices : (i) Cerebralisation. Roots with initial n or s are given with n ors which are later turned into n and s. This is done primarily to facilitate the statement of changes of sounds in these roots. (ii) The use of dental n. In the roots of the first nine classes which end with a consonant and have a penultimate nasal they are shown with an n in the place of the nasal which is later changed into an anusvāra. Similarly roots like att., add. etc. are given with a dental as their first consonant. This helps the system to avoid an unusual change of a cerebral into a dental when the last consonant falls away at the end of the word. (iii) Accentuation is used, both on the root syllable and the anubandhas, to indicate the use of the connecting vowel i in various formations and the distribution of the two padas. (iv) In addition, the

Page #421

412
Amrita
dhātupātha uses the following sounds to indicate various peculiarities of formations : i, u, ļ, !, ā, ī, ū, e, o, k, n, ñ, ț, n, p, m, ș, ir, ñi, şu and du. All these are also referred to in the Sutras of Pāņini which lay down their function. The usual practice of the Sanskrit lexicographers has been to drop these various indicatory letters and give the root in the original form incorporating all the changes effected by them either in the root itself or by labels like P, A, U etc. It is also usual to insert the nasal in the roots which have the indicatory letter, the anudātta i. Only the older lexicons give the root and the indicatory letters to show their functions. GOLDSTÜCKER, on the other hand, with complete disregard for brevity, spells out the functions fully as in ad-adādi-anudātta-udāttet r. 2 d cl. par. all of which can be compressed into II. P.
BOPP pointed out in the preface to his glossary that the root in Sanskrit assumes different grades in different formations and the full grade appears to be the original form. Following this view, which was prevalent in comparative philology, BR decided to take the full grade form of the root as the lemma for their dictionary. This they were able to carry out in the case of roots of the pattern (C)aC, (C)rC etc. but found it difficult to do the same in case of the patterns CiC and Cuc. This procedure gave the roots quite a new appearance in their lexicons and evoked a sharp rebuke from WHITNEY. 'So long as we speak of the Sanskrit root vid- and not ved- so long it seems to follow that we ought to speak of the root vrt- and not vart-?. The following few comparisons will help see how these scholars differed in fixing the form of the root : Dhātupātha
WHITNEY BR arc (pūjāyām)
arc-rc
arc ard (gatau yācane ca)
ard indh (dīptau)
idh-indh idh-indh
rd
r-rcch
ar
; (gatau) rdh (vrddhau)
rdh
ardh
Later lexicographers mostly followed the form of the root as it was given in the tradition of the Dhātupātha but without the indicatory letters attached to them. Not only is this method the usual one but it is also necessary to follow it because whenever Sanskrit roots are referred to or their forms discussed in literature, it is always done with the form found in the

Page #422

Dictionary of Sanskrit on Historical Principles
413
Dhātupātha. It will materially add to the usefulness of the dictionary to include in brackets the actual form of the Dhātupātha with its indicatory letters and meanings.
Sanskrit lexicographers fall into two distinct groups by the method in which the preverbs of Sanskrit are treated by them. The older scholars dealing with the Dhātupātha and the lexicographers dealt with the roots and the preverbs along with the roots under a single dictionary entry and gave meanings and examples to illustrate them under the same head. This is the procedure which we find in WESTERGAARD'S Radicles Linguae Sanskritae (Bonn 1841), the dictionaries of H. H. WILSON, GOLDSTÜCKER, BÖTHLINGK-ROTH, BURNOUF, BENFEY and CAPPELLER. Others like MacDONNELL, MONIER-WILLIAMS, APTE, SCHOUPAK and others give the root with a prefix as a separate entry and place it in the alphabetical order. The former further arrange the preverbs in an alphabetical order but in a reverse sequence so that the first prefix will be again preceded by a second and a third prefix as subsets of the first prefix with a new alphabetical order in each sequence. This can be illustrated with the example of the root vrtfrom WESTERGAARD (pp. 150-54).
vrt
ati-vrtanu-vrt
sam-anu-vrt
apa-vrtabhi-vrtsam-abhi-vrt
ā-vrt
apa-ã-vrtabhi-ā-vrtupa-ā-vrt
pari-ā-vstprati-ā-vrtvi-ā-vrt
sam-ā-vrt- etc.

Page #423

414
Amrita
One result of this procedure is the separation of all the nominal and adjectival derivatives from the verbs with prefixes and putting them in different places which obscures their inter-relations and entails repetition of meanings. Thus abhidruh as an adjective will be an entry under abhi while the verbal base abhi-druh- will be found under druh-. Secondly the user of the dictionary will have to know the analysis of the form before he is able to locate it in the lexicon. The logical conclusion of such a procedure will be to give most of the vocabulary of Sanskrit under the basic roots from which it is derived with the addition of a small number of underived words, an idea with which H. H. WILSON toyed for some time. But this is ruled out even on purely practical grounds.
Once we take the decision that we shall enter into the dictionary verbs with prefixes as separate lemmata, a number of consequences follow. The prefixed verbs must be arranged strictly in the alphabetical order. Thus samanuvrt goes under S and ävrt under Ā. The prefixed forms of the verbs will be thus scattered throughout the dictionary according to the alphabetical place of the initial sound of the prefix. This is likely to lead to some amount of repetition and obscuration of the connection between the verb and the prefixed form, but the nominal derivatives and the basic verbal form will be brought together. The problem thus boils down to either grouping the verbs and prefixed forms together and separate them from the nominal derivatives or to group the verbs with their nominal derivatives together and separate them from the prefixed verbs and their nominal derivatives. Semantic considerations are certainly in favour of the second alternative and the convenience of the user will also be served by it. A descriptive dictionary will have no difficulty in choosing between the two. The historical development of vocables however is so complex that no general rules can be formulated and no single procedure will always help. Hence in this case the best course will be to follow the descriptively adequate and practically convenient procedure to enter the verb with prefixes as separate entries and to list all the prefixes used with a given verb at the end of the article on the verb or in an appendix. The student of the history of the language can then look up all the entries of the verb with different prefixes if required by him. Such a method of cross-reference is the only possible device to do justice to both the aspects of the problem.
This brings us to the question of fixing the list of the upasargas which should be given a separate entry in the lexicon. We know that the usual

Page #424

Dictionary of Sanskrit on Historical Principles
415
prefixes. were originally independent words and could be separated from the verbs in the older language. Only as a result of natural development and in course of time they became inseparable from the verbs. But there will be no difficulty in indicating their older usage in the Veda even if the lemmata are given with the preverbs as is done in the case of the separable prefixes in a language like German. The Vedic citations which will occur will point out when the preverbs are separated and when they have become a fixed part of the 'verb. In fact the exact period when this happened in the case of a given verb and a given prefix will be an essential part of information which a historical dictionary of Sanskrit is expected to supply. As regards their number the differences among the Sanskrit grammarians are of a minor nature. Thus according to the Bșhaddevatā (5. 95), Sākatāyana regarded accha, śrad and antar as upasargas and Kātyāyana recorded marut as an upasarga in place of accha. It is best to confine oneself to the list of 22 upasargas enumerated by Panini in i. 4. 59 as the basic list and derivations from it should be so marked. The real problem in this connection is to decide which of the elements, prefixed to the verb, can be given a separate entry and which should be treated as sub-entries under the verb. Sanskrit grammarians from Pānini onwards make a sharp distinction between the upasargas and the gatis which also are prefixed to the verbal forms. It is true that by definition all upasargas are included in the larger category of gatis, nipātas etc. But what we are concerned with here are the gatis which fall outside the scope of the upasargas proper. The effects of being considered an upasarga or a gati are different and pertain to the rules of Sandhi or other phonemic changes or accentuation and, in rare cases, the form of the word itself. On the whole the cohesion between the upasarga and the verb is greater than between the gati and the verb. The list of Gatis is much larger and includes both closed and open sets. Besides individual combinations of this nature dealt with by Pānini in i. 4. 60-79 they cover a closed set called the uryādigana (P. i. 4. 61) of 38 items and open sets (ākrti-gana) like the sāksātprabhrtini (P. i. 4. 74) and all cvi-formations, dāc-formations (P. i. 4. 61) and all imitative elements placed before verbs (P. i. 4. 62). The best procedure to be followed in this case would be, again, to use the form of gatis as separate entries in all open sets and to have an indication of the members of the closed sets under the verb as well. This will bring about a uniformity in the treatment of both the types and serve the purpose of historical development admirably.
Pānini defines the term dhātu in two ways. In i. 3. 1 bhūvādayo dhātavah the word dhātu stands for the primitive roots of the language as

Page #425

416
Amrita
enumerated in the Dhātupātha attached to his grammar. Whether all the roots enumerated there are primitive in the sense of irreducible minimum elements is a different question and it finally depends upon our analysis of these elements based on comparative evidence from other related languages. As BENVENISTE does, we may set up various kinds of extensions of roots and separate from the root such elements and also suffixes and ultimately come out with a uniform type of triliteral root. In this sense many of the roots in the Dhätupātha including the whole group of the tenth conjugation can be further analysed. Except for noting the emergence of such derived roots at a particular period of Sanskrit this procedure will have no other bearing on a historical dictionary of Sanskrit. It is the legitimate task of an etymological dictionary to handle these problems in earnest.
Pānini also defines dhātu in a wider sense in his sūtra iii. 1.32 sanadyantā dhātavaḥ. By this he clearly means the verbal stems formed by adding a set of suffixes to roots and also to nominal stems. When these suffixes are applied to roots they give rise to desideratives, frequentatives and causatives and these can be best treated with other derivative verbal forms from the same root. When applied to non-verbal stems they give rise to a large number of denominatives with various meanings. Besides individual cases these formations include two closed sets, called .bhróādayah (iii. 1.12) of 29 items and sukhādayah (iii. 1.18) of 12 items and two open sets lohitādayah (iii. 1.13) (?) and kandvādayah (iii. 1.27). As it will not be possible to enumerate these cases, it is advisable to handle the denominatives by giving them after the non-verbal elements from which they are derived. A few verbal derivatives like gopāya can be set up as verbal entries in the dictionary. This is the only procedure which can be justified historically. In the case of secondary roots which have become primary due to usage and meaning, a cross-reference to the primitive root from which they are derived is all that is needed to bring out the historical facts fully. Compound denominatives which become frequent in the later phases of the language will have to be dealt with only in this manner and will have to be given along with the entry of that particular compound word.
There will arise occasions when the lexicographer will have to set up roots which do not find a place in the existing Dhātupāthas and hence their form cannot be fixed on the basis of tradition. Such cases will arise in the Vedic works where many verbal forms cannot be traced to the roots of the Dhātupātha and many verbal forms from later Sanskrit or Buddhist "Hybrid Sanskrit which have arisen under the influence of some Prakrit dialect will

Page #426

Dictionary of Sanskrit on Historical Principles
417
need setting up of new roots. In the first case, under favourable conditions, it may be possible to set up a root on the same principles which the
pātha followed and then rules of traditional Sanskrit grammar would become applicable to them. Or in the absence of sufficient evidence, we can set up a root having an close a form as possible to the hypothetical form of the Dhātupātha. The Sanskritised roots of the later language are not likely to give any real trouble in setting up some convenient form, mostly of a Sanskritised nature.
The remaining technical aspects of articles on roots can be easily handled. After the root can be given the technical form of the Dhātupātha with anubandhas and the meanings assigned to it in the different lists with an indication of the differences, is of some importance. The gana and pada will be indicated, followed by the important and irregular forms, both accented and unaccented, with an indication of their earliest occurrence. If the root is historically the same, the changes in the gana and the pada will be dealt with and forms separated accordingly. As the derivative conjugational forms will have separate entries, only the causative will be included under the root. Since all the participle forms, infintives and gerunds will have separate entries the article on the root will be confined to the finite forms only. The listing of the forms can be done in a fixed order so that nomenclature and space can be saved.
The problems involved in the nominal forms called subantas are simpler but equally important. The major part of the vocabulary is of this nature and the historical development in nouns is greater and of far more importance than in verbs and forms the core of a historical dictionary. It is better to set up the lemma in the form of the prātipadika, and not in the Nom. sg. form, as is done in many lexicons and which is also the practice of Greek and Latin dictionaries. Since one important aspect of the growth of Sanskrit vocabulary pertains to the finals of nouns like thematization, kaextension, the use of root nouns and nominalisation, and another to the changes of gender and number, all of which are inseparably mixed up with the case-ending, it will hamper the clarification of the growth in these aspects if the Nom. sg. form is taken as the entry form. The saving of space in indicating the gender and type of declension, which has favoured this practice, is only marginal in the case of a large-size dictionary. What is most crucial here are the consonant-ending nouns which cannot be consistently set up in their Nom. sg. form because of very extensive neutralisation of Sanskrit consonantal phonemes in the word-final position, obliterating all kinds of

Page #427

418
Amrita
essential differences between voicing and voicelessness, aspiration and deaspiration, places of articulation like the velar, palatal and cerebral and also the fricatives and aspirate. In fact Sanskrit allows only very few consonants at the end of words or before the Nom. Sg. suffix s and hence a choice of such an environment for giving the lemmata is most inadvisable. The prātipadika is the form which is best suited for this purpose.
This decision will naturally lead to the wiping off of the distinction between the nouns and the adjectives. But the other method has also to indicate this difference by giving the adjectives in all the three genders and it is more economical and clearer to indicate the difference by the use of letters like m. n. f. and adj. In fact there is no real difference between these two categories in Sanskrit and there has occurred extensive interchange between them in the course of the history of the language. Changes in gender, number and endings can be better represented by keeping the stem and the grammatical labels distinct.
There is no separate formal category of adverbs in Sanskrit and the case-forms of Acc., Inst., Abl. and Loc. are used to perform this function. The degree to which such forms have become adverbial differs from case to case and often the same form continues to be used both as a regular nominal or adjectival form and as an adverbial form. Hence the most convenient way to bring out this fact is to illustrate the uses of such cases of a nominal stem under that noun if the adverbial usage is only incipient and to set up a subentry if the adverbial usage has developed a good deal. If, however, the form has a suffix other than that of a case, like -tas, or has a peculiar form like uttarāhi, it must be set up as a separate entry. In addition the Acc. sg. of adjectives used as adverbs are given independently.
Various participles derived from verbs function as adjectives or nouns in Sanskrit and they must be treated in the way in which other vocables are treated. Here again, BR have stuck to the full grade form while the Indian tradition and other lexicographers have preferred the weak grade form. As we have adopted the weak form for the root as lemma, it follows that their weak form should also be used as the lemma. But in cases of doubt or ambiguity as in the case of words ending in -añc, -mant,-vant and -mat and -vat it may be useful to give the more differentiated form of the strong grade as exceptions. In this dictionary the nasal is used only in the transliteration and is put in paranthesis.
As regards the indication of the feminine formations which are treated

Page #428

Dictionary of Sanskrit on Historical Principles
419
by Panini in iv. 1. 3-81 and which are considered by him as not giving rise to prātipadikas, the best procedure would be to indicate this with the addition of ä and i in case of adjectives, if such a form is actually found in the material. In the case of nouns the feminine form should be given as a separate entry, due to the nature of specialisation of meaning and the amount of information to be included under it. A difference of gender in nouns is taken to justify a different entry unless it is clearly occasional or secondary. If the alphabetical places of the masculine and feminine are clearly separated as in the case of rājan and rajñi, two separate entries are useful. If the inclusion of the feminine form under the corresponding masculine leads to a disturbance of the alphabetical order, as will be the case if a word like rukmiņi is included under rukmin and follows words like rukminiša or rukminisvayamvara, it is better to enter the word rukmini in its proper alphabetical place. A cross-reference is all that is needed to indicate the relationship.
Words often have a different form as samāsānta and as the first member of a compound. In both cases it should be recorded with a hyphen either preceding or following it. BR do not do so in the case of the first member of a compound when a weak base is found there and indicate only the samāsānta form. In case of alternative stems or heteroclitic nouns the stem other than the one chosen for a full treatment should be indicated by means of a cross-reference.
A basic problem for a historical dictionary is to separate the cases of homophony from polysemy. A descriptivist can easily pick up one or more formal differences among them and set them up as different elements or he may follow consistently a policy of indicating their differences as belonging to different parts of speech in a fixed order. But a historical dictionary has to deal with changes in the formal features used for this purpose and show the development of meaning from one to the other. The use of man or fish as a verb is certainly later than their use as a noun, while equally clearly the use of sleep or fall as a verb is older than their use as a noun. To bring out or even to emphasise such a relationship, which is peculiar to each vocable, is the legitimate task of a historical dictionary. The only course which the lexicographer can follow is to set up two items as homophonous if their historical origins have different shapes, or their origins can be traced to different sources if no natural growth in their meanings can be shown. In all other cases, the items must be considered the same and an attempt made to follow the changes in their meanings. A convenient way to indicate the

Page #429

420
Amrita
homophones would be to number them with index numbers. This procedure, of course, applies to all kinds of words like verbs and particles, though the problem is more acute in the case of nominal stems.
The clearest cases of homonyms are those known as sabhangaślesas to Sanskrit rhetoricians. This means that the words are split differently for different meanings and as such are different items, both originally and in isolation, but chance to have the same shape because of tactical union. Thus a-japa is different from aja-pa and samudra from sa-mudra. Sanskrit literature is full of such cases and for this type of word-play some amount of phonemic latitude is allowed to the poets, giving rise to such statements as dalayor abhedah, by which jada and jala are taken to be homophones. Writers on Kavisiksā give a detailed account of such situations and teach how to produce them on an extensive scale. The second type of homophones consists of words which are clearly different in meanings and also different in their origin. One may compare akşa dice' with aksa ‘axle'. In all such cases the difference is noted by numbering these items with a prefixed Arabic numeral. When the sources cannot be shown as different, which often happens, due to lack of evidence, a careful distinction has to be made between homophones of this type and real cases of polysemy where a word has developed distinct and probably remotely connected senses.
On the other hand, formal differences may indicate different words in many cases. Such differences pertain to gender : śaraḥ, śaram; angaḥ, angam; number : kuravah, vangāh or case-forms : āre, dūrena etc. It may pertain to accentuation of a single word or of a compound where a Tatpurusa may be distinguished from a Bahuvrīhi by a difference of accent. Here again, in many cases such formal differences may be due to historical development and hence may not lead to the postulation of homophones. Each case requires a separate study based on the evidence available to the lexicographer.
Two more problems arise in listing compounds, both as main and as subordinate entries. Compared to other languages, Sanskrit has developed the tendency of forming long and complicated compounds which virtually correspond to the use of subordinate clauses of the European languages. A lexicographer may not be interested in the varieties of clauses but he has to deal with compounds fully in the dictionary. For one thing, they are not a mere sequence of free forms to be assembled at will to form syntactic units of a language but form much more closely connected units and possess most of the formal features of a simple word. They also show a number of

Page #430

Dictionary of Sanskrit on Historical Principles
421
characteristics by which they are distinguished from the analytical expressions of the same sense called the vigrahavākya. And they do show a marked semantic development which qualifies them for a special treatment in the lexicon of a language. Particularly in Sanskrit, they show both flexibility and formal unity which make them qualified to have separate entries in a dictionary. The inner structure of a long compound is of a hierarchical nature, based on the principle of dichotomy. All two-member compounds will, thus, form the basis of larger compounds and as such must be entered in the lexicon. Compounds of three members may be needed as separate entries for special reasons like a peculiar syntactic relation between the members, but still longer compounds can be noted only for special semantic considerations. The pratice of Sanskrit · lexicographers in this regard is not uniform. The inclusion of compounds as part of the vocabulary of a language is comparable to the inclusion of various idioms and phrases which are found in the dictionaries of European languages and it is not possible to lay down strict rules about their inclusion.
The long dvandva compounds produce a further difficulty. The grammatical theory .embodied in the rule, dvandvānte śrūyamānam padam pratyekam anveti, gives rise to a large number of binary compounds, many of which may not be otherwise attested or even permitted. Nor can we include the whole long compound as a single entry. The course followed here is to enter the last binary compound as a separate entry and give the rest as individual words, in order to preserve the grammatical information about compounding in the dictionary
The second problem pertains to the nomenclature of the compounds. The traditional system is complicated enough and may not suffice to cover all the cases. One way to handle the problem would be to leave it out of account and to expect the reader to make up his mind with the help of the meaning supplied. Accentual differences will demand a separate entry.
2. Parts of Speech. A historical dictionary has to indicate the part of speech to which a given vocable belongs, in other words, its grammatical status. Unlike the usual analytical dictionaries it must also take note of the changes the vocables undergo, in course of time, às regards their grammatical status. Hence it will not do to separate items merely on the basis of their different grammatical categories, which is likely to obscure changes in this regard. It will not do for example, to set up two different entries like urvi as the feminine of the adjective uru and urvī as the substantive, as this is likely

Page #431

422
Amrita
to obscure the change of the adjective into a noun by a process of specialization. Nor can we separate nouns from the denominatives produced from them, particularly when compound nouns are so used, cf. pralayajvalanāyate bhāmi Vi. 2. 82, dinakausikāyate RasGan. 2. 53 etc.
The system of classification of parts of speech is likely to differ from language to language. It is often argued that the system should be based exclusively on the structure of the language to be described. The use of parts of speech as found in the classical languages like the one given by Dionysios Thrax (1st Century B.C.) for Greek into noun, verb, participle, article, pronoun, preposition, adverb and conjunction or the one formulated for Latin by Priscianus (6th Century A. D.) in which the article is dropped and interjection is added, will not do for other languages, including Sanskrit, in spite of the close similarity between Sanskrit and the classical languages of Europe.
In the Indian tradition, the oldest systematic classification of vocables is the one given by Yāska in his Nirukta 1.1 tad yāni catvāri padajātāni nāmākhyāte copasarganipātās ca tānīmāni bhavanti. He classifies them into four categories : naman, ākhyāta, upsarga and nipāta.
He enumerates 20 upasargas, ā, pra, parā, abhi, prati, ati, su, nis, dus, ni, ava, uta, sam, vi, apa, anu, api, upa, pari and adhi, (Nir. 1. 3). To these Pāṇini adds nir und dur in i. 4. 59 to explain some forms like nilayate, dulayate which are of doubtful value. The same is true of his two sūtras vii. 2. 46 and viii. 3. 102. the Brhaddevatā 5. 95 tells us that sākatāyana added accha, śrad and antar to the upasargas. Kātyāyana adds marut, srad and antar under definite conditions to the list of the upasargas. cf. Vār. 4, 5 on P. i. 4. 59 and Vār. on P. iii. 3. 106. It is better to confine the list to the original 20 and use it for setting up lemmata with pre-verbs as separate entries. To Yāksa, as to his predecessors like Śākatāyana and Gārgya, these were used with both verbs and nouns because he says nāmākhyātayosa tu karmopasamyogadyotakā bhavanti and Durga illustrates some of them with nominal examples like subrāhmana, durbrāhmana, nirdhana etc. At least in the case of su and dur it is likely that they were originally used with verbal nouns and then carried over to the finite verbal forms as in the Taittirīyasamhita sujāgrhi I. ii. 3. 1. Similarly we find pra and ud extended to nouns which may have no verbal root from which they can be derived (cf. P. v. 4. 119). Twelve upasargas have developed a special use with nouns. These are called karma-pravacaniya (P. i. 4. 83) and they govern the accusative, ablative or

Page #432

Dictionary of Sanskrit on Historical Principles
423
the locative case of the noun (P. i. 4.84-97). The remaining indeclinables are called nipāta by Yäska (Nir. 1. 4-11) and are classified into three groups according to their meanings, those expressing comparison (upamārtha), those expressing some other meaning (karmopasangrahārtha) and those merely filling in the metrical line without any significance of their own (padapūrana). These are overlapping groups and Yāska already notes the difference of usage between the Vedic and the later language for some of them. His list runs as follows: iva, na, cid, nu, ca, ā, vā, aha, ha, kila, mā, khalu, śaśvat, nūnam, sim, eva, tvad, kam, im, id and u and further combinations of these like na, kila, nanu kila, net, na, cet are also mentioned. Most of them are included in the cādigana of Pānini (i. 4. 57) while ā, na, mā, sasvat and kam are included in both the cadigana and svarādigaņa (P. i. 1. 37); cid, hi, tva, and id are lacking in Pāṇini's Ganapātha though he knows the use of hi and cid; tva and tvat are included by him among sarvanāman. Obviously Yāska's list is not exhaustive and he himself is doubtful about the status of tva.
Pāṇini's analysis of Sanskrit makes use of classification of two types of elements, morphemes and words. The second alone corresponds to the parts of speech of Western tradition. The basic elements of Sanskrit, according to him, consist of dhātu, pratyaya and prātipadika. He uses either enumeration or mutual exclusion as his method for defining these elements. The roots (dhātus) are enumerated which he can do for Sanskrit because of the extreme transparency of its formations. The prātipadikas coming under the scope of unādi (P. ii. 3.1.) are generally considered by his followers as avyutpanna although an ancient school tries to derive them from roots. These are the primitive nominal stems. The roots have given rise to derivative roots with the addition of suffixes called san and others (P. iii. 1.32). To the dhātus are added two types of terminations called tin which give rise to finite verbal forms and krt (P. III. 1. 93) which give rise to nominal stems or nonfinite verbal forms. To the nominal stems are further added secondary suffixes called taddhita (P.iv.1.76) which do not include the feminine suffixes. The stems other than the roots with primary and secondary suffixes are called prātipadikas to which are added the suffixes called sup giving rise to nominal forms (P. iv. 1. 1-2).
The scheme of formation or derivation gives Pānini the basis for classifying all the words in the language which he call padas. The main division is between subanta and tiñanta (i. 4.14). The second group includes all the finite forms of verbs and are called ākhyātas. The first group includes every other word in the language and can be called a nominal form.

Page #433

424
Amrita
However, the sub-antas are further classified mostly in the fashion of concentric circles. The smallest group is formed by the karmapravacanīyas, 12 in number, used with nouns and can be called adnominals. They are also included in the next larger group called the upasargas (i. 4. 58, 59) which are enumerated in the prātigana. While the additional ten can only function as pre-verbs, others can function both as adnominals and preverbs, this being a matter of functional difference of basically the same items having the same historical origin. The next larger group consists of all these and a few other elements which can be prefixed to verbs to produce complex verbal forms and are called gatis. The next group includes all these, both the verbal prefixes and the adnominals and with the addition of the cādigana, it is given the name of nipāta (i. 4. 56). With the further addition of words forming the svarādigana, they constitute the next larger group called avyaya and the sup terminations are said to be lost after them (ii. 4.82). All these and many other nominals including numerals, pronouns etc. constitute the subanta group, contrasted with the tiñanta one, into which all the words of the language are divided. This classification can be made to correspond to the parts of speech of the Western tradition by classifying subantas, other than avyayas, into substantives (nāman), pronouns (sarvanāman), numerals (samkhyā) and participles (krdanta). The verbs correspond to tiñanta; prepositions, conjunctions, and interjections to nipātas and the, adverbs to avyaya to some extent. The article has no place in Sanskrit and the category of adjectives is ill-defined for it. It is quite different from gunavacana and is probably visesana. We can thus fix up the grammatical status of words in Sanskrit as follows:
1. Verb, tiñanta, ākhyāta :
Symbolised as Class I-X Ā (tmanepada), Parasmaipada), U(bhayapada)
m.f.n.
Pron.
Num.
2. Noun, nāman : 3. Pronoun, sarvanāman : 4. Numeral, samkhyā : 5. Adjective, višeșana : 6. Indeclinable, avyaya : 7. Particle, nipāta : 8. Adverbial, gati :
Adj.
Ind.
Part.
Gati.

Page #434

Dictionary of Sanskrit on Historical Principles
425
9. Preverb, upasarga :
Prev. 10. Adnominal, karmapravacanīya : Adn.
All these are further grouped into more comprehensive parts of speech which are systematically indicated in the Dictionary and cover all the words in the language. These are verb, noun, adjective, adverb and indeclinable.
3. Orthography. The marking of the formal features of the vocables of Sanskrit, like orthography, accent, length etc., is not complicated. The orthography is mostly fixed and variations pertain to some sounds, like v and b which is mainly a regional distinction confirmed by modern languages, and I which is partly historical and partly regional, d, dh, and 1, lh which is confined to the Rgvedic tradition and to the Nirukta, ś and s in free variation in many items like kośa and kosa, n and n which is due to the limitations of the influence of cerebralization in words and compounds and words of non-Sanskritic origin, k and g or ph and bh, a tendency seen in the southern regions under the influence of the Dravidian languages, and a few other sporadic variations. Many of the Siksās, a few Prātiśākhyas and some specialised Kośas give information on such points which will help in fixing the orthography in many such cases and will need recording in the dictionary.
Normalisation of Sanskrit writing to some extent is necessary due to the fact that the scribal tradition was not very consistent and could not be so as it belonged to different regions and different times. It covered a vast field and was carried on by persons of varied scholarship as is well-known to the editors of Sanskrit works. Nor is the teaching of grammarians quite uniform. It may therefore prove acceptable to use parasavarna before stops in place of anusvāra, which should be confined to the place before the sibilants and the aspirate and further, as a matter of convenience, before the semi-vowels and laterals. The visarga should be used only before the velar and labial voiceless stops and before the sibilants and should replace the theoretical signs of jivhämūliya and upadhmānīya wherever used. Only in rare cases will the visarga occur before dentals when followed by another sibilant as for example before a word like tsaru. The writing of a geminated consonant in a cluster should be confined to those cases where it is etymologically justified and not in cases like adharma etc. The nasalised vowels should be indicated when specifically taught. The peculiarities of writing anusvāra or other sounds like a nasalised y etc. will occur only in the citations of the Vedic schools where they are specifically used and not in the lemmata. The transliterations of words alone will be given with acute or grave marks, while the citations will

Page #435

426
Amrita
use the traditional mode of marking the accents. When a word shows a change or variation in accent, both will be indicated and in case of final consonants of the prātipadika the original value will be used. Other niceties and peculiarities of sounds and symbols will be treated under their technical names like yama, nāsikya, jihvāmüliya, ya-śruti etc. but will not be used all along.
There is no problem of hidden length in Sanskrit as in Latin or Greek. But clusters are of two kinds, which may or may not make position according to the rules of metrics. An indication of this fact will have to be given in the case of such words. Only in very few cases is the orthography at variance with the actual sounds as in pāvaka or tarasita where the original phonetic facts would be duly shown.
4. Etymology. The etymology of the words of a language is the prime concern of an etymological dictionary. Sanskrit has a couple of such etymological dictionaries which are completed and a couple of others have remained incomplete. As we have seen earlier, all the three approaches to lexicography, descriptive, historical and etymological, are distinct and give rise to three types of lexicons. But among them the historical and etymological are close to each other, both being diachronic in approach. They are related to each other as history is related to pre-history. Their aim remains the same viz. to follow the earlier stages of the linguistic elements as far as one can go. But they differ in the evidence used. While the historical dictionary makes use of the evidence in the language itself, the history of which it follows, and can make use of the method of internal reconstruction, the etymological dictionary makes use of the evidence of the 'cognate languages and using the comparative method reaches greater depth in time. But there is no real difference between the two and the methods used are not mutually exclusive. In fact, their concurrent use is likely to give us greater depth in time and, what is more important, greater confidence in the results. A good deal of historical value can be known from the internal relation of words in Sanskrit, both in their phonemic shapes due to morpho-phonemic alternations and their grammar based on irregular paradigms and a historical dictionary can utilise this information to the full. The complex relations between the allomorphs of a morpheme are due to historical changes which lie behind them and a careful interpretation of these may reveal the changes which have caused them. Thus, some at least of the historical changes which can be detected by the more powerful comparative method, can also be guessed from the internal evidence. This can be clearly seen in the masterly exposition of Latin

Page #436

Dictionary of Sanskrit on Historical Principles
427
phonology which NIEDERMANN gives in his Phonetique historique du Latin, and the same is true of the grammar in CHANTRAINE'S Morphologie historique du grec. A similar attempt can be made while dealing with the etymology of Sanskrit words.
The historical dictionary will have to bring out all such morphophonemic relations found in the words of a language and also deal with the derivations in full. It will also treat the relations between cognate words in Sanskrit like variations in verbs and nouns, though they cannot be regarded as derivations in Sanskrit itself. But should such a dictionary go further and include in its scope the history of the vocables up to their IE originals by the use of the material from the cognate languages ? To do this will amount to including in its scope the work of an etymological dictionary although both the OED and the Latin Thesaurus do so. But it must be realised that the level of confidence between these two fields is different and a careful perusal of the two Latin etymological dictionaries, one by A. WALDE and the other by ERNOUT and MEILLET, will indicate clearly the difference between a proved etymology and a merely probable one. As the etymologies of the first type are so few and so well-known, there is no real gain in going over them again and the probable ones are subject to such quick and drastic changes that a work of the type envisaged should better keep them out so as not to become outdated soon. Moreover, the historical dictionary is meant to supply to the etyymologists only the authenticated material so that he can build his etymologies on solid foundations. Naturally the historical material should not be coloured by any thought of preconceived etymologies.
There are, however, a couple of points on which the help of etymology will be urgently needed by the historical lexicographer. To decide whether or not to set up homonyms, he will have to use the etymology of the items under consideration as one of the means when semantic considerations either fail or are indecisive. Sometimes etymology may help him regard two items as identical in spite of the differences of meaning. Secondly, the etymology of a word may give him a clue to the more basic and historically original meaning when the purely chronological evidence leaves him in doubt. Meanings of cognates from the related languages will be of some help to him in this regard. In the present dictionary we will have to be satisfied with merely giving references to WALDE-POKORNY or POKORNY for IE and to TURNER for Indo-Aryan developments. For etymology proper, reference can be made to MAYRHOFER and for borrowals to DED of BURROW and EMENEAU and a few other standard dictionaries. Like the

Page #437

428
Amrita
ordinary desk dictionaries, the borrowings will be merely indicated by a reference to the donor language.
5. Forms. A recording of the attested forms of nouns and verbs in the literature is expected of any large-scale dictionary. The obvious purpose of this is to distinguish between forms which are theoretically possible and those which are in actual use. This is necessary because the theoretically possible forms may not be always usable on account of semantic and grammatical restrictions. The latter type of restrictions are really a part and parcel of the grammatical description of a language but will have to be included in the dictionary as a matter of convenience. Collective nouns, nouns meaning uncountable material or abstract nouns may not be found in the plural. But even here languages differ from one another and the actual items which come under these categories cannot be always delimited by a simple rule and often need enumeration. This fact makes it necessary for the dictionary to take note of them. That the word śara in the sense of 'water is found only in the plural as in Instr. saraih and Loc. śaresu is of interest and needs recording. It may be partly due to its association with similar forms of sara meaning "arrow used in s slesa and partly due to the association with āpah always used in the plural
The purely semantic constraints are difficult to capture. That saraih 'with arrows' is the most frequent form can be easily explained on the basis that it is a weapon used in large numbers. That sarau as dual is due to its association with the bow or the bowstring is also understandable, as also the peculiar situation where two arrows are referred to in the astronomical calculations in a circle. But a dative like sarāya is simply due to its deification (namaḥ śarāya, Tantrasāra 98. 26) and the locative singular śare is due to the fact that names were carved on the arrows (sare nämāksarāni vācayitvā, Abhiseka 1. 26]). The ablative sarāt is due to the Tantric figure represented by it and implies that things are taken out of it.
The recording of attested forms is important for the history of the language when they relate to alternative forms and usages. The Nom. pl. śarāsaḥ is older than sarāḥ and a historical dictionary has to note it. The frequency of forms which are ambiguous as regards the gender of the word explains the change of gender of a word like śaru from feminine to masculine. Hence information in this regard is essential. What forms of a noun or a verb should be recorded will depend upon considerations like a difference or shift of accent, alternative forms for the same grammatical

Page #438

Dictionary of Sanskrit on Historical Principles
429
category and forms which are irregular, archaic or important for some specific reason.
The historical dictionary is also expected to contribute to the problem of the attestation of the roots as listed in the Dhātupāthas and indicate when a particular root came into use much more precisely than what can be found in WHITNEY'S rough and ready classification of roots into stems found only in the early language (of Veda, Brāhmaṇas, Upanisads, Sūtras), stems found in both the early and later language and stems found in later language (Epic and Classical) only. Many of the so-called unattested roots may be found attested in the vast Sanskrit literature, though some are bound to remain without an attestation. It is equally important to know what possible grammatical forms of the roots are actually in use in literature and which are only theoretically possible. Here may arise a situation which may not be found in the case of nouns. The verbal formations of Sanskrit are most intricate and hence they have attracted the attention of grammarians from early days. Pāṇini himself gives roots which are not included in the Dhātupātha and are hence called sautradhātus. While dealing with the verbal formations based on the Astādhyāyi, later compendiums like the Prakriyākaumudi use the list of the Dhātupātha and illustrate their various formations with the application of the Sūtras. In addition, it appears that systematic collections of verbal forms were undertaken from early days and went by the name of dhātupar-āyanas. Later on this activity took the form of writing elaborate commentaries on the Dhātupātha itself and we have three well-known works of this nature on the Dhātupātha of the Pāninian system : · Ksirasvāmin's Ksiratarangini, Maitreyarakṣita's Dhātupradīpa and Mädhava's Dhātuvrtti, in which a large number of verbal forms are listed. But all such forms cannot be called attested in the sense in which we generally use the term. Hence it is necessary to make further distinction in this regard. The following classification of possible verbal forms is expected to meet the situation :
(i) forms attested in literature, Vedic and post-Vedic.
(ii) forms attested in literary works which are intended to be also illustrations of grammatical rules, the so-called Dhātukāvyas and Dvyāśrayakāvyas, abundantly available in Sanskrit. The use of a form in one of them is subject to some kind of limitation.
(iii) forms found in systematic works of grammar and the commentaries on the Dhātupāthas.

Page #439

430
Amrita
(iv) forms which are theoretically possible but not attested.
In addition a verb may be used by a commentator merely to give an etymology to a given word.
If forms of the third group are to be included in the dictionary they will have to be marked as such and the first two categories will have to be also kept separate. As the inclusion of all such forms is likely to become unwieldy, some method of reduction will have to be used. In the present dictionary, only important forms from all these categories are given with proper indication of their status.
6. Scholia. The use of scholia in Sanskrit is an important point. The etymological and grammatical activities began very early in India and both the Nirukta and the Vyākarana were included among the Vedāngas. Among them, the Nirukta appears to have flourished earlier and was considered important because the etymology of a word, in accordance with the primitive belief, was thought to be its very essence and thus giving the real meaning. This attitude explains the numerous attempts of the Brāhmaṇas to explain the essence of a deity, a rite or an object by trying to give the etymology of the name with the use of the abstract suffix tva. Yāska's Nirukta is the oldest known attempt at giving systematically the etymologies of Vedic and incidentally of non-Vedic words either to reveal their real significance or to explain their sense when it was obscure. Grammar was mostly concerned with the analysis of words and their formations and meaning played a subordinate part in it. The tradition of giving the meaning was continued in the later Dhātupāthas and Kośas up to the very end of Sanskrit literature. The mine of information found in all such works must be fully exploited by a modern lexicographer, even though some of it may prove to be of little value. A systematic statement of views, whether true or based on popular ideas, is required because it tells us what the early writers in Sanskrit thought about the meanings of nouns and verbs. If the material proves to be too extensive, the editor can state in full the view which he accepts as correct and give mere references to others. How much of this material requires to be given will depend on individual cases.
The use of the commentarial literature is another problem which needs a careful handling. A large part of Sanskrit literature is of this type. The commentaries range from virtually independent works nominally connected with some old work to mere marginal notes and explanations of a few words of a text called the tippanis. All varieties of intermediate types can be found

Page #440

431
Dictionary of Sanskrit on Historical Principles
and it is not possible to classify them with accuracy. Many of them make real and independent contributions to the subject while others are mostly. explanatory. In the field of philosophy many of the so-called Bhasyas and Vivaranas are virtually independent works and will have to be so treated. Commentaries on grammatical works or literary books are generally of an explanatory nature but even here some of them have great value like the Manubhasya of Medhätithi or the Abhinavabhärati of Abhinavagupta on the Natyaśāstra. Many commentators often take a view independent of the author whose work they are commenting upon, and so they must be used for their own sake. The most natural place to quote them would be along with the text on which they have commented, irrespective of the chronology of the books. This is particularly so if they are of an explanatory nature. If there is more than one such commentary, their historical sequence is important because the interpretation of a text in different times is likely to reveal some changes in usage as well as some changes in attitudes. Any one who takes the trouble of scanning the large number of commentaries on Manu can easily see how the interpretation differs due to the date and place of the commentator and his views on the customs and rites described therein.
The following scheme is adopted for quoting the remarks from the commentaries:
(a) to give an explanation of a word or a passage with which the editor agrees. This will be indicated by merely quoting the remark in brackets. with the word comm.
(b) to give additional information, like types, classes, other interpretations, etymologies etc.
(c) to give an explanation with which the editor does not agree but thinks it worth consideration.
(d) to indicate a difference of reading in the text.
(e) to inform that the commentator misunderstands the text.
The name of the commentator or of the commentary will be given with the use of an abbreviation if it is other than the one included in the edition used for extraction. Otherwise a mere comm. would suffice for this purpose.
7. Summary of Meanings. Whenever the article is found to be longer than a full column of the Dictionary and the meanings analysed are complex,

Page #441

432
Amrita
it is thought desirable to add a brief summary of the meanings at the beginning to help the reader get a general view of the article and locate quickly the meaning required by him. For the general reader this part of the article will suffice for all practical purposes.
8. Analysis of Meanings. The meanings of an item are related to each other in a complex manner and can be looked at from different points of view. Each approach may classify the meanings not only in a twofold division but in a multiple manner to bring out their maximum relevance.
One can relate the meanings of an item historically, the way in which they arose and developed. Naturally this is the basic approach of a historical lexicon. In a sense it is based on some universal concepts of semantic development. An item may show different lines of development and these can be arranged in a historical sequence. It can be easily seen that such an arrangement may not be chronological in the strict sense of the term, if we only consider the literary evidence for them and the dates of the works in which they occur. Śara as a name of a king or a sage is illustrated from the RV., while other meanings come from much later sources. The developments of meanings do not always run parallel and we have to admit that a meaning attested in earlier periods may be later in origin. It is equally obvious that even as a proper noun sara as the name of a mantra is later in date. In any case the final order has to take note of the other relationships which are discussed below and a mechanical chronol-ogical sequence of citations will not do for a historical dictionary. Normally the historical sequence of the citations will agree with the development of the meanings as can be seen by comparing the meanings of sara a reed, an arrow, a thorn, an iron pike, five, versed sign of an arc and one of the nābhasayogas. In others, as in the meanings like injury, a kind of an animal and the wind in autumn, no historical sequence can be suggested for lack of evidence and hence they have to be shifted to a sideline as it were. We are generally sure of the chronological sequence of the citations pertaining to a given meaning and here they have to be arranged in a strict temporal order.
Meanings may be related as connotative or expressing common things or ideas and denotative or expressing proper names. The second set of meanings implies some amount of arbitrary use and cannot be considered as being on the same plane with the other meanings. It may be a matter of convenience or even of arbitrary procedure to relegate all the denotative meanings as proper names to the end of the article on a word but it is

Page #442

Dictionary of Sanskrit on Historical Principles
433
necessary
Another way to distinguish between different meanings is to use the criterion of frequency. Some meanings are found to be very frequent while others are rare. An analytical or descriptive dictionary may use this feature as its basis and arrange the meanings according to their decreasing frequency. A historical dictionary cannot do this though it cannot also neglect this aspect, because the attestation of a rare meaning is bound to be very limited and may explain its occasional nature. It is equally true that a less frequent sense may be older than a more common one. In the case of śara, meanings like arrow or reed may be called frequent while the remaining meanings can be considered as rare. Vigraha in the sense of body or war is more frequent but late.
The vocabulary of a language falls into three main divisions, technical, semi-technical and non-technical or common and many vocables also show similar distinctions in their meanings. The meanings of śara like five, name of a composition and of a mantra may be called technical, meaning like a poisonous drink semi-technical and others non-technical. There is usually less semantic development in the technical senses than in the non-technical ones. The transferred meanings often give rise to a semi-technical sense. It is the non-technical common vocabulary which shows great fluctuation and must be adequately illustrated. There is no great gain in illustrating purely technical meanings with a chronologically arranged array of examples and probably citing the clearest definition would serve the purpose as well.
. The meanings also differ in their life-span. Some of them are found used throughout the period surveyed while others are used over shorter periods of time, becoming obsolete at some point. This can best be shown by the number of citations given from different periods. Naturally rare meanings get better illustrated because the very frequent senses cannot be illustrated with proportionally large citations for fear of increasing inordinately the bulk of the dictionary. Considerations like doubtful interpretation and limited use may also weigh in giving more examples for words than is strictly due to them. Śara, in the meaning arrow is used hundreds of times in all periods of Sanskrit but only one case is noted of its use in the sense of injury or killing. The citations also generally group themselves according to the meanings. Technical and scientific literature supplies the more technical senses while the common meanings are mostly illustrated with examples from belles lettres.
Finally, the meanings can be considered as either attested in usage or

Page #443

434
Amrita
not so attested. Obviously the latter category can be based only on the express statements of ancient lexicons or some stray remarks of an ancient author. No examples so far are found for the meanings of śara, as wind in autumn, a poisonous drink and a thorn.
A historical lexicographer has to consider all these different approaches and arrive at a general scheme of arrangement which will reflect the historical development of a given item best.
The analysis of meaning is based on a careful study of the citations collected for the dictionary and naturally these will have been classified into different sets according to the meanings of the vocable occurring in them. The citations are the primary material and the fixing of the meaning is the work of the editor. His task, then, is three-fold. First, he decides in how many different senses a word is used in the citations, which differences in the meanings are due to the specific context and which are due to a specific use of the word, which senses will be considered as distinct, which as mere overtones or shades of the same meaning and how to define these meanings or translate them in the object language so as to bring out the exact shade of the meaning. Secondly, he has to arrange these meanings into a historical sequence primarily based on the dates of the material used and making adjustments for occasional cases where the historical sequence, of meanings may not be accurately reflected in the chronology of the citations. He will also try to fix the temporal limits in both the directions, the earliest and the latest occurrence of a word in a given meaning. Thirdly, he will make a good selection of the citations on which his analysis is based and present them in a systematic and chronological order to display, to the reader the basic evidence on which the article is based.
The art of fixing the meaning of a vocable by studying its use is basic to all kinds of philosophical studies and has been practised all along where interpretation of the text is a part of such studies. Some philosophers of language go to the extent of suggesting that the meaning of a linguistic item is nothing but the way in which it is used and this is certainly true as far as historical lexicography is concerned. If a living language is acquired in its natural surroundings and in its daily use in communication, the meaning of an item ought to be perfectly grasped by the users after some practice and. facility in the language. Two things contribute to achieve this : the linguistic context in which the vocable is used and the non-linguistic context of situation, which accompanies it. But in the study of classical languages, the

Page #444

Dictionary of Sanskrit on Historical Principles
435
two are present only to a limited extent. The context of situation gets merged into the linguistic context and is not independent of it. Hence the problem of interpretation, in cases where the material is confined to older documents, requires effort. To sort out the non-linguistic context of a vocable, the lexicographer collects as many diverse excerpts as possible to weed out the non-essential from the essential and then tries to delimit the situation in which the vocable occurs. The larger the number of excerpts from diverse sources the better the result. In the limiting case of a single excerpt the work comes to an end and hence the difficulty of deciding the meanings of hapax legomena of uncertain etymology. The study of the purely linguistic context helps the lexicographer to refine further the meaning and its nuances by observing the other linguistic elements forming collocations with it and elements which cannot so collocate. This also helps him formulate the grammatical structure of that vocable and its various syntactic and stylistic peculiar-ities. Lastly, he compares the vocable with others of approximately the same meaning and sharpens the boundaries between them. This is the well-known' procedure of comparing the synonyms and antonyms to help define more accurately the meanings of words. All this procedure can be called the comparative method of 'semantic reconstruction' which confines itself to the collection of examples in a given language in which the vocable occurs and which is hinted at by MALKIEL
The lexicographer has very little to do with the epistemological and philosophical problems involved in definitions. There are various types of definitions which are in pratice. The Aristotelian definition consisting of the genus and the differentia is well known. Its value in lexicography, however, lies in the type of differentia that is used. A banal type of definition which is found in some lexicons should be avoided where the differentia is colourless like a kind of, a variety of, or a type of.' CAPPELLER gives 25 names of comparisons like samśayopamā, nirnayopamā, prativastupamā, given by Dandin in his Kavyādarśa with the same explanation (a kind of comparison'. Explanations like a kind of a tree, bird, animal or any other object, are too frequent in the lexicons and should be avoided as far as possible. The ostensive method which points out to an object which may be represented by a picture is no doubt useful and special dictionaries of this type called die Bildwörterbücher' are current for many living languages. For the classical languages they have a limited use and are rarely undertaken. Obviously the historical dictionary cannot make use of illustrations and pictures to any large extent. Very often the definition has to be couched in

Page #445

436
Amrita
the form of a rule, as when 'self is said to refer to the subject or 'l' refers: to the speaker. Some elements of the language, particularly the pronouns and their derivatives, admit of only such a definition. The denotative method consists in giving samples or tokens of a particular thing and has also a limited use in lexicography. The method of analysis is close to this and enumerates the various elements concerned and can be used to some extent. This was one of the early methods used in Indian philosophy and we often come across statements like kim jñānan, jñānam pañcavidham prajñaptam. The method of synthesis goes a step further and after enumerating the members of the set also indicates the special conditions of applicability to each one of them. But the methods which are basic to a historical dictionary are two : the method of synonyms from the language of exposition or from the same language and the contextual method which gives suitable examples to explain : the significance of the item under consideration. The lexicographer will use, in addition to these, any other method which he feels can help the reader to a better understanding of the vocable.
The problem of the selection of citations arises because of the bulk of the dictionary. If 1500 books are excerpted a common word is expected to be found in most of them and if the word has a couple of meanings the examples are likely to run into thousands. It is clearly impracticable and unnecessary to include them all. It will not only increase the bulk of the work beyond measure but may also prove to be a hindrance to the reader in finding out the meanings of words out of this vast collection. Hence a judicious selection of the citations and their systematic arrangement. The citations must include some of the clearest cases for deciding the meaning of the item, some of the well known quotations which have become proverbial, some of the more difficult and obscure or doubtful ones as a subject of further study and a number of citations to represent the use of the word in various literary branches and in different periods of the history of the language. They should give a fair picture of the provenance of the item. The chronological order is essential and will need modifications where imitation or secondary use is involved and where the citations from the commentaries are used. What minimum gap is to kept between two citations will depend upon the specific nature of the word under study. The citations should also serve another purpose, namely to supply as much cultural information as can be got, out of them. In some cases the attention of the reader should be drawn to such facts after the citation, by adding remarks in brackets if necessary: It is not possible to translate the citations when they are so numerous and the

Page #446

Dictionary of Sanskrit on Historical Principles
437
literature so vast. As far as possible the quotation should be self-sufficient and understandable without outside help. But in many cases words in brackets will have to be added to indicate the context and in more intricate cases a brief statement of the context may be necessary.
The ideal way of fixing the number of citations would be to keep the same proportion in which they are found in the data, a kind of reduction in scale. But this is clearly impossible and rare meanings and usages are bound to be represented on a scale larger than the more common and frequent usages, as in the case of the width of a river or a road in a geographical map. The imbalance can be set aright with the help of specific statements which will be required also for fixing the status of the vocable.
In the case of difficult words, hapax legomena, words of unknown meaning etc. an attempt will be made to present the whole evidence available to the lexicographer because this will be the starting point for further study in the lexicology of the language. No dictionary can hope to solve all the problems in a language and all that is expected of it is the presentation of relevant information on such cases which require further study.
9. Status of words. The science of lexicography emphasizes the fact that all vocables of a language have not the same status in the sociolinguistic context. Labels like slang, elegant, archaic, popular, common, scientific, technical, suited to a particular type of discourse, belonging to a specific register, dialectal, standard and many others are often attached to words to indicate their status. A good dictionary, historical or otherwise, has to take note of this labelling and help the reader judge the value and the use of the word correctly. In the case of a classical language like Sanskrit, lack of adequate information on all vocables in this regard restricts the scope of such an approach severely and such information can be given only for a limited number of items and to a limited extent. Dialectal and regional vocables are few and far between and our information about them is of the scantiest. On the other hand, the literary type from which the word comes as also its temporal limits can be ascertained with great accuracy. Hence we can describe the words with labels like 'in philosophy, in grammar, in ritual, in dramas, in historical poems, in satires' and so on. Among the earlier lexicographers of Sanskrit, it is GOLDSTÜCKER who has used such labels extensively though not consistently. Thus he speaks of Vedic, in Vedic grammar, in the Vedas; Jains, in Jain doctrine, in the mythology of the Jains; Buddhists, in Buddhist literature, in Buddhist doctrine, in the Buddhist

Page #447

438
Amrita
doctrine of Nepal; in rhetoric, in rhetorical terminology; in philosophy, medicine etc. It is necessary to follow this lead more consistently and label the words or their meanings, mostly in case of the technical vocabulary with the appropriate reference to the branch to which they belong. This is the only aspect of words of which we can be certain. How fine the distinction can be made in this regard is a matter of choice. As far as I am able to judge a minimum of 15 and a maximum of 75 labels can be used to describe a vocable or a meaning as belonging to this or that branch of literature. Only in rare cases can we surmise the application of a sociolinguistic lable like belonging to a particular region, to a particular style, to a particular register, in a particular context, vulgar or dignified etc. Occasionally we get words of address different from words indicating relationship, or expressions of phatic communication, words used by a vita or a ganikā, words meant for the stage. etc. On the other hand, words with a specific meaning can be indicated as belonging to a particular school of philosophy or a śāstra. In fact, a good procedure in analysing the senses of vocables would be first to separate all such special and technical usages and then take the residue as belonging to the common usage. A few artificial or playful words are also met with and shortening of words plays some role in this respect.
The OED gives the following scheme for describing the status of words :
(1) obsolete vs current
(2) natives vs naturalised foreigners vs aliens.
(3) common vs literary (archaic, foreign, scientific) vs. colloquial (technical, slang, vulgar, dialectal, local)
(4) nonce words.
From among these only those italicized will have some use for Sanskrit vocabulary. Others will be needed only on rare occasions.
10. Idioms. There is very little that can be said about the peculiar usages of words and the idiomatic expressions they give rise to. Very little information about them is available in the Sanskrit literature as a whole and it will have to be diligently collected and presented. Classical languages are comparatively poorer than modern languages as far as idioms are concerned.
11. Cultural information. The inclusion of information about the things denoted by the words is another matter. Cultural, historical, archaeological,

Page #448

Dictionary of Sanskrit on Historical Principles
439
mythological, geographical and other information of a varied type will be needed to elucidate the senses of numerous words. Much will depend upon the inclusion of the names of countries, mountains, rivers, towns, villages, parks, forests, gardens etc., the names of kings, queens, famous persons, gods, goddesses, mythical beings, rites, customs, practices, beliefs etc. and the length to which their explanations can be given. Nothing is gained by describing a rite like abhiseka at great length or describe a complicated sacrifice like agnistoma in all its features. Identification of places should be done as briefly as possible and mythological and other persons should be described with the greatest brevity. The so-called laukikanyāyas, which are so frequent in Sanskrit, should be specially noted and explained.
12. Compounds. When both simple words and compounds are given as lemmata, the question of listing words with which a given vocable forms compounds gets 'simplified. All the compounds with a given word as the first member will naturally follow each other in the alphabetical order and need no special treatment. But compounds in which a given vocable occurs as the second member will be scattered throughout the dictionary and hence it is desirable that at least a list of such words which are important from the semantic point of view be given at the end of the article. This, however, requires the working up of the whole corpus before even the early letters can be edited. Hence it is thought best to give such lists at the end of the dictionary in the form of an Appendix.
13. Statistics. Full statistical information cannot be given in a dictionary of Sanskrit of the scope implied in the above discussion. This is due to lack of sufficient material in this regard. Full indices of the major works are not available and to prepare such indices for works like the Mahābhārata or the Rāmāyana or even some medieval digests of law will be a task of greater magnitude than the dictionary itself. Most of the available indices are confined to the Vedic literature, particularly the Samhitas, and only a few works from the later period. Hence such a treatment can be given for a few important and selected words and that too for the Vedic period. Most of the remarks on frequency (like common, popular, favourite etc.) are impressionistic. 6. THE TECHNICAL ASPECTS OF A SANSKRIT HISTORICAL DICTIONARY
Like other types, a historical dictionary has to make use of a large number of technical devices for its easy use and to economize as much space as possible without impairing its value. The larger the scope of a dictionary,

Page #449

440
Amrita
the greater the need for compactness so that the work does not become unwieldy and hence unusable. Even then, most historical dictionaries run into many volumes because the richness of the material of a language which is cultivated over many centuries or even millenia cannot be altogether reduced to small proportions.
The devices used to facilitate the use of a dictionary are: 1. the alphabetical arrangement of the vocables, 2. numbering of homonyms to keep them apart, 3. numbering of the various distinct meanings of a given word, sometimes in a series of subdivisions to bring out the interrelations between them, 4. transliteration of the lemmata, 5. use of italies for definite purposes, 6. cross references and a few others. The devices to effect compactness are abbreviations of titles of books and authors, abbreviations of grammatical terms, a compact mode of reference to original sources, punctuation marks and their pregnant use, and use of different types for the different parts of the material.
1. The alphabetical arrangement used in dictionaries is of recent origin and has now been well-established. It has proved so useful that it is now extended to other works of reference like encyclopaedia, compendiums, directories and many others. As far as Sanskrit lexicons are concerned, we know that the oldest such collections were arranged according to the meanings of words as in the first three chapters of the Nighantu. Words with the same or similar meaning, forming a kind of semantic field, were put together and were called synonyms. The textual tradition of the Nighantu has two sets of expressions for this purpose : for the nouns a genitive of the leading word iti rūpasya 3. 7, iti prajñāyāḥ 3. 9, or iti sodaśa rūpánāmāni, iti ekādaśa prajñānāmāni and for the verbs only one type like daśāttikarmanaḥ 2. 8, iti kruddhyatiikarmāṇaḥ 2. 12. These collections later grew into synonymous Kośas like those of Amara, Halāyudha, Hemacandra and others. The naigamakānda of the Nighantu, however, enumerates individual words and Yāska makes an attempt to explain their meanings with the help of their etymologies. When Yāska feels that two words are of the same form but different meanings i.e. homonymous he refers to the second with the word idam api itarat as in the example madhu somam ityaupamikam madyateh / idam apītaran madhvetasmad eva Nir. 4. 8. Such collections must have given rise to the later homonymous Kośas where various meanings of a given word are collected together. Yāska is explicit in this regard at the beginning of the 4th adhyāya : ekārtham anekaśabdam ity etad uktam / atha yāny anekārthany ekaśabdāni tāny ato' nukr-amisyāmah Nir. 4. 1.

Page #450

Dictionary of Sanskrit on Historical Principles
441
• The Sanskrit lexicons make use of three distinct principles to arrange their material. The most external and mechanical principle is to use the length of the explanation, whether it covers the whole stanza or one half of it or only a quarter. This method is used by Ajayapāla in his Kośa, by śāśvata in his Anekārthasamuccaya and by some others. A second principle used by the lexicographers is the number of syllables a word contains as, for example, in the Anekārthasaṁgraha of Hemacandra. Here some of the Kośakāras regard the final consonant as a separate syllable because of the method of writing which is mostly syllabic and in which the last consonant has to be written as a separate letter though.marked with the sign of virāma. The third principle which they use can be called partial alphabetization. Here many of them arrange their words according to the last letter or sound of the word. Amara in his Nānārthavarga uses this principle of grouping together words ending in k, kh, g, gh etc. upto h. Sometimes the initial sound is used to arrange the words and Hemacandra goes to the extent of utilizing the first consonant and the following vowel in this respect. Often these principles are used together while some, like Halāyudha in his Abhidhānanamamälā, appear to follow no principle at all in arranging the words. In any case, a thorough alphabetization was not known to these early lexicographers and only recently has this method been adopted consistently and thoroughly for the preparation of the modern type of dictionaries.
The medieval and still later works, particularly the ekāksara-kośas and the mätrkākośas often use an alphabet consisting of two additional vowels, am and ah at the end. But the generally accepted alphabet for this purpose is the one consisting of the vowels from a to au and the consonants from k to h. Among these must be placed the written symbols of anusvāra to be used before the sibilants and h and also before semi-vowels y, v, r and l. In all other cases the parasavarna has to be used and alphabetized accordingly. The best way to locate it then is to consider it as the first consonant with an independent place in the process of alphabetization. The order of the sounds y, r, I and v is conventional and is generally followed though the older order appears to be y, v, r, I, as can be guessed from the order of the corresponding vowels. The visarga, whenever it is written, is best regarded as coming at the end of the vowels and is given an independent position in arrangement which is suited to its mode of writing. Again conventionally it is written before the velar and the labial voiceless sounds and dentals followed by a sibilant and is so arranged in these cases. Though it is optionally written before the three sibilants it is alphabetized as if it were homorganic with the

Page #451

442
Amrita
following sound. Final visarga is restored to its original sound s orr as the case may be and is so alphabetized. Anunāsika vowels are few and can be best alphabetized after the simple vowels. In the forms of the Rgveda, ! and lh are alphabetized immediately after d and dh. But variation of l as ! in southern MSS is alphabetized after I. Other signs like jihvámūliya, upadhmānīya, yama etc. are explained under these vocables but are not used in the lemmata of the lexicon. Similarly the geminated consonants are so written and alphabetized only if they are etymologically justified. However, deviations from this principle are duly noted whenever authentic-ated statements to the contrary are found. Variations of spelling due to interchange of vowels or consonants are also noted.
The real problem of alphabetization lies in deciding whether a. strict alphabetical order is to be followed throughout or some exceptions are to be made in order to bring together words having a close connection with each other. Nobody denies the validity of the principle that words closely related or derived from the same source should be brought as close as possible. But this may lead to the disturbance of the normal alphabetical order. It is not possible to put together all the derivatives and related vocables in one place or under one entry, least of all in Sanskrit. The result will be a dictionary arranged according to the basic roots with all the derived words under them and only a few underived words scattered in between. MONIER-WILLIAMS informs us that H. H. WILSON once intended to compile a Sanskrit Dictionary in which all words in the language were to be scientifically arranged under about 2,000 roots and that he actually made some progress in carrying out that project. Such dictionaries are known for some languages but their scope and purpose are different and they do not constitute what is usually understood by a historical dictionary. Now and then, Madhava in his Dhātuvrtti makes an attempt of this type in some places (cf. 1.13) where the following vocables and forms are noted : mandate, mamande, mandită, mandurā, mandiram, mandram, mandārah, mandaraḥ all under the root madi, or (1. 91) where añgati, angam, angaḥ, anganā, agniḥ, āgneyam, agnāyī, agram, agrima, angārah, angulih, angurih are brought together under the root agi (=ang). In the absence of such a listing of all derivations at one place which will then require a long, elaborate and dictionary-like index to locate them, one has to make a choice of some vocables to be given close to the main entry and the rest in their normal alphabetical place. Any one who uses the dictionary of MONIER-WILLIAMS knows how often he has to locate a word with the help of remarks like 'see page and column.' The remedy which

Page #452

Dictionary of Sanskrit on Historical Principles
443
Monier-Williams) uses for this situation and which is followed by others is to set up one major series of alphabetization and under some of these items to set up another subordinate alphabetized series. Thus in the series gabha, gabholika, gabhvara, gam 1, gam 2, gamātra, gamb, gambhan, gambhärikä etc. all of which are given in the Devanāgarī script, immediately after gamis inserted another subordinate alphabetized series like Gat, Gata, Gataka, Gati, Gatika etc. in Roman script and then under Gata is inserted in the same article a third series of the second members of the compounds with gata, okalmasa, okāla, okirti, oklama....up to Gataujas in which the words beginning with vowels and forming sandhi with the preceding a come at the end. The device can be repeated as often as needed but will lead to a very complicated system difficult for the reader to follow with ease. Secondly, the decision about what should be included in the subordinate series and what in the principal one, will remain subjective and the reader, unless he chances to have exactly the same views as those of the lexicographer in this regard, will find himself often thwarted in his attempt to locate words.
A símpler solution is often used to solve this problem. A strict alphabetical order is followed for all the entries of the dictionary and then with the help of cross references or explicit statements the interrelation among words is indicated. BÖHTLINGK’S shorter dictionary follows this device and we find in the same basic series of alphabetized elements sequences like gat, gata, gataka, gatakrama, gabhīra, gabhvara, gam etc. The Random House dictionary of English also uses this device and covers by it such morphologically related elements as the verb 'to get', its past tense got, its participle'gotten', or the singular 'foot and its plural 'feet' with a crossreference or explanation in so many words. With a far simpler morphology which English possesses, this can be attempted for that language but will not be practicable for Sanskrit. If attempted, the dictionary will turn into an endless list of forms, alphabetically arranged with grammatical analysis and description of each word, with no gain whatsoever. A historical dictionary of Sanskrit has to use the second method and add sufficient cross-references to elucidate the interrelations of words. A rough and ready rule would be to include all the tiñanta forms under the root and all those which can be regarded as subanta forms be treated as separate entries. Of course, borderline cases will arise and some arbitrary but convenient decisions will have to be taken for forms of infinitive, gerund, namul etc.
cond me. A rous
and
is, of
2. The homonyms are usually indicated by adding prefixed numbers to them so as to keep them distinct from the numbering used for the various

Page #453

444
Amrita
senses of a word. What is difficult to decide is when to consider the items as homonymous and when polysemous and hence showing various meanings. This has been already broadly discussed above while dealing with the general principles of a historical dictionary. Here we may indicate a kind of a practical procedure to do it. All the meanings which are associated with a given phonemic shape should be put together and then classified. A word like nidāna will show the following meanings, irrespective of the considerations of homonymy, 1. a rope to tie the cow's legs; 2. a rope to tie the calf; 3. cause; 4. primary cause; 5. reason; 6. occasion; 7. basic texts; 8. occasion for an event; 9. occasion for the use of a mantra; 10. its application; 11. one of the causes in the chain of causation; 12. a kind of wicked meditation; 13. a kind of evil resolution; 14. the diagnosis of a disease; 15. name of a book of diagnosis; 16. kind of Buddhist sacred work; 17. end; 18. cessation; 19. closure; 20. purification. All these constitute its contextual meanings based on its use in a specific situation. Before proceeding further, one has to take up the question of setting up homonyms i.e. the number of basically different items of the same shape nidāna. A formal criterion would be to look into the derivations of the word from roots and to see whether they are different, Thus nidāna can be derived from
i. ni + dā to bind
ii. ni + do iii. ni + dai
to cut to purify.
As some of the meanings do show affiliation with the meanings of the roots used, it is possible to set up three homonyms here. Thus purification belongs to item (iii), end, closure and cessation belong to item (ii) and the remaining belong to the first item. Hence three homonyms are necessarily required and should be so indicated. Of the remaining 16 meanings an attempt has to be made to see whether they can be connected with each other and regarded as development from each other or are nuances of the same meaning. If we succeed in this, we set up only one item for them all, but if it cannot be done with the available evidence, we will have to set up additional homonyms to account for them. Additional evidence like difference in gender, number, pada, vikarana etc. may be then looked into to confirm the analysis.
3. The numbering of the meanings can be done in many ways. We can simply number them serially from 1 onwards and, as stated in the Shorter OED, the sequence should be strictly according to the time of the first

Page #454

445
Dictionary of Sanskrit on Historical Principles
occurrence of the meaning in the data collected and hence presumably in the language. A good example is supplied by the Oxford Latin Dictionary which gives the meanings of the verb capio as follows: 1. to take into the hand, take hold of. b. to take or pick up (arms) 2. to put on, assume (clothing, etc. also armour) 3. to take (food or drink) 4. to seize hold of, arrest (a person or his body) 5. to catch by hunting, fishing, etc. b. (of birds or beasts of prey) c. to gather, reap, get (crops or other produce) 6. to capture, seize (by military or sim. action) a. (men) b. (a place). c. (land, booty, spoils, etc.) 7. to enter or take up (a position). b. to reach, make (a harbour, etc.) 8. to take for oneself, take possession of, appropriate b. to extort (money) c. to exact (tribute, a penalty) d. to obtain (by mental processes). 9. to choose, select (for a stated or implied purpose). b. (w. pred, ace,) to choose, appoint (in a particular capacity). c. to take (an example etc.) d. to choose (a time); to seize (an opportunity) e. to adopt or form a plan. 10. to accept (something offered); to take (a bribe). 11. to get, obtain, be given, win. b. to derive (profit, revenue). c. to acquire or inherit (property). 12. to assume, take on (a form, quality. etc.). b. to require (a state or condition). 13. (of persons or things). a. to incur, suffer (inconvenience, injury etc.) b. to obtain, derive (a benefit, pleasure etc.) 14. to make a beginning, begin (at or with). b. to come to an end, finish, stop. 15. to undertake, take in hand, carry out (an action, process, policy etc.) b. to undertake, enter on (an office etc.) c. to take up arms, begin hostilities. d. to rest, sleep etc. 16. to entertain (a feeling etc.), adopt (an attitude). 17. to take a hold of, delight, charm, captivate (usu. of things). a. (act.) b. (pass.) 18. to fascinate, enthrall with love, enamour. 19. to move (by entreaty etc.) win over. b. to win over, (w. bribes etc.), corrupt. c. to overcome (by habit). 20. a. to take in, delude. b. to catch out, trap. 21. to take a hold of, afflict, overcome. a. (of physical conditions). b. (of feelings, states of mind). 22. lame, crippled, paralysed; blind, deaf, deranged. 23. (usu. of things) to have placed, impinge etc. upon it (them), receive. 24. to have within or upon itself, to contain. b. (gram.) to have as meanings or inflectional forms. 25. to (be able to) contain, have room for, hold. b. to have capacity of, hold. 26. to give scope for, hold. 27. to keep under control, contain. 28. to be capable of having; also of enduring. b. (of persons) to qualify for. c. (of things) to admit of. d. to command belief, be credible. 29. to grasp mentally, take in, comprehend (alone or w. animo etc.) b. to apprehend (with the senses). Others may consider this feature of first occurrences as not an overriding one and may regard the classification of the interrelations between the various senses as equally important. This can be represented by arranging the meanings in the form of a branching tree with

Page #455

446
Amrita
main and subordinate branches growing out of the main trunk of the tree. This can then be represented by a process of numbering by means of a complex symbolism consisting of Roman capitals, Roman lower-case letters, Greek capitals and lower-case letters, Arabic numbers so as to produce a complex symbol like $21bBA. (WACKERNAGEL II.] p. 28), $158b-ca (WACKERNAGEL III p. 54) etc. Very soon such a system develops into a complex picture and puts heavy strain on the reader to understand it. An example from the Latin Thesaurus and the Dictionary of Greek Epics may give some idea of this method. Here is the analysis of the Latin verb 'mergo'.
I. Literal meaning A. in a strict-sense : to dip in liquid etc. 1. without intention or desire to destroy a. Generally
a. transitive B. reflexive 7. medio-passive
8. intransitive b. Specially a. stars merging in the ocean
i. medio-passive ii. intransitive
iii. transitive B. with the idea of washing
i. in the secular use ii. Pagan use iii. Christian use
iv. in specific places 7. in ocean, rivers etc.
8. in medicinal or other things. 2. with intention or wish to destroy a. transitive c. with living beings B. with things

Page #456

447
Dictionary of Sanskrit on Historical Principles
b. reflexive c. medio-passive d. intransitive B. Broader use 1. to put in solids a. generalis
a. transitive B. reflexive
7. intransitive b. special
a. weapons in the body B. in bones y agricultural implements in the ground 8. in lower regions
€. reflexion, in human beings II. (i) in imagination
(ii) in comparison III. figurative sense, to drown, obscure-men and things A. without intention to destroy
1. with darkness 2. with light
3. with other things B. with intention to destroy 1. especially to kill
a. transitive
B. reflexive 2. generally to oppress, to press down a. transitive
a. alone
B. with ablative b. intransitive

Page #457

448
Amrita
One can also compare the analysis of the Greek word anēr. I. anēr as a contrastive concept I. 1 (contrasted) with woman I. 1 a the contrasting concept woman explicitly stated I. 1 ac functional relation I. 1 ad aa sex partner I. 1 ac bb husband I. 1 aa bbaa actual husband I. 1 ac bbBB possible or prospective husband I. 1 ac cc men in general as partner of women I. 1 ad ccad the masculine gender I. 1 ac ccßß any group of men I. 1 aß no functional relation I. 1 aß aa man and wife act (suffer) in the same manner I. 1 aß aaaa singular I. 1 ap aaßß plural I. 1 aß bb men and women differently affected I. 1 aß cc relations between man and wife I. 1 b constrastive concept woman implied 1. 2 contrasted with child I. 3 contrasted with gold I. 3 a antithesis I. 3 ad essential difference emphasized I. 3 aa aa in evaluating expressions I. 3 aa aaaa in general expressions I. 3 aa aaßß in special expressions I. 3 aa bb in non-evaluating expressions I. 3 aß essential difference not emphasized I. 3 b connections I. 3 ba in general expressions I. 3 ba aa common origin

Page #458

449
Dictionary of Sanskrit on Historical Principles
1. 3 ba bb co-ordinate I. 3 bp special expressions I. 3 bß aa positive sentences, as well as I. 3 bß aaaa plural 1. 3 bB aaßß singular I. 3 bb bb negative sentences, neither nor I. 3 bB cc either or 1. 3 c relations between gods and men 1. 3 cô men as objects of divine actions I. 3 ca aa relations of gods together with men (a man) I. 3 ca bb individual gods or groups of gods to men or a man I. 3 cß gods are objects of actions of men I. 3 cy spatial and functional relationship I. 3 és as sex partners of goddesses 1. 3 cɛ gods resemble (appear as) men I. 3 cɛ aa an indefinite one I. 3 ce bb a named one I. 3 ce cc a closely determined but not named I. 4 contrastive concept to animals, plants and objects I. 4 a animals I: 4 ad coordinations and antithesis 1. 4 aa aa functional relations I. 4 aa aaaa horses and men I. 4 aa aaßß men and dogs I. 4 aa aayy men and cattle I. 4 aa bb no functional relation I. 4 aa bbad men and animals suffer in the same manner I. 4 ac bb BB opposed to each other I. 4 aß men and animals come in relation I. 4 aß aa general expressions I. 4 aß bb special expressions

Page #459

450
Amrita
I. 4 b objects or natural and cultural spheres II. simple anēr II. 1 as generic concept II. 1 a in sentences and general statements II. 1 b in generalizing statements II. 1 ba plural without plural idea II. 1 bß singular and plural as collective II. 1 bß aa plural II. 1 bb bb singular II. 1 c in comparisons, hypotheses etc. II. 1 d extended : humanity II. 2 as designation of acting person II. 2 a name not known or unimportant (eventually placed after) II. 2 ad singular II. 2 bB plural II. 2 b represents individual name or special group name II. 2 ba singular II. 2 bp plural III. anēr with attributive, relative sentences and in apposition III. 1 characterising adjective or participle and with relative sentence III. 1 a with characterising adjective or participle III. 1 aa person of the action III. 1 aß aa unknown or initially introduced without a name III. 1 aa aaca unknown person, no mention of name III. 1 aa aaßß unknown and known with mention of name III. 1 ac bb definite known person III. 1 aß anēr as generic concept III. 1 aß aa as predicate noun or noun in apposition III. 1 aß bb of imaginary person III. 1 aß bbaa in sentences, proverbs, similes III. 1 ap bbBB agent of action as representative of a type

Page #460

Dictionary of Sanskrit on Historical Principles
451
III. 1 aß bbyy in the situations imagined III. 1 aß bb88 in comparison as partitive genitive after tis, as
adnominal genitive of general meaning III. 1 b with relative sentence III. 1 ba agent of action III. 1 ba aa unknown person III. 1 ba bb definite known person III. 1 bß in plural, indefinite many members of a definite group III. 1 by as generic concept III. 1 by aa anēr as predicate noun III. 1 by bb of an imaginary person III. 1 by bbaa in sentences, proverbs, comparisons III. 1 by bb BB agent of action as representative of a type III. 1 by bbyy in imaginary situations III. 1 by bb88 partitive genitive after allos III. 2 with pronoun pronominal adjective III. 2 a with demonstrative pronoun
2 b with possessive genitive of personal pronoun III. 2 c with indefinite pronoun III. 2 d with interrogative pronoun III.: 2 e with pronominal adjective III. 2 ea hekastos III. 2 eß allos III. 2 ey ho autos III. 3 with designation of quantity III. 3 a with cardinal number III. 3 b with ordinal number III. 3 c with indefinite numerals III. 4 with apposition and adjectives denoting groups III. 4 a with apposition III. 4 aa apposition in the naming of rank, profession, function

Page #461

452
Amrita
III. 4 aß apposition is in ethnicon III. 4 b with adjectives expressing groups IV. anēr as genitive attribute
In this dictionary it is thought advisable to use a system of three levels of meaning-distinctions. They are indicated by the use of Arabic numerals, the Latin capitals and the Roman lower case numbers and their various combinations. The maximum distinction of meaning is expressed by the first, the minimum by the last and all the intermediate grades are put together in the second mode of presentation.
4. It is again purely a matter of arbitrary decision whether the lemmata be given in Devanāgarī script or in Roman transliteration. It has been decided to give all the citations in the Devanāgarī script and the method of accentuation will be the one current in the book from which the passage is taken. It naturally follows that the lemmata must also be given in Devanāgarī. But it may not be possible to mark its accent according to any one current system to the exclusion of the others and a new system of accentuation for Devanāgarī, as is done by Pt. VISHVABANDHU SHASTRI, adds to the confusion. Hence it is suggested that there should be a Roman transliteration of the entry word and it should be marked for udātta or independent svarita only, with the usual practice of writing the acute or the grave accent. This will facilitate the use of the dictionary by those who are not familiar with the Devanāgarī alphabet. For a better and fuller use of the dictionary however it is but natural to expect the reader to know Devanāgarī and the rules of writing Sanskrit.
5. In order to reduce the chances of confusion and yet to include as much of explanation as possible to make the understanding of the words easy, it is necessary to separate those words of the glosses which can also function as translations of the vocables from others added for elucidation. This is done by including this additional matter in brackets so that the reader does not confuse the meaning of an item with its explanation. This will prove of great service for technical terms.
6. The use of cross-references will have to be as extensive as possible. It helps to tie down the whole work into a compact and organic whole and helps save space by avoiding repetition. They will be used to indicate the relation between words, the use of synonyms and antonyms and other related concepts which are likely to explain a given vocable. In fact, to carry out this technical device to good use, it may be necessary to prepare articles of closely related

Page #462

Dictionary of Sanskrit on Historical Principles
453
Orban
words together and then decide how and where the information should be distributed among them. No general rules can be formulated for this purpose.
Other abbreviating devices like a raised cipher before or after a word or the use of a tilde before a word to avoid repetition are useful provided the matter to be supplied is close at hand. In the larger dictionaries it may prove to be much better to avoid them and give the whole expression without the use of such devices.
To economise space and to achieve compactness, the dictionary has to use a number of devices. It makes use of abbreviations of the titles of books and the names of the authors while quoting them. A historical dictionary quotes them in such an extensive manner that these citations form the bulk of the work and hence abbreviations are of great practical advantage. But while abbreviating the titles of the books and names of authors some precautions must be taken. Brevity must be weighed against understanding and it is equally irksome to be forced to refer again and again to an arbitrary list of abbreviations to understand what the reference is made to. If the exact dates of the majority of the quotations could be given in the body of the text, as is done in the OED, it may not be necessary to know what the titles of the books are. The chronological sequence is clear enough without it. But this is not true of Sanskrit. No such precise dates can be given for the majority of works and what is important is to know whether a particular word, form or meaning is known to, say, Kautilya or Manu or is found in the Rgveda or the Atharvaveda or comes into use in the Satapathabrāhmana or is peculiar to a Buddhist or a Jain writer etc. This makes it essential for a Sanskrit dictionary to emphasise the source of the quotations than its probable date and hence the abbreviations of the titles of books and names of authors must be kept as distinct and easily identifiable as possible.
One way to do this would be to take up the abbreviations which are already current like RV. AV. TS. VS. ŚB. etc. and then add others for additional books. This has proved useful for a smaller list of works but may ultimately prove costlier for a larger corpus of say more than a thousand titles. In the Latin tradition the name of the author is given prominence and it is used for references, with the name of the book as a secondary indication. This has helped the Latin Thesaurus to be brief. But in Sanskrit, greater value is attached to the names of books than to those of the authors, because in many cases the authors are unknown or only traditional or doubtful. Hence we must give reference to books as primary and to authors as secondary.

Page #463

454
Amrita
Only if the author has only one book to his credit and nothing else, as in the case of Bhāravi's Kirātārjunīya or Māgha's śiśupālavadha, it becomes immaterial which name is chosen for reference. The name of the book is preferred for the sake of uniformity.
We can examine a few current abbreviations and see how far they are compact and immediately suggestive of the book. The following ones from MONIER-WILLIAMS are typical :
Alamkāras 1 (arvasva, by Ruyyaka) Alamkāras 2 (arvasva, by Mankhaka) Alamkāras (ekhara, by Keśava-miśra) Dhanamj (aya-vijaya) Dhūrtan (artaka) Dhūrtas (amāgama) Gathāsamgr (aha) Katharn (ava) Kathās (aritsāgara) Krishnakarn (āmrita) Kuttanīm (ata) Mālatīm (ādhava) Suvarnapr (abhāsa) Vedāntas (āra) Vrisabhān (ujā nātikā, by Mathurā-dāsa)
This type of abbreviations is not economical and maximally compact. A more systematic approach to this problem has to be made and the following procedure is suggested.
Most of the names of Sanskrit books are compounds of two or three words, names having a simple word being rare. In the compound names, the first word is usually more significant than the second and the second more significant than the third. The last member of the name is either a term of generic nature indicating the genre of literature like kāvya, nātaka, campū, śataka, astaka or a word expressing the scope or nature of the work like saṁgraha, sāra, sūtra, candrikā, pradīpa, prakāśa, bhāsya, vrtti, tikā, etc. Naturally the maximum information is carried by the first, and less and less by the second and the third word. Hence some general rules of abbreviations

Page #464

Dictionary of Sanskrit on Historical Principles
455
can be laid down.
1. If the work or author has a name of a single word of two syllables it is kept intact e.g. Māgha, Vyāsa, Prabhā etc. If it consists of three or more syllables it is abbreviated to the first two e.g. Bhāska (rī).
2. If the name consists of two words, the first is shortened to two syllables and the second to a single syllable e.g. Raghu Vamsa), śiśu(pāla) Va(dha), Kävyapra(kāśa) etc.
3. If the name consists of three words, the first is shortened to two syllables, the second to a single syllable and the third to its initial sound or letter e.g. CāruCa(ryā)ś(ataka), Devibhā(gavata)P(urāņa).
4.. Common prefixes like Bịhat, Laghu, Mahā, Vțddha etc. are uniformly abbreviated to their first syllable while the last descriptive elements like Carita, Kosa, Nātaka, Purāna, Samhitā, Sataka, Tantra, Tikā, Upanisad etc. are indicated by the initial letter.
5.-Further adjustment is made by dropping the inherent vowel a of the Devanāgarī script to differentiate between two otherwise similar or identical abbreviations.
The abbreviations of grammatical terms offer no difficulty and the usual abbreviations are adopted.
As regards the mode of reference the following procedure is used. References to Sanskrit works require from one to four numbers to locate a passage. A Sataka work will need a single number for the stanza, a work where references are made to page and line will need two numbers for this purpose and for some books four such numbers may be needed as for Taittiriya Samhita and the Satapatha Brāhmana. References requiring five numbers are rare and are avoided.
The best way to make the reference both economical and clear is to use different types of numerals at the four places needed to cover all references. The last two places are given by small Arabic numbers. The third from the end is given by lower Roman numerals like i, ii, etc., and the first is given by capital Roman numbers like I, II, III, etc. In case a fifth place is needed we should use capital Roman letters like A, B, C. Thus a citation from RV. or AV. will show a number like ii. 40. 2; xv. 12. 7 etc., from SatBrāh. II. v. 9. 7; from Raghuvamsa 12. 5; from Pānini viii. 2. 4; and so on.
Among the punctuation marks, a simple comma is used to separate

Page #465

456
Amrita
different words to express the same meaning while numerals, simple or complex, are used to separate different meanings. Two references are separated by a semicolon. Square brackets are used to give the etymology and cognates while parantheses are used to add words either in the quotations or in the explanations. Other punctuation marks are used as and when needed but a colon is avoided in Sanskrit texts as it is likely to be confused with a visarga. It is replaced by a short dash. All abbreviations are followed by a full-stop and only single quotation marks are employed, when needed.
Annotations :
1. The Computer in Literary and Linguistic Research, Cambridge 1971, p. 3. 2. A Typological Classification of Dictionaries : Problems of Lexicography 1962, IJAL
p. 18 ff. 3. Rey-Debove : Le domaine de dictionnaire. La Lexicographic, 1970. Paris, p. 30
ff.
4. Lieder des Rgveda. p. VI.
000
Dictionary of Sanskrit on Historical Principles
DCPRI. 1976

Page #466

The Five Mahābhūtas
A Semantic Analysis
By the word mahābhūta we usually understand the elements out of which all the material things in the world are formed or to which all the material things can be reduced. In this sense, these can be considered as the basic physical elements, the building blocks of the material world. The history of Indian Philosophy has given rise to two distinct trends of thought : One of them, which can be described as materialistic, in which the existence of these elements, is admitted as real, is, the view of Cārvāka School (also called Lokāyata). The other, which is traditionally traced to Brhaspati whose postulates are often quoted for corroboration, admits the existence of mental or psychological entities as being equally real, and can be described as idealistic philosophy. These nonmaterial entities may be either considered as having a fleeting existence, as is done by the Buddhist philosophers or they may be considered as permanent, as is done by the majority of Indian philosophical schools. A further trend of thought was developed later in which the mental or spiritual entity alone was regarded as really existing, thereby implying the non-existence of the material things, giving rise to a kind of monism. But the majority of the philosophical schools in India admitted the existence of both the mental and material elements and drew a sharp distinction between the two, and hence may be designated as realistic in nature. Among them a further distinction can be made between those who regarded the psychic element as being inactive, the activity being confined to the material elements as is done by the Sāṁkhyas, and those who assigned all the activity to the psychic element only, thinking the material elements as being inactive because of being insentient. The concept of the mahābhutas and their prototypes called the tanmātras play a vital role in all philosophical systems of India.
Though these are generally admitted as real elements, the systems differ among themselves as regards their exact number and nature. Five of

Page #467

458
Amrita
these are most frequently named, probably under the influence of the analysis to which the microcosm is subjected and wherein the number five plays an important part. Sometimes only four of them are given on the basis of everyday experience and insistence on their concrete form as is done by the Cārvāka, Jain and Buddhist Schools. There are however some indications of an earlier list of only three elements as can be seen from the Vedanta doctrine of trivṛtkarana as against the later pañcikaraṇa.
Both these views are based upon the supposition that each element has something of the others in its composition. The Ionian Greek philosophers also refer to only three such elements, water, air and fire as the source of all things (Gr. arkhe), unless we choose to interpret epeiron of Anaximandros as standing for space which is endless. We may choose to consider only two-earth and waters-if we interpret the passage from Bṛhaḍäranyaka Upanisad, V. 5. 1: apaevedamagra äsuḥ/itä äpaḥ satyamasṛjanta satyam brahma brahma prajapatim prajapatirdevān as giving the whole list in this concept of cosmogony.
This paper attempts to concentrate on the words mahābhüta and bhuta and assess their nature from three different points of view, which are generally used in the study of ancient literature. As Bruno Snell points out, the study of early Greek philosophy, usually called pre-Socratic, is to be based on the interpretation of the Greek texts with a view to understand them from a mythological point of view, both as literary works and as showing aesthetic values. Another is a semantic analysis, by looking at the development of the meanings of the important words, leading to the history of ideas or concepts; or from a philosophical viewpoint by fixing the context of the complex system of thought which they reveal. Here we can attempt only one of these. approaches, viz., the historical development of the concepts associated with words bhūta or mahābhūta with the hope that we would be able to clarify them a little better. Here the key-words are studied not in the way in which a linguist or a lexicographer studies them, an approach in which emphasis is. laid on the form of the words and their usage, but to concentrate on the idea itself, trace its origin and follow its development and ramifications.
The method which can be effectively used for this purpose would be to try to ascertain as accurately as possible the etymology of the words pertaining to the ideas in general, then to fix the scope by considering the contexts in which these words are used. Their places in the context of similarwords in the same language or a group of closely related languages are also ascertained, so as to build a broad semantic field, and assign a function to the given word. Thus taking up the basic concept one tries to classify it along

Page #468

The Five Mahābhūtas
459
different scales, which are known as the semantic co-ordinates like concrete or abstract, an individual or a generic type, etc., and build up the history of the meaning of that word. Each of these steps has its own limitations and hence for arriving at a reasonably acceptable conclusion, it is necessary to supplement it with material obtained from what can be tentatively called a comparative study of word-meanings, which attempts to build up a prototype for its use as does the historical linguistics for the form of the word. Our attempt will be confined to the semantics of the given word. For the present, only one aspect or two of mahābhūta, namely, its predominant use in the field of cosmogony and cosmology and its numerical scope are taken up.
Let us translate the word mahābhūta as 'basic element. Its use is most conspicuous in speculations on the early cosmogonies and cosmologies in the Indian and Greek traditions. These material elements seem to have grown in number from three to four and then to five and their primitive meanings are preserved in the two well-developed systems of Buddhism and Jainism, the two so-called heterodox systems. In other words, they mark an earlier stage of the concept of the mahābhūtas as compared to the other systems of Indian thought, and their primary meaning is better revealed in the growth of these concepts in the early Greek Philosophy.
. The traditional list enumerates them as prthivi (earth); āpaḥ (waters); agni (fire); vayu (air); and ākāśa (sky). In its reverse order we find them at Taittiriya (2.1): "tasmad vā etasmād ātmana ākāśas sambhūtaḥ, ākāśāt vāyuh / vāyoragnih / agnerāpah / adbhyaḥ prthivi". Originally they were called merely as bhūtas. The earliest occurrence of the word mahābhūta is found in the Aitareya Aranyaka (3.4) : imāni ca pañca mahābhūtāni prthivi, vayuh ākāśah apo jyotims. It is again found in the apocryphal 14th chapter of Nirukta the meaning of which is not at all clear. Another relatively early use is found in Charaka's śarīrasthāna (1.27.28) : mahābhūtāni kham vāyuragnirāpah ksitistathā/ In the Pāli literature it is found in the Samaññaphalasuttar of the Dighanikāya : ayam kho me kāyo rūpi cātumahābhūtiko mättäpettikasambhavo odanakummāsūpacayo .... idam ca pan me viññanam ettha sitam ettha padibaddham; in the Samyuttanikāya, 22.82 : cattāro kho bhikkhū mahābhūtā hetu cattāro mahābhūtā pañcayo rūpakkhandhassapañcāpanāya, in the Pañcaskandha Prakarana of Vasubandhu (as rendered in Sanskrit) yat kincit rūpaṁ sarvam tat catvāri mahābhūtāni catvāri ca mahābhūtāni upādāya. Most of the earlier Upanisads and Jain and Buddhist works use the simpler word bhūtāni as in Taittirīya 3.1 : yato vā imani bhūtāni jāyante yen jātāni jīvanti yat prayanti; Chandogya, 1.9.1 : sarvāṇi ha
..

Page #469

460
Amrita
vă imāni bhūtānyākāśādeva samutpadyante ākāśam pratyastam yānti. (We should note particularly the words sarvāni “all” and ākāśa as the source); Brhadāranyaka, 2.4.2 : etebhyo bhūtebhyaḥ samutthāya tānyeva amvinaśyati (referring to idam mahadbhūtam), 4.5.13, Svetāśvatara, 1.2 : kālah svabhāvo niyatiryadrcchā bhūtāni yoniḥ puruşa iti cintyam /
In all other cases where the word bhūta is used in the eighteen Upanisads it means a creature or a being and not the material element. In the Pāli works, the word, usually used for the material element is not bhūta and dhātu. It is highly instructive to read Buddhaghosa's commentary Papañcasüdani on Majjhimanikāya which says : tatthāyam bhūtasaddo pancakkhardha - amanussadhātu - vijjamāna - khiņāsava - satta - rukkhādisu dissati. These seven meanings are explained by him with passages from the Pāli canon as (1) (pañcasu khamdhesu); (2) goblin' yanidha bhutani samāgatāni); (3) the four material elements (catusu dhātusu); (4) any inanimate object (bhūtasmim pācittiyam); (5) as the predicative use of the verb (bhū kāladhaso bhūto); (6) all beings (sabbe va nikkhipimsasti bhūtāloke samussayam); and (7) the vegetable kingdom, particularly the trees or plants (bhūtagamapātavyatāyā).
The conclusion from this evidence can be easily drawn. Mahābhūta is the latest term to be used for the physical elements, which are taken collectively and hence mostly used in plural. This should be distinguished from its other use, where it is found without forming a samāsa and means the great being' evaṁ vă are idaṁ mahad bhūtam anantam apāraṁ vijñanaghana eva - Brhadāranyaka, 2.4.12, immediately followed by the other use of bhūta in Aitareya : bhūtebhyah samutthāya. It thus replaces the earlier word bhūta in the same sense. The Buddhists found the word bhūta in the sense of the material element confusing, because the primary and usual sense of the word was 'living being'. Hence they replaced it with the word dhātu which has the required meaning of a material substance. In its turn the word bhūta had the meaning of an embodied being, a living creature and incidentally 'a tree' when it is thought to possess life. It may be noted that the use of the word bhūta in this sense does not make a distinction between the material body and the animating soul, and thus represents a stage where both were inextricably mixed up, a stage in which these two aspects were not separated.
In the Jain philosophical writings both in Sanskrit and Prakrit, the word mahābhūta is used while referring to the non-Jain systems of philosophy. Thus in the Süyagada, 1. 1. 1 called the samayajjhayana a reference to the Lokāyata system is found in the following two verses : santi

Page #470

The Five Mahābhūtas
461
āyachaddā
pañca mahabhūyā - ihamegesimāhiyā ete pañca mahabhūyā pujegātu / all of which refer to the Cārvāka system and again at 2. 1. 654: iha khalu pañca mahabbhūyā jehim no kijjai kiriyā / It also uses the word pañchamahabbuiye to refer to an adherent of this system. In all other cases the word bhuya or bhūyaim is used to refer to all kinds of living beings, mettim bhūeśu kappae bhūehim na virujjhejjhā and to groups of various grades of living beings Thanamga 3. 3 59: devāṇāgā jakkhā bhūyā and the word bhūyagāma is used as a collective term. In fact, the phrase sabbe pāņā, sabbe bhūyā sabbe jīvā sabbē sattā occurs hundreds of times in the Ardhamāgadhi canon. In the later philosophical writings exemplified by the Gaṇadharavāda of the Viseṣāvasyakabhāṣya of Jinabhadra the words pañcabhūya or bhūya are used to remove the doubts of the fourth, Gaṇadharam Viyatta kim manne pañcabhūya atthi va nathitti sansao tujjha (1649); pañcakkhesu na jutte tuha bhūmijalāṇalesu sandeho/ aṇilāgāsesu bhave so viņa kajjoņumāņão (1748). We may note in passing that Jinabhadra appears to believe that while the earth, water and fire are directly observable, wind and sky are only to be inferred being not perceptible. What he means can only be clarified in the context of the atomic theory of the Jain philosophers.
The usually accepted five material elements are rubricated in the classical philosophy of the Jains and are elaborated in the famous authoritative work, the Tattvärthasūtra of Umāsvāti in a peculiar way thus differing from all other philosophical systems. Four of them are included in the concept of pudgala which stands for matter in general and the last one called ākāśa gets its place along with two other elements called dharma and adharma as non-sentient entities and coupled with the sentient living beings, jīvas, they form the group of five astikāyas. This group is then opposed to kāla (time) which is regarded as an immaterial non-living thing. Altogether they make up the six dravyas and constitute the whole universe called loka tattvārtha, tatra lokaḥ kaḥ katividho vā kim samstho vā / atrocyate pañcāstikāyasamudayo lokaḥ (3-6). The space beyond the loka is called alokākāśa which is said to be completely void, which makes the place of ākāśa incompatible with the other four material elements. The latter forms a group by themselves to be contrasted with the living elements on the basis of their atomic structure. The current word used for them is astikāya 'having body'. Kundakunda, in his Pañcāstikāya points out that the aņu or paramāņu is the cause of the four dhātus (another term for the four material elements) and explains that they are constituted by the molecular units formed of the atoms. Each atom is supposed to possess one of the tastes tikta, kaṭuka, kaṣāya, āmla, and madhura, one of the five colours sveta, pīta, harita, aruna and kṛṣṇa, one of

Page #471

462
Amrita
the two smells sugandha and durgandha and at least two pairs of touch, karkasa, mrdu, guru, laghu, sitosna, snigdha-rūksa the last two pairs being compatible with each other, which means that a paramāņu can be either cold or hot, wet or dry. It is also credited with various shapes like parimandala, trikona, caturasra, āyata, etc. This description of the atom appears to have a close resemblance with the description given by Leukippos and elaborated by Demokritos. In their view an atom is round and warm, white and rough, black and smooth, sour and angular, and sweet and large, round. This can hardly be an accident. This stage of thought is nearer the concept of animatism than animism, the first and second stages of the semantic development.
This complex theory of the atoms was not known to the earlier Jain writings, and they used the material elements as a part of their doctrine of Chajjīvanikāya the standard description of which is found in the 4th chapter of Dasaveyaliya, a mūlasūtra. The text says these six are pudhavikāiyā, ādakāiya, tedakāiyā, vanassaikāiyā and tasakāiyā. This view is completely animistic in nature. The elements are conceived as animated beings and only as an afterthought the body is thought to be made of the various elements. The Vaisesika theory that all the elements have a threefold form viz: śarīra-indriyavisaya-bhedāt and the view that the earth-body is what we find here, while the bodies of other elements are located in the Varunaloka, the Adityaloka and the Vayuloka is a result of this secondary interpretation. Akāśa being incorporeal is excluded from this scheme. The interrelation between the views of the Cārvākas, Buddhists, the Jains and the Vaiseșikas is not yet fully worked out in detail.
Before we are able to decide the exact stage of semantic development of these elements called bhūtas it is necessary to fix the earliest meaning of this word. It occurs in all Indian philosophical systems and in the so-called scientific cosmogonies of the pre-Socratic philosophers in Greece. It may be pointed out that there is no common word found in Greek to cover all the elements together and they are considered as having their own phūsis. The surest example of the reconstruction of a root in IE period is supplied by the equation Gr. phūsis and Skt. bhūta and a number of verbal and adjectival forms derived from this root.
These can be listed as follows:
Skt. abhūta, Gr. ephū aorist 3rd sg., Skt. Bhūyät, Gr. phuiz benedictive, Skt. babhūvān, Gr. pephuos perfect part., and its fem. Skt. babhūvūsi, Gr.. pephunēa. From these forms it is easy to set up the equation IE bhe a - Skt. bhū, Gr. phūo. There is a minor point of sound change in this, because the vowel is long in Sanskrit but short in Greek. The most likely explanation for

Page #472

The Five Mahābhūtas
463
it is to consider the root as having a diss- yllabic structure. The root vowel was originally short, which has been set up as a long ū in Skt., for purely technical reasons that the root is to be given in the form in which it occurs in the past passive participle. That the vowel was originally short is reflected in the sūtra of Pānini bhūvādayo dhātavaḥ, 1. 3. 1. where, in spite of Patañjali, the V- sound was a mere glide between bhū and ādi.
The main purpose of this comparison is to find out the original meaning of this root. Without going into details I am inclined to think that the suppletion between the two Indo-European roots bhe a and es tell us their semantic relationship. While as - has the meaning of a verb of existence, the root bhe a - expresses the idea of growth, development, change. This is reflected in the Pānini's Sūtra 2. 4. 52 asterbhūh by which the root bhū replaces the root as in all non-conjugational tenses, which can be explained only on the assumption that as has a stative aspect, confirming the fact as it is, while bhe a has a progressive aspect, stating what the thing is growing or developing into. The first is static while the second is dynamic. This is quite clear in Sanskrit where bhavati means 'it grows', while asti means 'it exists'. Prof. Burnel appears to think that the corresponding root in Greek and its noun form phusis means 'the stuff out of which a thing is made', thus having nearly the same meaning which arkhe has. Most linguists however believe that the root means 'to grow', and the noun means 'growth'. The meaning of phusis as 'nature of a thing is a later development due to its contrast with the other word thesis which means what is attributed or assigned to a thing, in fact, the association of a word with its meaning which corresponds to the Sanskrit concept of vrtti or saṁketa. Its exact Skt. correspondence is bhūti which never has the meaning of existence but that of growth and hence prosperity. On the other hand, the passive participle of the root bhe a gives us in Sanskrit bhūta which has been all along used to refer to living beings. Only on this suffixation can we explain the meaning 'a tree of this word as attested by Buddhaghosa in Pāli and the meaning of the Greek word phuton to mean a plant or tree as can be seen from its use in words like zoophyte - 'a plant resembling an animal and phyto-graphy which means descriptive botany. This original meaning plays an important role in the context of cosmogony both in India and Greece where we are told that one thing gives birth to another or one thing merges into another, which are the basic concepts of the cosmogony of the elements.
The speculations of a cosmological nature which are found scattered in the major Upanisads have been clearly summarised by Prof. R. D. Rana his well-known work, A Constructive Survey of Upanisadic Philosophy, in which

Page #473

464
Amrita
he has also drawn attention to similar ideas found in early Ionian philosophers from the seventh century onwards. Thales the earliest among them considered that all things came out of water. Anaximenes thought that air is the primary substance out of which arose all others by the process of manosis 'thinning and pullnosis 'the process of thickening'. An-aximander considered the primary substance to be infinite or indeterminate, calling it a-peiron ‘endless', which was intermediate between earth and water on the one side and air and fire on the other, from which developed all these four elements.
Heracleitos of Ephesos championed the claim of fire as the source of the other elements by saying that fire first transformed itself into sea and this changed partly into earth and partly into air (prēstēr) as can be seen from his fragments 20 to 26. With these can be compared Brhadāranyaka, V. 5. 1: . āpevedamagre äsuḥ / tā āpaḥ satyamassjanta. Chandogya, IV. 3. 1 : vāyurvāva'', samvargo / yada va agnirudvāyati vayumeva apyeti / yadā āpaḥ ucchusyanti vāyumeva apiyanti. We find in Chāndogya, VI. 8. 4 the suggestion that the first evolute was fire from which came water and food (which stands for the earth) on the analogy of the root and the shoot of a tree : evameva khalu somya annena śungena āpomūlamanviccha, adbhiḥ somya sungena tejomālamanviccha tejasā somya sungena sanmūlamanviccha. There are a couple of passages where ākāśa (is it sky ? heaven, or space ?) is considered as the source of all the other elements. Chandogya 1. 9. 1 : sarvāni ha vā imani bhūtāni ākāśādeva samutpadyante ākāśam pratyastam yanti. As thought progressed, the Upanişads suggested some abstract or psychological objects as the real source of all these elements. Like Anaximander's a-peiron ākāśa is. taken as the source of all. Similarly asat (Gr. mēon), sat (Gr., on), prāna, ātman and still later a creator was taken as the original point of departure. All these speculations suggest some further progress of thought, but do not explain or clarify the older ideas of taking the elements as basic, which certainly was the earlier stage.
To understand the conceptual background of these earlier cosmogonies, we have to look at the history of the meanings of the words used for the elements. In other words, we have to fix the original meanings of the names of these elements. A few facts about them may be stated without much discussion. The words prthivī, ap, tejas, vāyu and ākāśa appear to be their original designations. Only at a later stage, when these words lost their cosmological affiliations are they replaced by other words having the same meaning. From being of specific connotations, they acquired more generalised meaning and then it became unimportant which word could be used for them. One can compare

Page #474

The Five Mahābhūtas
465
the otiginal words with the words used in a text like Caraka Samhitā : kha, ap, vāyu, agni, ksiti, mahi and bhūmi are used only in the Jaiminiyopanisad, 1. 10. 10, while ksiti is found in the older Upanisads. Maitāyani has the list
the form akāśavāyvagnyudaka bhūmyādayah 6. 4; vāta is not used in this context, as also udaka. In Mundaka, 2. 1. 3 we read : etasmāt (purusät) jāyate prāno manah sarvendriyāni ca / kham vāyurjyotirāpah prthivi visvasya dhārini/ which is a very late form of cosmogony and these things never play the role of a source. Long ago Meillet showed that while udaka means water in a secular sense, āpaḥ has religious and cosmological associations. Of the two IE words for fire, agni (IE ognis) and pur, the first is the older being found in the marginal areas while pur (e.g. fire) is an innovation.
The situation is very similar to this in Greek as well. For earth both gē and khthon are used but the first gets a place in the cosmology as one member of the primeval pair, and is used as an element. Khthon cognate with Skt. ksmā mostly refers to the surface of the earth, as a place of habitation. In the Iliad, XIX-259 it is associated with the seen and the Erinys and occurs in the utterance of an imprecation (gē te Kāi helios Kāi Evineies). For wind aveimos (root - ave - to blow) is used which suggests its origin in breathing, while Sanskrit vāyu corresponds to aēr and originally meant mist, or lower atmosphere, as against the upper vault or firmament, which corresponds to Sanskrit ākāśa 'the shining one'. The word used in the building up of the Greek mythology is however Ouranós 'heaven'. Aither is used as a feminine noun to refer to the upper air or heaven. Thus air and sky did not occur as
different elements in Greek cosmogony. Here either one or other element is • taken as the original substance which in the Greek mythology is concerned as a living thing, and in no way different from the anthropomorphic gods.
The situation is slightly different in the Upanisadic cosmogonies. While commenting upon Chāndogya, III. 14. 1 : tajjalāniti śānta upāsīta the cryptic word tajjalān is explained by Sankara with the words, kathaḥ sarvasya brahmatvamityata aha tajjalaniti / tasmād brahmano jātam tejo' bannādikramena sarvam/ atastajjam / tathā tenaiva jananakramena pratilomatayā tasminneva brahmani līyate tadātmatayā slisyate iti tallam / tathā tasminneva sthitikāle’niti prāniti cestate iti /
Whatever the value of such an explanation, taking the word-element ja from jāta, la from līyate and an from aniti, it is obvious that the cosmogonies use two methods of evolution and involution to describe the process of creation or dissolution, which is only partly true of the Greek cosmogonies. Obviously the process is built on the model of the birth, existence and death

Page #475

466
Amrita
of a man or a living being. A close scrutiny of the wording of the cosmogonies reveals the fact that in all of them, the process of birth and also that of absorption or end on the part of these basic elements is used in the active sense, as something which they do. This means they are animate things acting on their own. In this context we should consider expressions like: tā apaḥ satyamasṛjanta tadaikṣata bahu syāṁ prajayeyeti, tattejo'sṛjata while. creating and apyeti, prayanti, abhisamviśanti while merging, which attribute to them conscious acts of origination and dissolution.
Let us summarise the semantic development discussed so far. The four elements or mahābhütas were originally considered as living objects like men and animals, and hence could do all the activities which were expected of them as living beings. This stage of thought is given the name of animatism. by the anthropologists while dealing with the origins of religion. At this stage' no distinction was drawn between the living and lifeless. This was the stage at which these four or five mahābhütas were conceived.
In the next stage occurred a differentiation between those who were endowed with the power of conscious activity which gave rise to anthropomorphism and there emerged mythological stories when these elements, particularly the earth, water, air and sky were personified to produce the stories of creation. This marked the second stage of thought where a distinction was drawn between the sentient element and the nonsentient part which was conceived as the bodies of these elements. This gave rise to what are known as the scientific cosmogonies of the early Greek thinkers, and also some of the early cosmogonies found in the Upanisads.
The third stage of development further emphasized the part played by the psychic element in creation and the cosmologies were so adjusted as to give to these elements the real activity of production, while their bodies were regarded as inanimate by nature. This completed the distinction between living and lifeless things. Further growth of thought centred round the ephemeral or permanent nature of the psychic element called the soul. The semantic history of the Greek word psukhe epitomises this development in its three successive meanings breath, life and soul, while others like ge, herdór, púr, ánemos, aither, or ouranòs became fixed in their meanings at the end of the second stage of development.
000
The Five Mahābhūtas
Prakrti Vedic, Buddhist and Jain Traditions

Page #476

Key-Note Address for a Seminar on Lexicography
Before we discuss the nature of Lexicography and its theoretical aspect lexicology, it is necessary to take a broad view of the subject in relation to linguistics on one side and philology on the other. The basic structure of a lexicographical work consists of the body of words of a given language, and the meanings assigned to them in the works written in that language, if it is a classical one, or the use to which it is put if the language is a living speech. The two extremes of the situation are the thoughts or ideas or existential objects on one side and the sound-elements on the other. A series of closely connected studies form a bridge between the two. They begin with the study of the physical sounds and the physiological activities involved in the production of human sounds of articulated nature which have no direct connection with conveyance of meaning or expression of thoughts and the understanding of the external world in its extreme complexity and its own structure irrespective of any linguistic expression for them. It is true that they are inextricably mixed up and are essential for understanding. The Greek philosopher Parmenides took the view that to gar auto noeīn estīn te kāi einai thinking is the same thing as being', which is not completely true.
A number of studies deal with the intermediate chain and try to clarify the relation between means of expression and the contents of thought, which are inextricably mixed up and often show the dominance of one or the other in varying degrees of importance. This situation can be made more intelligible by using terms and specifying their precise use. Thus we speak of phonetics, phonology, grammar, syntax, style and semantics, by giving more and more attention to the contents of expression as the series progresses. For our purpose we may simplify the picture by covering them with two broad words of common use, grammar and lexicon, the one containing what are known as phonology, morphology and syntax collectively called discourse, and the other consisting of lexicons of various types. This is how Bloomfield viewed the

Page #477

468
Amrita
whole process of communication as consisting of grammar and lexicon, the basic principle being the presence or absence of motivation, in the sense that being purely arbitrary or partially so. The primary task of the lexicographer is to deal with unmotivated elements, while grammar deals with the various means by which motivation is effected in practice.
To clarify the picture to a greater degree it may prove convenient to make use of a graphic representation of these various aspects of linguistic sciences, to which lexicography belongs. For this purpose we may adopt a scheme which was originally suggested by Prof. Ullmaun in his semantic studies with necessary modifications. We make use of three co-ordinates for such a representation. Along the vertical co-ordinate we distinguish three stages from bottom to top and call them, phonology, morphology and syntax, Along the horizontal co-ordinate from left to right we distinguish two stages named form and substance and along the diagonal two more stages called synchronic and diachronic in the ascending order. The degree of motivation increases as we proceed from bottom to top and from left to right.
Theoretically speaking the lexicographers should give more attention to motivated elements and only for specific reasons the more motivated elements of a language, leavings the other things to the grammarian of the language. In theory at least this will help the lexicographer to decide what he should consider to be the primary function of a lexicon and what the lexicological reasons is for this purpose.
This principle is probably implied by Bloomfield when he says that language consists of a grammar and a lexicon which now includes, the alphabet or the elements of the phonological system, the basic morphemes of which words are formed and the syntagmic rules by which they are combined in longer units viz., words, phrases and sentences. This is the significance of what is called addyoccāraṇa in Sanskrit, (grā), the given elements, viz. the sounds, the roots, the various grammatical morphemes called pratyayas the rules of syntagmas and their meanings on which the whole structure of the language is built. On these given elements the complicated structure of grammar rests. A better way to characterize Pāņini's Grammar would be to take up the given elements and with their help build valid strings of utterances. Given the basic unmotivated elements, the lexicon of a language, the grammarian builds the linguistic utterances to meet the heavy demands of expressing complex ideas found in actual life. I believe that Panini makes use of meaning to the extend needed for the explanation of formal differences.

Page #478

Key-Note Address for a Seminar on Lexicography
469
•Another principles of lexicography arises out of the limitations found in the rules of grammar which describe the grammatical features of the language. The best observed rules of a language are universal and are regularly observed by its grammar. They are the categories found in a language like gender and number distinction and are uniformly observed and can be best regarded as compulsory for the speakers of a language. Once a distinction between numbers like singular or plural and in some cases dual is made, it must be observed regularly. This is also true of distinction of tenses and moods and some times of aspects of verbal action, which once noted, are generally observed throughout the language. They are obviously compulsory rules of a given language. But very often actual cases occur where they are observed partially and are sporadically found. There are numerous causes to account for such irregularities. As remarked by Sapir all grammars leak and no grammar is completely free from them. Rules have exceptions to them and are found in almost all the languages and complete regularity is not found even in aritificial languages, much less in natural languages. The causes for such a situation are many, some of them are due to borrowing of words which keep the pattern. found in the doner language but are absent in the borrowing language and such exceptions are due to the borrowings. Other exceptions are due to archaisms where some patterns of the earlier stage continue in a few limited cases. Still others are due to analogical formations confined to a few cases. Such irregularities abound in all languages, making their predictability uncertain. In all such cases their legitimate place is in the lexicon where all the words in a given language are expected to be included. Thus irregular features as well as limited dialectal features must be included in the scope of a dictionary, so that the user of a language gets the necessary information about them. Such cases came under the designation of closed sets as against open sets, and must be included in the lexicon of a languag good and exhaustive grammar thus implies a good and comprehensive dictionary of that language which will help the grammarian to state these features or rules more accurately. An exhaustive and complete grammar as in the case of Sanskrit grammar of Pānini presupposes a complete dictionary of that language as a prerequisite, which may be in the form of collections as seen in the Dhātupātha and Ganapātha of Sanskrit grammar or produced by testing the ability of the native speaker to decide whether a given form is acceptable or not. Sanskrit grammar makes use of two expressions, a general one lokāt avagantavyam and anabhidhānāt when the form is not allowed. This can be called the principle of exhaustiveness of Lexicography. Unpredictable forms or features of a language thus fall in the scope of a lexicon, as in the

Page #479

470
Amrita
case of determination of genders of Sanskrit nouns. This explains why most Sanskrit Kosas state their task to be nämalingānuśāsanam.
In the attempt made so far to delimit the scope of a lexicon from the formal point of view, the word motivation is used in a very specific sense and needs a clearer definition. The wider view of the use of motivation in Philosophy leads to the problem of determining the cause, purpose or aim which prompts action. Greek uses two distinct verbs to differentiate the meanings of the concept of motion. One is the verb teléō 'to end, to finish' and is usually used to mean 'to die'. It comes from the IE root kwel = and corresponds to the Sanskrit root car to move' and develops the meaning of aim or purpose. The other root is Greek kineō to be in motion, which is cognate with Latin ciere to put into motion. This leads to the formal point of view, the word motivation is used in a very specific sense and needs a clearer definition. The wider view of the use of motivation in Philosophy leads to the problem of determining the cause, purpose or aim which prompts action.. Greek uses two distinct verbs to differentiate the meanings of the concept of motion. One is the verb teléō 'to end, to finish' and is usually used to mean 'to die'. It comes from the IE root kwel and corresponds to the Sanskrit root car to move' and develops the meaning of aim or purpose. The other root is Greek kineo to be in motion, which is cognate with Latin ciere to put into motion. This leads to the question of deciding the relation between the two Sanskrit roots car-to wander and car-to be in motion, both derived from the IE root kel-, distinguished from each other in having a labia-valor and a palatalized velar sound. Leaving this problem of ultimate unity or difference of the two roots and concepts aside, the principle of motivation used above may be explained by making use of the Sanskrit theory of nimitta 'cause' as applied to words and expressed in the two phrases utpatti nimitta and pravrtti nimitta, the one stressing the source of a word to ascertain its meaning and the other its use in actual practice of the speakers of a given language. The lexical principle of motivation used earlier is more closely associated with etymology than usage. As such it cannot be allowed to have the fullest scope to delimit the bulk of a lexicon. Philosophers like Wiltgenstein, Moore and others have emphasized the usage at the cost of origins, and this causes the lexicographer to find a way to reconcile the two, because he is in duty bound to deal with both the types of meaning in his work. I believe that this can. be done. Making use of the principle of motivation we can delimit the lexical units in such a way as to include much of the grammatical machinery which in popular parlance in less than a word and excluded numerous items which
-

Page #480

Key-Note Address for a Seminar on Lexicography
471
go by the names of derivatives and compounds. This imbalance has to be remedied in the interest of usefulness of a lexicon without of course making the theory on which it is based unnecessarily muddy.
To do this, let us take up for consideration the other essential work the lexicographer has to perform, besides fixing up the corpus of his undertaking. This is the obvious task of assigning meanings to the items included. This is indeed the real task to which the lexicographer must apply himself, and is the one for which the lexicon is used. Presuming that the lexicographer can collect all the ideas which are expressed in a language as he collects all the words, which is not easy, we may lay down a principle of inclusion and exclusion of such elements in the lexicon. Just as the lexicon will not include all the sentences which are used in a language in a serial order and give their meanings in the language of explanation as does a traveller's guide, it cannot include all the meanings to which the words are used in the given language. It is essential to put a limit to them in one way or the other but it must be based on some acceptable principle. This may be tentatively called the semantic principle to be observed in this regard and it must be justified on theoretical grounds and not only on practical considerations.:
This task can be done by finding out and using methods by which meanings in use can be sorted out for inclusion or exclusion in the dictionary.
To carry on this task, the theory of word meanings developed in Sanskrit works on rhetorics will offer a well-come help. The meaning current 'in language are classified into three groups, and are designated as vācya, lākṣanika and vyangya. Of those the last called 'the suggested meaning is generally excluded in the dictionaries when they are not linguistically controlled. While the so-called Vācca meaning based on the expressive power of words finds a legitimate place in a lexicon. The case of the so called indicated meanings or laksyārthas are partly included and partly excluded. The so-called nirudha lakṣaņā has a legitimate place in a lexicon and is generally given as a figurative sense of words. But the other type of indicated meaning which is called prayojanavati laksanā based upon some definite purpose in a particular situation should be excluded from the scope of a dictionary. Unfortunately these two varieties of indicative meaning do not correspond with the distinction which the western lexicographers draw between what is called metanomy and figurative sense.
To draw a clear line of demarcation between the various meanings in

Page #481

472
Amrita
which words are used for inclusion in the lexicon, we should make use of the theory of environment of meanings. It is a basic principle that the meanings conveyed by words have no natural relationship with words themselves as physical entities. But this relation is not completely arbitrary. It is a well known fact that the bonds between the phonetic form of a word and the meaning it conveys has no natural relationship but a merely arbitrary one based on association over a long period of time in the course of the history of a language and is made firm by its acceptance by the linguistic community usually consisting of large numbers. No doubt there are a number of meanings assigned to a word in all languages which are purely arbitrary, but they become valid only when the community of its speakers accept it. Hence they are included in the scope of a lexicon.
It is an observed fact that the chain of phonetic elements of an utterance can be easily separated into their individual forms, but the meaning of the whole utterance is not so easily and surely segmentable. The problem of deciding the question of where the central concept of a sentence-meaning lies either in the finite verb, or the door of an action or the purpose of the utterance is differently solved by the grammarians, the logicians and the ritualists in Indian Philosophy but it plays no part in lexicography and has no relevance for the understanding of the language. In its place a more practically useful linguistic principle must be used.
Such a principle can be found in the idea of a focal point and its environment. Meanings of words are not isolated units but form a part of a wider field of semantics, and just as words are separated from the continuum of sounds to be segmented into units, so also the whole sense of an utterance must be separated into units of meanings to be assigned to corresponding units of sounds to become legitimate units of language. The resulting picture can be compared to a wave motion in a medium like water, or air. Just as the continuous wave is divided into segments called the wave length either between the two adjoining high or low points in it, so also the meaning of an utterance is divided into semantic bundles of some definite scope which together make up the sentence-meaning. When we take a small segment in this continuum as a point of reference, the others contribute to its environment, which mutually decide the scope of both. Using the principle of: trial and error a convenient arrangement is arrived at for use. It is now easy to see that any segment of utterance associated with a segment of meaning gives rise to a meaningful lexeme. This is not purely arbitrary but is conventional being based on a long history of the language, coupled with the

Page #482

Key-Note Address for a Seminar on Lexicography
473
modes of thought current in the linguistic community.
The segmentation of the sound continuum of an utterance into convenient segments is a comparatively easy matter but not so the cutting of the units of the thought continuum. Here the basic facts of this semantic material is already segmental into various units according to the long tradition of a given language and are generally accepted when a language is learned. Any precision for this depends upon mutual understanding and adjustments, where a chosen element gets associated with all other relevant elements collectively called its environment. This environment of a linguistic element called a lexeme is of two types, linguistic and extra-linguistic, the second of which cannot be used to fix the meanings of lexemes. Hence the meanings given to words in a lexicon will be such as can be determined by the linguistic environment stated above and no extra-linguistics information can be used for this purpose. This clear-cut decision lays down the limits which the assigned meaning of a word must observe and must be restricted to it by the assigned meaning of a word must lexicographer.
OOO
Key-Note Address for a Seminar on Lexicography
Unpublished

Page #483

A Comprehensive and Critical dictionary of
Prākrit Languages (Introduction)
1. THE TITLE
The full title of the present lexical work is 'A Comprehensive and Critical Dictionary of the Prākrit Languages with special reference to Jain literature'. This designation is chosen to indicate the main features of the work. Compared to the available dictionaries of the Prākrits, old and new, this dictionary is intended to cover all the available lexemes in the Prākrit languages and as such aims at comprehensiveness as far as the Prākrit literature is concerned. It is meant to be critical in the sense that while using the literature for this purpose, a critical attitude is consistently followed, and each item is examined and all quotations for it are thoroughly corrected wherever necessary with all the available material, and are fully interpreted. This is essential because most of the Prākrit works which are at present available are not critically edited and properly translated. In the case of editions of the canonical texts, there are too many differences among them, because they are published at different times and are worked with divergent principles which are often contrary to each other, being based upon different views of their nature. Equally chaotic is the method used for the purpose of giving references to passages and many of the entries are based on memory of the dictionary-makers. Sometimes an attempt is made to put a Sanskrit word into Prākrit without any thought being given to its actual occurrence in the Prākrit language. Short of preparing complete critical editions of the works used for this dictionary, an attempt is made to present the words and examples in as critical a form as possible with the available material. A consistent and easily verifiable mode of reference is used which will help the reader to trace the passages in different editions. The meanings given are based on a thorough consideration of all the evidence available including the traditional interpretations and for this purpose extracts from the commentarial literature are given in their proper places with references. By the expression

Page #484

475
A Comprehensive and Critical dictionary of Prakrit Languages (Introduction)
Präkrit languages' it is intended to take the word in the sense in which it has been used in Indian Classical literature as a whole and more particularly its use as found in the traditional Präkrit grammars, works on rhetorics and dramaturgy. In this sense the word 'Präkrits' is not co-extensive with what the modern linguists call the Middle Indo-Aryan languages (MIA). Hence we exclude from the scope of this work the whole of the Päli literature and extra-Indian Prakrit dialects and some of the early inscriptions in MIA like those of Aśoka. This exclusion is based on two considerations: (1) excellent dictionaries are available for Päli literature and an exhaustive one is already in progress. Glossaries for the excluded dialects and inscriptions are also known and meet the needs sufficiently well. (2) Technically the different modes of writing of these dialects make it difficult to include the words in the alphabetical order of the Prakrit Dictionary and their inclusion as separate entries will only increase the bulk of the work without real advantage. However, whenever an etymology of the Prakrit word is attempted, this material is given due consideration. The reference to Jain literature is intended to draw attention to the fact that an extensive material is available. here and that the semantic analysis of the peculiar and technical words from this literature is expected to yield a rich harvest and is in need of some amount of special explanation to understand them. The position which the Apabhramsa literature occupies in the history of Indo-Aryan languages fully justifies its inclusion in the Dictionary. All the material bearing on the socalled Vibhāṣās 'sub-dialects of Präkrits' being scanty will be included in the scope of this work. But the bulk of the Dictionary will be formed by the chief Prakrit languages namely Ardha-Mägadhi, Jain Mähärästri, Jain-Sauraseni, Mähärästri, Sauraseni, Mägadhi and Apabhramsa and illustrations from them will be used throughout. A complete list of books used for this Dictionary is given separately along with the abbreviations used and the mode of reference followed. It contains nearly 500 works and thus covers almost all the material available in the Präkrit languages.
II. LEXICAL MATERIAL IN PRÅKRIT
For a better understading of the nature of the present dictionary, it may be useful to give a brief survey of the lexical material available for the Prakrits included in this dictionary. As compared with Sanskrit and Pāli, the traditional lexical material for the Prakrits is definitely scanty and not of much significance. It is true that the vast commentarial literature on the AMg. canon and post-canonical works both in JM. and JS. often cite a few passages which appear like bits of Kosas of the traditional type, just as they quote a few grammatical rules which have given rise to the hope of finding old

Page #485

476
Amrita
grammars of Prākrit written in Prākrit itself, as is the case with the Pāli: language. But in both cases the expectation is not justified. In the earliest exegetical works in Amg. and Pro-canonical works in JS. and more particularly in the so-called Nijjuttis and Bhāsas, where such bits of statements occur, those are due to the tradition of these works to explain the texts in a peculiar manner. Of the various devices used to elucidate the meanings of the sacred texts, there is one called egatthas i. e. giving words all of which have the same meaning. As the idea is to classify the concept which underlies. a word more than its exact sense, these bits of eggatthas are built on a very loose idea of synonyms and a collection of words is put together to include as many aspects of the concepts as possible, which can hardly be called a Kośa as usually understood in Ancient Indian literature. A fairly evolved example of this can be seen at the beginning of each chapter of the late canonical book Panhāvāgaranāi. In the very first chapter we come across the word pāņivaha and its 30 names like pāņivahań, ummulană sarirão, avīsambho, hiṁsavihimisā, akiccam, ghāyaṇā, māraņa, vahaņā, uddhavanā, tivāyanā, ārambha-samārambho etc. called goņnāņi nāmāņi. The Nijjutti on the Suyagada gives purely phonetic variants of its name as : tass ya imani nāmāņi/ suttagadań, suttakadaḥ, sūyagadaṁ ceva gonnāi 2.. The later work called Angavijjā abounds in such collections of synonyms and a modern work : called Ekārthakośa puts them together. In spite of the usefulness of such passages, they can hardly be called Kośas or excerpts from Kośas in the usual sense of the term as seen in a very large number of works in Sanskrit beginning with the work of Amara. Real Kośas in this sense are only two, Dhanapāla's Pāialacchīnāmamālā and Ratnāvali (later known by the name Deśīnāmamālā) of Hemacandra. The first is a small work of 279 gāthās and deals with nearly a thousand words in Prākrit containing mostly tatsamas and tadbhavas along with a few Deśī words. The author tells us at the end of the work that he composed it in VS. 1329 and mentions a famous incident which occurred in that year, viz. the town of Mānyakbeta was attacked and looted by the king of Mālavā. The author himself lived in Dhārā and wrote his Kosa for his sister Sundari. He also indirectly gives his name in a line by the simple device of listing words, the end-syllables of which make up his name Dhanavāla. The choice of the words is made on the basis of their usefulness for writing poetry. It is obvious that he wants his work to be considered as an aid to poets, which thus belongs to the genre called Kaviśiksā.
The device used to arrange the words is equally simple which explains the limitations placed on the choice of words. Synonyms which number many but not exceeding a gāthā are given first, then those which can be accommodated in one line, or one carana or the fourth part of a gāthā, two

Page #486

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
477
lines of which are divided into two parts each of unequal length. The list is concluded with words with their meanings in a single word. It can be easily imagined that such a plan would give very little scope both for the number of words to be included and the explanations to be given for them. Hence its use for a Prākrit lexicographer is very little.
On the other hand the Deśīnāmamālā of Hemacandra is of capital importance for Prākrit lexicography. While the tat-sama and tad-bhava words of Prākrit are easily identified with their Sanskrit counterparts and offer no difficulty of explanation, the so-called Deśī words, which occur throughout the Prākrit literature, are enigmatic as regards their source. and often their meaning cannot be decided, when not extensively used. Hemacandra has devoted one kośa of his to collect and explain this stock of words in his work originally called Rayanăvali but renamed by its first editor R. Pischel as Deśīnāmamālā as being more expressive of its nature. It consists of eight chapters called vargas based on the division of the alphabet into groups according to a nongrammatical astrological tradition of India aufhall E 7 2116huruf for a weilfa- UHSETEGI p. 272. This divides the alphabet into vowels from 37 to 37:, ka-varga, ca varga, ţa-varga, ta-varga pa-varga and yavarga. Hemacandra has split the last one into two, the first containing y, r,"l, and v, the second s, ş, s and h. Naturally the Deśī words do not show the sounds visarga, nasals n, ñ and y initially and hence they are not necessary for arranging the words on the basis of their initial sounds. He remarks
RG E 441444f rafa 7 faGl: p. 208; reet Micasa a 14:
4 p. 236; 378497 mod TOT 7 4490 sfa 164: u p. 313. Besides arranging the words in the alphabetical order of the first syllable he further arrange them according to the number of syllables they contain, i, e. as having two, three, four or five syllables in each group and this is again repeated twice on the consideration, whether they are having a single meaning or many (ekārtha, anekārtha). The ekārtha group naturally includes words having the same meaning, a synonymous Kośa, while the other is a polysemous Kośa. This brings his classification in conformity with the basic division of the Sanskrit Kośa works. Hemacandra has put together as many as 6000 words of this nature in 783 stanzas.
To his text he has added a Sanskrit commentary explaining the meanings of these words and illustrates their use in a large number of stanzas composed by himself. Pischel is very critical about these as having no literary merit, even after giving due consideration to the constraints under which they were composed. In spite of this limitation they are often of use in deciding the meanings of words when they are polysemous in Sanskrit itself and hence

Page #487

478
Amrita
they are often cited for this purpose in the dictionary.
Hemacandra makes two more points worth nothing. He explicitly says that he has not included in his work numerous words current in different parts of the country on the ground that they are too many to be listed. He has confined himself to such words as are actually found used in Präkrit literature. He also defines what be means by Deśī qualitatively in his stanza 3. There he lays down the criteria that they should not be justifiable by the rules of grammer, or not known to Sanskrit lexicons and not justifiable by interpreting their sense by the device of a figurative or transferred meaning. More particularly words derived from the socalled traices are also excluded. He thus makes a distinction between Deśī words not derived from Sanskrit words and Dhātvādeśas which are not genetically connected with Sanskrit roots, a distintion which is reflected in the further history of such words in .. the modern Indian languages.
All the modern Dictionaries of Prākrits are arranged on the alphabetical principle, but the way in which the alphabet is arranged differs from dictionary to dictionary. Particularly in the treatment of the Anusvāra no definite principle is followed and hence groups of words involving this sound are found at different places in these dictionaries. Nor is the principle followed by a particular dictionary used consistently throughout and hence the Anusvāra, if it occurs in the middle of a word, is not given the same treatment as is given to it, if it occurs in the first syllable of a word.
The largest modern dictionary which professes to deal with Prākrits is called Abhidhānarājendra, composed by Vijayarājendrasuri (1826-1906), edited by his two pupils Dīpavijaya and Yatīndravijaya and published between 1910 and 1924, in seven folio volumes and contains more than 9000 pages of a large size in two columns. The lemmata are given in Prākrit, but that does not mean that all such words actually occur in the Prākrit literature. For example the entry 345ca on page 3 is based on its Sanskrit counterpart 3facil and its meaning as current in the Mimāṁśā is given with a quotation from Vācaspatya. Its use in Prākrit is unknown. It quotes long passages from Prākrit works along with the Sanskrit commentary on them when available and even whole chapters are included in this respect. This procedure accounts for its bulk. It is more of the nature of an encyclopaedia and includes both Prākrit and Sanskrit material, which the author considers as useful for studying the Jain canonical works and their commentaries. It is obviously modelled on Sanskrit works like the 16Gh645h and the anger and as such has both the merits and short-comings of these works. In spite of its obvious defects, the work has been carefully scanned to get some items and meanings which are

Page #488

A Comprehensive and Critical dictionary of Prakrit Languages (Introduction)
479
not otherwise available. The material utilised is mostly based on the memory of the author and the MSS. used were not critically edited. Its bulk has certainly affected the judgements of both F. W. Thomas and W. Schubring when they speak of this work in their reviews of 1924 and 1935. The strictures passed on it by Pt. Hargovind Das Seth in the preface of his Dictionary in 1928 appear to be not fully justified, as claimed there that a name of a work of Yaśovijaya 345rehenfchel in Prākrit is coined by the author.
The author has used about a hundred books in Prākrit and Sanskrit and has also listed the words under which the bulk of the information is to be found in the introduction. But their number is very small. The introduction also includes the whole of the Prākrit grammar of Hemacandra with a new commentary in Sanskrit verses written by the author, which however gives no new information.
During the thirties of this century two more dictionaries of Prākrit were published. Unlike the work mentioned above, they were planned as regular dictionaries of the modern type, arranged alphabetically and the meanings of the words explained with a few examples. The various meanings which the words conveyed are systematically numbered and separated from each other. The usual parts of speech are indicated and the verbal bases are clearly marked. The words are collected from the literature and they are given with their Sanskrit equivalents based on the rules of changes of sounds which the Prākrit grammarians have formulated. Of these two, the ArdhaMāgadhi dictionary was planned to include all the words in the Jain Svetāmbara canon or sacred books, overlooking the legitimacy of the books called . Prakirnakas. All the 45 books, which form the canon, were utilised. However, in spite of the fact that Ardha-Māgadhi as a Prākrit is confined to the canonical works only, some other works like the six Karmagranthas, Kammapayadi, the Višesāvaśyakabhāsya, the Pañcāsaka and Supāsanāhacariya were included. The reasons for this is hard to find. Originally planned by a non-academician Mr. Bhandari, who gives a graphic description of the difficulties met by the project during its execution, the basic work was prepared by the monk Shri Ratancandji Mahārāja, a versatile Sanskrit scholar famous for his extraordinary memory, containing an estimated 50,000 words taken from the canonical books, who also gave their meanings, both common and technical, in Gujarātī. As the plan originally conceived had decided to
give the meanings in three languages, Gujarātī, Hindi and English, in order to · have the dictionary a wider circulation, the Gujarāti meanings were rendered
into the other two languages with the help of a number of scholars who

Page #489

480
Amrita
themselves were not acquainted with the original texts and hence worked under a real handicap. The bulk of the dictionary, which consists of four volumes with a total of more than three thousand pages, is primarily due to this trilingual translations and the actual lexical items are far less than can be expected from the bulk of the work. Most of the shortcomings of the dictionary are due to the way in which the plan was carried out in very adverse circumstances. It appears that the original intention to confine the work to only the Ardha-Māgadhi Prākrit as found in the canon, which is also apparent from the fact that a new grammar of this Prākrit was prepared by Dr. Banarasi das and given at the beginning, was altered and a fifth volume of 857 pages was added, which included all the words which were found in the Paiasaddamahannavo of Pandit Hargovinddas Seth. This was done with the intention of making it cover all the Prākrit dialects, but which were not found in the first four volumes. The words in this fifth volume are given with only a Gujarāti rendering. This has deprived the dictionary of the possibility of giving all the meanings a word has, at one place, a great inconvenience for the user. The use of uncritical editions of the books, the insufficiency of the references given which makes it often impossible to trace them and lack of first-hand knowledge of the source material on the part of the numerous translators, have made the work less useful in comparison with its bulk and the labour spent on it. Quite recently the whole work is reprinted as it is, without making any attempt to remedy even its scribal mistakes.'
The other Dictionary of the Prākrits of the twenties is the most frequently used work called Päiasaddamahannavo, compiled by Pt. Hargovinddas Seth all alone. This work has served its purpose well during the last 60 years and more, and a second edition of it was published in 1961. More recently an abridged edition of the same for the use of students was issued under the same name. The abridgement was mainly done by dropping a fair number of words, mostly derivatives or compound forms, which can be easily produced by the rules of grammar. Otherwise all the three editions are identical in every respect and no attempt was made to either correct its mistakes or modernise it by additional material.
This works consists of about 1000 pages, each page having three columns, and includes about 75000 words. Meanings are given in Hindi and citations are few, a few more references being added to them. It uses as many as 170 books and represents the major Prākrit dialects except works from Jain Saurasenī. However 3RTERAR and carine are included, the purpose of which remains unexplained. Only a few Apabhramsa works were used for the simple reason that most of them were published after the Dictionary was

Page #490

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
481
completed. Citations from the Māhārāstrī epics are given mostly by a mere reference to the name of the book, when they have a word-index. (Probably the index to Setubandha was not available to the author.) The dramatic Prākrits are poorly represented and only an unpublished list of technical Prākrit terms in the dramas was made use of, and hence other words of the dramatic Prākrits were not included. In fact this has become a good dictionary for the three Prākrits, Ardha-Māgadhi, Jain Māhārāştri and Māhārāstrī.
All Prākrit words are followed by their Sanskrit equivalents, but no attempt is made to include their etymology and often the Sanskrit given does not appear to be correct according to the Sanskrit grammar. All words taken from the Deśīnāmamālā are called Desī, even when they have easily ascertainable Sanskrit cognates.
From the technical point of view the dictionary leaves much to be desired. The abbreviations of the works show that no uniform principle is followed : sometimes the name of the work is abbreviated; sometimes the name of the author. Thus the hymn Rsabhapancāsikā is indicated by Dhana which is an abbreviation of the author's name Dhanapāla. pi stands for Pischel to refer to the famous work Grammatik der Prākrit Sprachen. The collection of the Prākrit stories prepared by H. Jacobi is indicated by the siglum mahā, because they are said to have been in Māhārāstrī, a double confusion. There are many such curiosities which make it difficult for the reader to remember the abbreviations. .
The references are quite accurate when they are taken from the works · which the author himself has edited as in the case of Visesāvasyakabhāsya or
Surasundarikathā. The same is true of the voluminous work Paumacariya of Vimala. On the other hand, references to works like Vyavahāra or Brhatkalpa are very insufficient, the number of the uddeśas alone being given and in many other cases only the name of the book is all that is found. When MSS. are used it is but natural that the author could not give the reference to folio, side and line, unless he also specified the exact MS. which he has used.
A long introduction of 64 pages is added which contains a detailed grammar of the Prākrit in general and the characteristic features of the various Prakrit dialects like Saurasenī, Māgadhi, Paiśācī, Cülikā-Paiśācī and Apabbramsa, in which the author follows very closely the Grammar of Hemacandra.
While giving the meanings of words, numbers 1, 2, 3 etc. are used to indicate their different senses. But in most cases, lack of examples on which

Page #491

482
Amrita
this was based makes it very difficult to decide their correctness. The author has put great faith in the traditional interpretation given in the commentaries in Sanskrit, which has led him admit meanings which cannot be justified. But it must be admitted that, in spite of the deficiencies of the work, it is the only dictionary of Prākrits taken as a whole and it has served its purpose well during the last 60 years and more.
The only other dictionary which needs to be referred to is a publication called Alpaparicita-Saiddhāntikaśabdakośa consisting of five parts, published between 1954 and 1979 thus taking 25 years to complete. It contains a total number of 1256 pages and an additional supplement of 56 pages devoted to listing the words of the Deśīnāmamālā of Hemacandra in an alphabetical order with the meanings assigned to them by Hemacandra himself, and hence has no independent value.
This work is prepared by Muni Sāgarānanda, a profound student of the Jain Āgamas, and responsible for the editions of the major works of the canon along with the Sanskrit commentaries published by the Agamodayasamiti, These are the only reliable editions available today for many of these works. The commentaries on the canonical works were written by Haribhadra, Śilāńka, Hemacandra, Abhayadeva, Malayagiri, śāntigüri, Dronācārya and Candrasūri, covering a period of more than five hundered years, which was mostly devoted to the writing of Sanskrit commentaries. This dictionary takes note of all the 45 books of the canon along with the commentaries written on them and incidentally a few non-canonical works like Dasaveyāliyacunni, Paumacariya of Vimala, Uvaesamālā of Dharmadāsa, Tattvārthasūtra of Umāsvāti along with his so-called Svopajña Bhāsya and Višeşāvaśyakabhāsya of Jinabhadra. The later volumes were edited by Sāgarānanda's pupil Kāñcanasāgara.
Though intended to deal with the words found in the Agama texts, which are less known as is indicated in the title, the work is fairly comprehensive as far as the canon is concerned. By the word Alpa-paricita, it appears, the author took it to mean, not generally found in the classical Sanskrit literature or less known outside the Sanskrit works written by the Jain scholars.
The real peculiarity of the work lies in the fact that it is a PrākritSanskrit dictionary which also includes a fair amount of Sanskrit words occurring in the commentaries and their explanations given there. As an editor of the texts of the Āgamodayasamiti, one can well understand the procedure which the author has followed. For each word included (either in

Page #492

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
483
Sanskrit or Prākrit), he adds by way of explanation the relevant statements of the commentators in Sanskrit as they are actually found in the editions. As the same word occurs in a number of canonical books, the author has put together the explanations offered by the different commentators and even by the same commentator at different places. The references again are in the form of the number of the folio (the side being not indicated) because nearly all the books are in the form of folio-bundles or pothis. As is to be expected from the editor of these books, the page number is mostly accurate and one can locate the word and its explanation by reading at most the two sides of the folio. .
The author has made no attempt to weigh the explanations offered by the commentators and has not suggested his own views in this matter. But this is not an important loss because the reader has before him all the excerpts which he needs for this purpose. This method often gives rise to some problems like a choice of the meaning among the divergent explanations in Sanskrit. For example ambādei in Uttarādhyayana is explained by the word tiraskurute and in the Āvaśyaka by upalabhate. When a word like aṁsalaga is explained as aṁsagata, the etymological or derivational process gets obscured because the word consists of the stem amsala and ga from the root gam-, while in the other the stem is amsa and the other element is gata- a past passive participle of the same root gam-. When añāiyam in Viy. is explained with the words anātīta, anādikam, ajñātikam, ļņātītam and anātitam, the reader can make up his mind only by going through the original texts. When the Prākrit form aggei is explained as agneyi, (Tageszafacer) at Avaśyaka 315, the other form aggeyi as agnikona at Bhagavati 493 and Thānanga 133, difficulties arise because of the lack of context, which become clear only when we go to the originals and look at the whole context. Aggibhūi is explained as Agnidyotajīva at Āva. 172. When the passage is read in full one realises that Agnibhūti had a number of lives one after another and of them was Agnidyota, a Brahmin, hence the explanation ought to mean the soul of Agnidyota Brahmin' which was the same as that of Agnibhūti.
In spite of all such curiosities, the work is a very useful compilation to serve as an index not only to the canonical texts but also to the voluminous Sanskrit commentaries, which could not have been done by any one else better than Sägarānanda, who carefully edited all these works over a life-time. . There is also a fair amount of lexical material for some dialects of
it which should be noted here. This material is in the form of indices and glossaries with or without meanings, some of them being also without an

Page #493

484
Amrita
exact reference to the passages of the works on which they are based. In this regard a reference should be made to the indices of the Ardha-Magadhi canon published from Ladnun and to the indices of individual works, attached to the editions published in the Jain Agama series published from Bombay. But even here no uniform plan has been followed and the original divisions of the canonical works into Śrutaskandhas, Adhyāyas, Uddeśakas and Sutras is neglected. In both these publications the Prakirṇakas are neglected. Most of the published Apabhramsa texts possess good word-indices and recently an attempt is made to put them together in one work called a dictionary of Apabhramsa. Unfortunately the compiler has not taken trouble to avoid the deficiencies and mistakes of the individual indices and has simply repeated them in his work. Unindiced works have been simply neglected, thus impairing its value. Another small dictionary confined to the works of. Kundakunda in Jain Śauraseni is also available and is of some use.
In view of this situation the present dictionary has made it a point to scan all the words in these works and particularly the Ardha-Māgadhi Dictionary, the Paiasaddamahanṇavo and the Alpa-paricita-saiddhantikasabdakosa and to indicate their differences or mistakes under the corresponding words of its own.
III. LINGUISTIC BACKGROUND OF THE DICTIONARY
To understand the derivations and etymologies given in the dictionary, it is quite necessary to keep in mind the exact position the Prakrit dialects occupy in the general development of the Indo-Aryan languages. Frequent references will have to be made to the earlier stages of the languages as also to the later developments which the Prakrits have undergone. It is also necessary to fix the scope of each Präkrit used for this work and to indicate which works are included under a given name. It is hardly possible to explain the derivation of Prakrit words without comparing them with their OIA. equivalents and numerous words can only be understood in the light of their use in the NIA. languages. Hence a brief survey of the development of Indo-Aryan is given below. Relation of the Prakrit words with those found in the other Indo-European languages is rarely found and only incidentally discussed here.
1. Historical Linguistics
The science of language has succeeded in establishing what is called the genealogical classification of languages and thus grouping them into different families. This idea of a family of languages is mainly based on historical facts. As it is chiefly a historical concept, according to which a parent language undergoes successive changes resulting into the emergence of

Page #494

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
485
divergent but related languages at a later stage in which each member of the family is, in reality, nothing but a more or less modified form of the original speech, the method which can legitimately help build up such groups cannot but be historical, in which the earlier stages of the given languages are followed upto their common original form. In the absence of such material, however, the comparative method has to be used either to supplement it or, in rare cases, to take its place. We compare languages as regards their geographical location, their inherited vocabulary (naturally excluding borrowed words), inflectional peculiarities, types of syntactical groupings and, above all, phonetic correspondences amenable to more or less rigid application, to show their relatedness. But the extreme modifications, which a language may undergo in course of its evolution, make this procedure only lead to results of a positive nature. We can prove two languages as related but cannot assert that any two languages were not so related in the past.
If the similarities found in the languages help us in grouping them into families, sometimes of wide extent, the facts of common innovation lead us to put two or more languages into a more compact group, forming a subdivision of the bigger family. A language grows and assumes different forms, and each one of them may further repeat the same process. Thus we can well speak of the genealogy of a language with its nearer and distant relations, provided we remember that this is, after all, a figurative way of expression. 2. Indo-Aryan Languages
The Indo-Aryan languages (so called to emphasise the fact that they are of Aryan descent and belong to India, so as to be distinguished from other languages of Non-Aryan descent of the country and from Aryan languages current outside India) belong to the Indo-European (IE.) family (possibly itself forming a branch of an older family with Hittite), which includes other groups like the Hellenic, Italic, Keltic, Germanic, Slavic etc. They form an important branch of the Aryan group along with the Iranian branch with the inclusion of a possible intermediate group called Dardic. In the Indo-Iranian group itself the two main branches differ from each other more on account of innovations and rapid changes on the side of the Iranian languages than on the part of the Indo-Aryan group which on the whole shows itself more conservative.
The changes on the Indo-Aryan side, though less radical, are of greater importance to us. They include the change of ai and au to e and o (Sk. vēda Av. vaēdā, Sk rocáyati Av. raocayeiti); of ai and au into ai and au; the change

Page #495

486
Amrita
of palatal 2, zh into i, h (Sk. yajate Av. yazaite, Sk. hásta Av. zasta); the loss. of voiced sibilants like 2, § (Sk. nedistha Av. nazdiśta Sk. duruktám Av duzūxtam); the creation of a new series of retroflex sounds including the nasal n and the sibilant s; the creation of groups like cch and ks from various earlier groups. In morphology we have the form mama 1st per. pro Gen. sing. for mana of Avesta; the root brū -'to speak for the original mru- and many analogical forms like the 1st per. act. thematic ending -mi. The two branches also differ in the treatment of (Sk. sakrt Av. hak.rdt) and in the grade of the termination of the medial present part. (weak in Av. mna : bardmna, strong in Sk. -māna, bharamana).
While most of these early stages are reconstructed with the help of the comparative method we are on more secure ground when we enter the history of the IA. languages. From Rgveda (RV.) onwards we have a 'mass of.. documentary evidence marking the different stages in the growth of these languages. But the evidence is of varying value. Based on broad considerations, it is customary to divide the history of these languages into three stages, which are not strictly chronological in view of the fact that languages of the earlier stage continued to be artificially cultivated after they had become purely literary idioms. Thus we speak of the OIA. or Sanskrit, the MIA. or Prākrit and the NIA. or the stage of the modern languages. The OIA. : includes the Vedic language (wherein we can distinguish the language of RV. and of the other Samhitās), the Classical Sanskrit (of Pānini-Patañjali), the Epic language (of the Mahābhārata and the Rāmāyana), the Sanskrit of the Jains and Budhists and what we can infer about the spoken idioms of those days. The whole of this group derives its unity more on account of a rigid and well defined phonetic pattern to be adhered to, than common morphological structure, which on the contrary, shows a steady process of growth and simplification. This is strongly brought out by the attempts of later writers at hypersanskritisation and the use of such foreign words only as are in agreement with the Sanskrit phonetic system. Even then each one of these speeches has its own features which distinguish it from the others. Naturally the majority of them are grammatical as distinct from phonological.
The MIA. group includes a number of languages and dialects, all showing a strong family likeness and forming the next stage in the growth of the Indo-Aryan. These languages can be arranged into different groups according to the purpose for which they were used (those used for religious preaching and those for secular literature like dramas, epics and popular tales), according to locality of the dialects on which they are based (those of the North, of the Midland, of the East and South), according to the form in

Page #496

A Comprehensive and Critical dictionary of Prakrit Languages (Introduction)
487
which they are preserved for us (as literary, inscriptional and on other monuments like coins and idols) and chronological according to the more or less archaic or developed forms shown by them. Each one of these classifications has its value and helps us in judging the linguistic nature of the Prākrits in the form in which we now possess them. We can see how the use of Māhārāstrī for the composition of songs and lyrics and writing long epics making use of Yamakas and other tours-de-force, is partly responsible for the form in which we find it. The use of Sauraseni and Māgadhi with their subdialects and rarely Ardha-Māgadhi and Paiśāci (?) in the dramas side by side with Sanskrit has led to the effacement of finer dialectical differences and some overlappings between them. The inscriptional Prākrits are influenced to some extent by the form of writing in which the gemination of the consonants (length) remains unmarked and the order of the members of consonant-groups remains uncertain. Even the treatment of these Präkrits by the Präkrit grammarians in conformity with a definite plan has left some traces on their present form, the most important of which is the tendency to form wide generalisations on the basis of a few examples. With due consideration of all such points, the main task of the student of the MIA. is to ascertain the local distribution of these Prākrit dialects and to arrange them in a chronological sequence so as to mark the various stages of their growth.
In spite of the meagreness of the evidence, its conflicting nature and the theoretical difficulty of distinguishing between regional and chronological features, it is possible to arrive at some broad conclusions. To the North-West can be assigned the Mānsehrā and Shabhāzgarhi versions of Asokan
inscriptions, the dialect of the Kharosthi Dhammapada and possibly Paisāci .. and its sub-dialects. To the East lie the Aśokan inscriptions of Gangetic basin
and the mouth of Mahānadī, the Ramgarh inscription and the Dramatic Māgadhi with its sub dialects. To the West are found the Girnar version of Asoka, the Pāli of the Buddhist canon, the inscriptions of the Sātavāhanas and Western Kşatraps and the Māhārāstrī Prākrit. To the Midland belongs Sauraseni and to the east of it Ardha-Māgadhī, while a similar dialect is seen in the inscriptions of Asoka in the Deccan. Of course, this distribution neglects some amount of conflicting facts and leaves some dialects out of consideration. Attempts have been made to reconstruct the features of the earlier stages of some of the Prākrit languages available to us in their present forms. Such has been the case with Pāli as it is known to us at present, with Ardha-Māgadhi as it is handed down by the tradition and for Māhārāstri which is regarded as being closer to Saurasenī or even identical with it. But the results in this direction have been very meagre and pertain to a few phonetic features and grammatical forms. These remain mostly speculative in

Page #497

488
Amrita
nature and do not justify rewriting the traditional texts.
The classification in time is based on a firmer foundation and is more comprehensive. To the older stage belong the various inscriptions, Pāli, ArdhaMāgadhi and Paiśācī. A låter stage is formed by Saurasenī, Māgadhi, Jain Māhārāstri and on the threshold of the NIA. are found the Apabhramsa dialects and Avahatthā...
In contrast with the OIA, the MIA. group shows quite a different appearance on account of a drastic change in its phonetic structure. In grammer and syntax, the older synthetic mode of expression continues nearly to the end of the period, although the richness of forms of the older stage is lost and now and then, especially in the Apabhramba dialects, we see the rise of the analytical mode of expression. In their vocabulary, the MIA. dialects are mostly dependent on the OIA. but with a few additions of the socalled Deśī words of uncertain origin. The percentage of these depends both on the date and the nature of the work in which they are found.
The NIA. languages cover the whole of the Northern India and parts of the North-Western region, Deccan and Shri Lankā. They naturally admit of a ready geographical distribution. Their relation with the various MIA. dialects. however, is not easy to settle; nor can we group them in cognate groups with equal ease. Most of them date from a nearly identical period, though the first available documents in them belong to different centuries from the 10th to the 16th, while many have no written literature. Having continuous areas to cover, there are naturally many border dialects like Bhojapuri, Kachī. etc., the affinities of which are bound to remain doubtful Some of the languages of NIA. form distinct groups either on account of isolation, separation or other historical incidents. Thus Simhalese is separated from the main body of the IA. by the Dravidian tongues and follows its individual line of development. The Dard languages (including the Kafir dialects, the Khowar of Chitral, the Kohistānī dialects and the Shina group) are found to the extreme North-West and may owe their peculiarities either to an independent origin on par with Old Indian and Old Iranian or may be the result of a strong Iranian influence. Out of them only Kashmīrī came under the influence of Indo-Aryan civilization. On account of the nomadic life of the speakers, another of the Midland or a North-Western dialect assumed different forms in different countries. These are known as Gypsy dialects (Romani).
The Indo-Aryan languages proper are usually grouped into an inner and an outer band with a few intermediate languages, a classification which

Page #498

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
489
is much disputed but to which new support is sought in the treatment of the aspirates in the different NIA. languages. It is, however, more convenient and safer to follow a much more limited type of classification based on both geographical and historical considerations. To the North-West are found Lahndā and Sindhī, to the South we have Marāthi, to the centre belong Punjābi, Rājasthānī, Gujarātī and Western Hindi to the East of it the Eastern Hindi dialects, to the North of the Midland the Pahārī group and to the East belong the Biharī dialects, Oriya, Bengali and Assamese. 3. Prākrit Languages
The dictionary quotes extensively from the seven main Prākrit dialects to illustrate the meanings of words included and also makes a systematic attempt to trace the origins of the words in the Indo-Aryan branch of the Indo-Europian family of languages. For this purpose a brief survey of these languages may prove useful. The question about the peculiar features of each dialect and the relations among them make it necessary to name the dialects specifically when quotations from them are given. This is done by naming the dialect at the beginning and giving the name of the book at the end.
The bulk of the citations naturally comes from the Prākrits known as Ardha-Māgadhi (AMg.), Jain Māhārāstrī (JM.), Jain Sauraseni (JS.), Māhārāstrī (M.), Sauraseni (S), Māgadhi (Mg.) and Apabhramsa (Apa.), because the literature in these dialects is found in large measure. The main problem which a lexicographer meets in this case pertains to demarcating clearly one dialect from another and to decide to what extent minor dialectal differences in a given Prākrit can be ascertained and indicated. Obviously a given quatation from a given book must be assigned to the language in which it is written, at the same time taking note of the fact that many books use different dialects in different parts of the same work. Thus many books written in JM. include long passages and stories in Apa., and many quotations from the canon occur in JM. works. In fact it is often difficult to decide whether a given work is to be considered as being written in AMg. or JM. In the present case only such books as form the part of the Svetāmbara canon are cited under the siglum AMg. and others are considered to be in JM. This is obviously arbitrary and leads to such results as considering the Ogha and Pisda Niryuktis as being written in AMg. while the AvNi. and AyārNi. in JM. This problem is acute in the case of the Prakīrņakas and again an arbitrary decision is taken by including all such works attributed to Yeders in AMg. and those of definite authors in JM. In case of doubt, the reader should refer to the classification of books on the basis of their language, given separately. All Prākrit works written by the Digambara writers are given the designation

Page #499

490
Amrita
JS. while under Apa. are included works which are written by both the. Digambara and Svetāmbara writers, and a few verses found in the works of non-Jain writers.
In the case of other dialects, under Māhārāstrī are included the well known older epics, and more recent ones from the south and the majority of Prākrit stanzas quoted in the works on rhetorics. But the problem which needs a clarification pertains to the stanzas in Sanskrit dramas and the Sattakas and the use of Saurasenī, Māgadhi and various 90s and fans: which may occur in them. A brief history of this problem may help us to understand the situation, though the problem cannot be completely solved due to lack of necessary evidence in this regard. (i) Ardha-Māgadhī
From among the MIA. dialects the Svetāmbara writers have used two closely related languages, Ardha-Māgadhi and Jain Māhārāstrī, to compose their canonical and post-canonical literature. The relation between AMg. and JM. can be understood better by considering the history of these languages. Slowly AMg. has come under the influence of Māhārāstrī so as to develop into what is now called Jain Māhārāstrī; and in the course of a further development, this language is modified by the influence of Sanskrit on the one hand and the Apabhraṁsa dialect on the other.
The oldest form of AMg. can be seen in the first books ‘of Ācārānga and Sūtrakstānga and the older parts of other works like Uttarādhyayana etc. It is characterised, in its phonology, by the change of -r- to -l- to a greater extent (anelisa Ayār, 1. 9. 1. 16, Sūy. 1. 15. 2); the loss of initial y (ahāsuyam āyār. 1. 9. 1. 1; āvakahāe āyār. 1. 9. 1. 2; āvanti Āyår. 1. 4. 1. 2; āhattahiyam Sūy. 1. 13. 1) and the change of -kh- to -gh-(āghāi Āyār. 1. 4. 2. 1; Sūy. 1. 11. 23; äghāyam Sūy. 1. 1. 2. 1). In its morphology we have a number of old forms: the Nom. sg. in -e also in stanzas (cf. uvahānasuyam Āyār. 1. 9); forms of the past tenses (atarimsu Sūy. 1. 11. 6; rütthā Sūy. 9. 1. 1; ahesi Āyār. 1. 9. 3. 6); forms of the future (agamissā Sūy. 1. 15. 25); a present part. formed by -mīņa (abhivāyamīne āyār. 1. 9. 1. 8); forms of Gerund with -ntā (hanţā Sūy. 1. 8. 5) and īyāna (ārusiyāna Āyār. 1. 9. 1. 3). Nouns often govern cases (kālamākankhi Sūy. 1. 11. 38) and a few peculiar words are met with (añju 'straight' Sūy. 1. 9. 1; Āyār. 1. 91. 7; acca- body Ayār1. 9. 1. 11; Sūy. 1. 13. 17; adu for' Sūy. 1. 2. 2. 2; nhusā daughterin-law Sūy. 1. 9. 5).
T
The later and more usual form of AMg. is distinguished from JM. by the additional features like the lengthening of final -am before eva (evāmeva

Page #500

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
491
samanāųso Nāyā 1. 9. 53); i representing iti (Goyamā i samane Uvās. 86); forms of prati without the final -i (pāduppanna Suy. 2. 1. 15); Nom. sg. prose (ceie Vivā. 1); Inst. sg. in -sā (kāyasă Dasave. 6. 27); Dat.sg. in -ttae (neraiyattāe Niraya. 63); Loc.sg.in -ṁsi (sayanijjamsi Kapp. 32); Nom. sg. of -t stems in - (bhagavaṁ Kapp. 1; vannavaṁ Utt. 3. 18); forms like Kammunā (Utt. 1. 17), kāladhammunā (Vivā 49); verbs like äitkkhai (Kapp. S. 64): Kuvvanti (Sūy. 1.4. 1. 16); duruhāmi (Uvās. 108); past tense forms in --tthā and -imsu (vippalaithā Nāyā. 1. 1. 159; harimsu Sūy. 1. 14. 3) and a number of peculiar forms of the infinitive (pavvaittae Nāyā. 1. 12. 39) and Gerund (cicca Utt. 7. 28, vijahittu Utt. 8. 2; laddhur Dasave. 8. 29). In its syntax and vocabulary AMg. resembles Pāli and is nearer to it than the later Prākrits. The language of the canonical works like the Paiņnas, however, is hardly different from the older stage of JM. (ii) Jain Māhārāștri "
The various Nijjuttis and narrative works like Paumacariya, Vasudevahindi and others may be taken to represent the archaic form of JM, the language of the non-canonical books of the Svetämbara writers. Here the Nom. sg. always ends in -0. The Dat. sg. has a -ya ending (vahāya VasuHi. 169. 19) and a Inst. sg. is formed with -īņa (naravaiņa PaumCa. (V.) 5. 110). The Loc. sg. of Fem. words also ends in -āyaṁ (disāyaḥ VasuHi. 280. 27); mayā is used along with mae as Inst. sg. I Per. pron.; a form in -as (jāņam VasuHi. 174. 26) is found for 1 P. sg. present, and there are future forms like bhunjiham (VasuHi. 22. 28) and pavissehaṁ (PaumaCa(V) 8. 191). There is a constant confusion between the Inf and Ger. This form of the language also shows a large number of syntactical peculiarities. A number of words and forms of AMg. are also utilized. In its classical form, as represented by Haribhadra's Samarādityakathā, JM. comes nearest to pure Māhārāștri and shows the influence of Sk. in the use of cvi-forms, compounds ending in - bhūya (-bhūta) and the literary words. The late form of this language is best seen in the stories of Devendra in his commentary on Uttarādhyayana and the anthology called Vajjālagga. Here we see a strong influence of the Apabhraṁsa dialects making itself felt both in its phonology and grammar. This is but inevitable when we find writers like Somaprabha composing works in all the three languages, Sk. JM. Apa. without any plan and mixing them in all possible ways. In this context, it must be clearly understood that this so called Apa. influence originates from the spoken languages in all probability the mother tongues of the writers) and traces of it can be detected even in the AMg. canon (acchahim Utt. 22. 16) and the older works in JM. But this late form of JM. shows the influence of the literary Apa. with which

Page #501

492
Amrita
the writers were thoroughly acquainted. In its phonology, we find the confusion between written -m- and -va- (samarā Vajag. 62); simplification of the conjuncts (hosai Vajlag. 54); a shortening of the final vowels (ju Vajlag. 234; annaha VajLag. 129); the change of -s- to -h- (dehi VajLag. 296) and - y- in place of -v- (taheya Vajag. 383). In its morphology, we have the forms of the Nom. and Acc. falling together (dala Acc. VajLag. 1_4; āvai Acc Vajlag. 273); pronominal forms like pai (Inst. Vaj.Lag. 190; Loc. Vajlag. 704) and gerunds in -evi or-ivi (langhevi Erz. 78. v. 181). It shows a further contraction of syllables (avantassa Erz. 82. v. 255; atthantassa VajLag. 132) and all traces of the Sk. -aya- formans of the verbs are lost (cintai Erz. 81. v. 240; vāranti Vajlag. 56). (iii) Jain Sauraseni
The name Jain Sauraseni is given to the Prākrit dialect which is seen in the works of the Digambara writers, most of whom come from the south. It is modelled on the name Jain Māhārāstrī which is used to designate the Prākrit found in the non-canonical works of the śvetāmbara writers. The Digambara view is that the original canon, containing the teachings of the last prophet Mahāvīra, was lost along with the language in which it was composed, when the community migrated to the south, sometime before the beginning of the Christian era. Their religious needs were however met by the writings of later scholars like Puspadanta, Bhūtabali and Gunadhara in the form of Sūtras called Satkhandāgama and Kasāyapāhuda. Later on these works were commented upon by scholars like Virasena and Jinasena, in extensive works called Dhavalā, Mahādhavalā and Jayadhavalā, mostly written in Prākrit but with Passages in Sanskrit as well. In the meantime the religious and the philosophical teachings were written down in manuals in verse' by scholars like Kundakunda, Vattakera, Yativrsabha and others upto the 12th century. All these writings are included in the scope of Jain Saurasenī.
A
The language found in all these books is fairly uniform but differs considerably from the Jain Māhārāstrī of the writers from the north. It changes the Sanskrit sounds ta and tha into da and dha and preserves the original sounds da and dha. It changes the dental nasal into na and knows only one sibilant sa. Sometimes jña is changed into nha. Peculiar grammatical forms are a Loc. sg. ending in -mhi and the adverbial ablative ending -do. The verbal terminations like ti and -te are changed to -di and-de. Passive is formed with -iya and the gerund ends in -dūņa. The name given to this - dialect is intended to suggest that it is a peculiar form of Saurasenī used by the Jain writers of the south. The peculiarities of this Prākrit need further confirmation by the editions of these works in a more critical form.

Page #502

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
493
(iv-vi) Māhārāștri and the Dramatic Prākrits
The nature of Māhārāstrī is described by the Prākrit grammarians like Vararuci, and others following him in detail because they considered it as representative of all the Prākrit dialects being the most comprehensive. From the date of Dandin onwards it is named Māhārāstrī and is thought to be the best Prākrit in which famous epics were composed. But its relation to Dāksinātyā of Bharata, as used in the dramas, and to Sauraseni in particular, which is the most extensively used Prakrit in the dramas, remains doubtful and the problem needs a historical review.
During the beginning of the 20th century when Präkrit studies were in their infancy and were primarily related to the dramatic Prākrits, a controversy, arose about the distribution of dialects among the dramatic personages and the distinction between the various Prākrits as described by the Prākrit grammarians. It was natural and inevitable to proceed with the information supplied by the writers on dramaturgy and poetics as regards the use of the dialects by various characters in the drama and to rely on the Prākrit grammarians for the distinguishing characteristics of the dialects usually enumerated in this context. The result was a kind of disagreement between the two views leading to the problem of deciding the Main Prākrit of the dramas. However, it must be noted that neither the Prākrit grammarians were of one opinion about the features of a given dialect, nor did the works on dramaturgy show complete unanimity in the use of a language by a particular type of character in the dramas. The result was a kind of historical reconstruction of the growth in the dramatic practice, which · distinguished various stages : (i) a pre-classical Prākrit called old Saurasenī, as the dominant language in the earliest stage of the Sanskrit drama, as seen in the fragments of Buddhist dramas discovered in Central Asia, edited and analysed by lüders (ii) a slightly later stage as seen in the use of Sauraseni in the Dhruvās given by Bharata in the 32nd chapter of his Nātyaśāstra and the absence of Māhārāstrī in the list of the dramatic Prākrits, this language being grammatically analysed by Jacobi and supported to some extent by the recently discovered dramas attributed to Bhāsa (iii) a classical stage represented by the dramas of Kālidās, Harsa, Śūdraka, Viśākhadatta, Bhavabhūti and others upto the dramas of Rājasekhara and the dramatic theory that Sauraseni was the language of the prose spoken by ladies in general and Māhārāstrī to be used in their stanzas, a view held by Pischel, Konow and others. This position was, however, objected to by some scholars and the use of Māhārāstrī was denied for the dramas as was done by Hillebrandt and others, or it was considered to be the same as Sauraseni but

Page #503

494
Amrita
in a more developed form as M. Ghosh and others did, thus casting some doubt on the validity of the views of the dramaturgists and the grammarians. In particular, two problems arose (1) whether the stanzas are to be in Māhārāstri or in Sauraseni and (2) whether the generally accepted characteristic feature of Saurasenī, the voicing of the dental stops is valid or not. This led to a further question, what language is meant by 'Prākrit when it is used by the grammarians as a cover term for the whole group of Middle Indo-Aryan languages with which they deal.
A lexicographer has to decide how to designate the Prākrit passages which are found in the Sanskrit dramas. Hence he has to take some decision about the various Prākrits, dialects and subdialects called Bhāsā and Vibhāsā. A closer examination of the original data is essential to decide the issue. I have a feeling that much of the controversy on this account is based on some assumptions of a linguistic nature which are not correct and some interpretations of the passages on which they are based appear to be erroneous. To the first group belongs the view that distinctions in the literary dialects is mostly based on phonological differences. This may be true where we can actually analyse the spoken languages which are fairly uniform in this respect. But while dealing with ancient and medieval languages which are available only in their written form, this may not do. 'They are not uniform and use material belonging to different stages of development and hence are misleading. Nor are the sounds uniformly used in all the words in which they are expected to occur. A striking example is supplied by the opinion of Lüders, who sets up a language called Ardha-Māgadhī for the dramatic fragments found in central Asia on the evidence of a single form of future tense kāhāma for this purpose. His other assumption, that the three forms of Prākrit revealed in these fragments are representative of their older stage on the ground that the intervocalic stops are preserved and are not lost, is also doubtful. Phonetic changes in the various languages and dialects do not proceed with the same speed and older and younger phases can be found at any given stage of the language. That intervocalic stops in Latin are either completely lost or voiced or kept in tact in the different Romance languages does not give us any right to say that one is older and the other is younger in their written form and the same is true in case of the Prākrit languages Lüders himself has argued that the original language of the Pāli canon voiced the intervocalic stops, where the later Pāli uses the unvoiced consonants. This has an important bearing on the nature of the language of the canon of the Svetāmbara community, usually called Ardha-Māgadhi but the older name of which was Addhamāgahā Bhāsā. The MS. material of the canonical works, both older and younger, is such as to allow side by side the use of the

Page #504

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
495
intervocalic stops and their loss even in the same sentence or verse. This has led to a regular duplication of all words containing such sounds which cannot be true of any language whatsoever. For example the 3rd person singular active form of a verb, which is very frequent cannot be having both -t+i and -i as its ending throughout the language in all its occurrences. And to make the matter worse, textual evidence, whether the MS. is on paper or palmleaves, whether old or young, is equally uncertain. There appears to be a feeling among the editors of the Āgamic texts that giving preference to the practice of retaining the consonants would make the language look older and nearer Pāli which is considered as an older speech. This does not look natural and hence cannot be accepted. To ascertain the relative ages of closely related languages, the only valid criteria are of a grammatical nature. The synchronic nature of the Asokan inscriptions and the phonological differences found in them are responsible for extending their use to the literary languages as well. Besides morphological criteria one can also think of a semantic criterion as well for this purpose. For example, the meaning of the verb atipāt- (causal) and the noun atipata shows the meaning 'to injure, to kill only in Pāli and AMg. A reference to the article atipāt- in the second volume of the Sanskrit Dictionary of the Deccan College will show that it is rarely used there in this sense and the only examples for it are from the works of Āyurveda, Suśruta and Caraka, the language of which shows a close affinity with Buddhist usage.
Returning to the dramatic Prākrits, we should note that a number of misinterpretations and conclusions drawn from insufficient evidence have led to the present position, which, to say the least, is confusing. For example it is often stated that the 12th chapter of Vararuci's Prākstaprakāśa is an interpolation on the ground that there is no commentary of Bhāmaha now available. There is, however, good internal evidence to show that this chapter forms an integral part of the original work. There is no doubt that the first 9 chapters deal with some Prākrit, which is not named at the beginning. In fact the very first Sūtra runs as 3 : which means of the vowel which is initial and constitutes an adhikara for the following Sūtras. This is followed by the next Sūtra 371 Hzcy which shows that the initial vowel of words like
to and others is changed to a giving rise to the form i s in this language. But then what is the name of this language ? The author says nothing. It may look reasonable to think that it was called 'Prākrit' because the work bears the name Prākrtaprakāśa. But then all the other languages
It with in chapters 10, 11 and 12 are also included in the scope of the title. We may further ask, does the author use the word Prākrit in two senses, Prakrit as a cover term for all the dialects dealt with and Prākrit as a specific

Page #505

496
Amrita
dialect treated in the first 9 chapters? In this first alternative, what is given then in these chapters is not the discription of any one dialect, but a common core of all the dialects taken together. But then, following the usual practice only the deviations from this common core should have been given for the other dialects which follow. But this is not the case. There are two Sūtras each, at the beginning of the last three chapters, the first giving the name of the language and the second its basis. The Sūtras run as follows पैशाची, प्रकृतिः
A, AMET, ufa: ist, sich yofa: 14. This means the procedure followed in describing these languages is first to describe Sauraseni and then mark the deviations from it in case of Paiśācī and Māgadhi. But the description is hardly sufficient for this purpose unless we include all the matter of the first nine chapters in it. Thus the full description of Sauraseni would consist of two components, one, the special rules laid down in the 12th chapter and the other, all the rules of Prākrit as found in chapters 19. This part of its grammar is indicated by the last rule of the 12th chapter which runs to Heretag. This means that the Prākrit language described in the first 9 chapters under the general name Prākrit is here specified as having the name of Māhārāstrī. Has not Dandin specifically said HERICISI 410 uri ugodi fars: ? Thus the traditional explanation that the name of the language described in the first 9 chapters is Māhārāstrī is correct.
The 3rd Sūtra of the 12th chapter lays down the rule that non-initial, intervocalic ta and tha are changed into da and dha. Here the basic sounds are to be taken from Sanskrit and not from Prākrit because they do not survive there. Hence the earlier rule ufa: Hot4. This implies that these two sounds did not exist in the Prākrit of the first 9 chapters which thus differs from sauraseni in this diagnostic feature, which fact is true of Māhārāstrī. Sūtra 2. 7 states Halley at a: as an exception to the general rule 2. 2, by which both 7 and were dropped, but survived in a group of words as exceptions. For this gana a further change is indicated by 1. 29 3C rally by which r is changed to u which is found in the final form 35. Thus the voicing of dental stops was a regular feature of Saurasenī while it occurred in a few words in Māhārāstrī as exceptions. This complex procedure is chosen because no single Prākrit dialect can be taken as a basis for others.
Once this position is realised the other problem about the statement of Hemacandra 1. 209 375 faç #alfcg & sfa 31rC : I R retary ta erud sfa tand gets clarified. This passage has troubled all those who have worked on it, beginning with Th. Bloch, Pischel, Niti-Dolci, Alsdorf and others. The question is raised, whom does Hemacandra refer to by m al and why did he introduce this discussion in the commentary of the Sūtra which

Page #506

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
497
lays down the change of fac to 500. The scholars have blamed Hemacandra quite unnecessarily by saying that he has copied this passage from somewhere and without understading it. More specifically it is argued that the words terally are a wrong copying of some such expression as frally for standing for a reduplicated substitute in the system of grammar followed by Trivikrama and others.
A little more attention to the plan of Hemacandra's Prākrit grammar can clear up this point. This passage occurs in the section dealing with the Main Prākrit, which fact is of some importance. The system followed by Hemacandra in stating the changes of non-initial intervocalic consonants is as follows. After stating the general rule of dropping such consonants 5-11-2-3
-2-4-4 and a (1. 177). Hemacandra notes the sporadic changes of the Sanskrit consonants in their traditional order beginning with onwards. When these substitutes are more than one for a given consonant the Sūtras are arranged in that group according to the traditional order of the substitutes. To take an example the consonant go, its change to T is given in 1. 181, to 7 in 182, to a in 1. 183, to 4 in 1. 184, to # in 1. 185 and to & in 1. 186. Here the substitutes follow the traditional order. When it comes to 7, besides its general loss the other changes are given as follows. It changes to 7 and 59 (1. 204), to Z (1. 205),.to 3 (1. 206-7) to U (1. 208209) and to (1. 210), to 07 (1. 221), to a (1. 213) and to (1. 214). Here the substitutes are also in the traditional order. After stating the change of a to Uo (1. 209) and before its change to is given (1. 210) Hemacandra naturally finds that a possible change of 7 to a has to be dealt with because such a change has been admitted by someone as occurring in some sporadic words. Hence 37 afect does not mean in the topic of the Sūtra 1. 209, but in the alphabetical order of the substitutes i. e. between and which is the place for . In fact the sound which undergoes change is here contextually fixed as a (cf. 1. 204 a: = e) and the Sūtra is shortened to the essential parts Traifay :. If we add the word 7: by the process of 377fa, the Sūtra becomes #rallagata which is actually the Sūtra of Vararuci 2. 7 in his Prākrtaprakāśa. It is immaterial whether Hemacandra should have given the name of Vararuci or refer to any other grammarian which uses such a rule in the description of the basic Prākrit or uses the word to refer to a whole school of Eastern grammarians. It only means that his view is different from them, and he does not admit such a change in the principal Prākrit. It helps him then, to follow a uniform procedure of basing the rules of change on each preceding dialect, which he does by changing the order of the dialects to शौरसेनी [शेषं प्राकृतवत्]. मागधी [शेषं शौरसेनीवत्], पैशाची [शेषं शौरसेनीवत्], चूलिकापैशाची retu yra), 37942T ( id iacu. He also takes note of the words and forms

Page #507

498
Amrita
which Prākrits have taken from Sanskrit, mainly words, by his last Sūtra 4. 446 T Hitac fHSH. Instead of blaming him on the count of copying without understanding, we must admire his skill in following a far more systematic and rigorous procedure in this case.
With these interpretations and clarifications, we should be able to state the exact relationship between the two Prākrit dialects Māhārāstrī and Saurasenī. It is not a question of one language developing into another, or that one is an earlier form of the other. They are two distinct dialects which were used for literary purposes at different stages of their development. In case of the dentals, their voicing was a regular feature of Sauresenī, while such a change is confined to a few words in Māhārāstrī, which can be considered as either archaisms or borrowings. There is however one more difficulty in accepting such a position, which must be cleared up. The Prākstaprakāśa of Vararuci appears to show a change of this type even in the morphological form which cannot be thus limited to a few words, or thought of as an isolated phenomenon. Vararuci 5. 6 reads SHRUGIGE4:, which Bhāmaha explains as stating that the Abl. sg. of a-ending nouns in this Prākrit is formed by adding the terminations 341, ai, į or fe. The ending 371 is nothing but the base itself with the loss of the final consonant 7 in the original Skt. form. But both the endings and with a voiced dental, which though regular in Saurasenī, are unusual in Māhārāstrī. Bhämaha's examples are GI, Torg but no such forms are recorded in this Prākrit. No other grammarian mentions them. Nor can we regard them as archaisms. It appears that these forms are due to some misunderstanding of the wording of this Sūtra, either on the part of the earliest editor Cowell or possibly on the part of Bhāmaha himself. Though Vararuci does not specifically mention the mechanism of stating grammatical rules, he generally follows the method followed by the Sanskrit grammarians which was accepted as a regular technique of the science of grammar. Whenever a single vowel is to be stated a 7 is added to it, to fix its quantity and to make the use of its cases easy. Thus 7: stands for the gen. sg. of 7 which is equal to its explanation RT. In a series of terminations which begin with vowels this becomes essential. In the above Sūtra the suffixes enumerated are 371, 371, 3 and fe and to combine them, a
is added to the first two to give 34776 ATCL, and following the usual Sandhi rules, the reads 3116ige: in the plural. Failing to understand this, some one wrongly analysed it into 37 g 64: and to justify it gave the examples in conformity with it as वच्छादो, वच्छादु. This can hardly be the mistake of the commentator, when we find that E 5. 22 and many others are correctly analysed by him. Once Cowell included this in his edition, most of the later scholars have continued to repeat it to the present day. Let me add here that

Page #508

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
499
•
The
this problem is simplified for the sake of clarity.
The idea that all the stanzas spoken by those characters in the Sanskrit drama who use Sauraseni in their speech, are to be taken as composed in Māhārāstrī, is another case of misunderstanding. Bharata's statement that seven dialects were used in the dramas and in addition seven more subdialects were also used is substantially correct. This list does not include Māhārāstrī, but his treatment of Prākrit in general is nothing but a brief summary of what the later grammarians give. The 17th chapter of the Nātyaśāstra is repeatedly edited by scholars like Dr. Ghosh and Dr. Vaidya. In this sketch there is nothing which is new and all that is found in it is also included in the first 9 chapters of Vararuci, and the first 3 Pädas of Hemacandra, and therefore they naturally deal with the same language. Whether it should be called by the generic term Prākrit or by the specific designation Māhārāstrī is really of no importance, and in the editions of the Prākrtaprakāśa, with the commentaries in South India, this situation persists. More particularly Bharata's statement in V. 13 372489 Gal: would suggest that the sound occurring in the Main Prākrit was a slightly different sound than the same phoneme in Saurasenī and Sanskrit. Phonetically it may be slightly fricativised and thus' would approach a sound like a stage which immediately preceeded its complete loss.
In this context it is necessary to refer to the traditional allotment of the various Prākrit dialects and sub-dialects to the different dramatic characters. Bharata gives a detailed statement about it in his chapter 17, stanzas 32-40, and the Daśarūpa and the Natyadarpana merely summarise them, as does also Viśvanātha in his Sāhityadarpana. He, however, makes a slight addition to it with the words आसामेव तु गाथासु महाराष्ट्री (sic.) प्रयोजयेत् to which nothing corresponds in the other works. Both Pischel and Konow have given more importance to this statement by taking the word gāthāsu in a wider sense so as to include all kinds of stanzas in Prākrit and making a sharp division between the prose in Saurasenī and verses in Māhārāstrī. In case of all other dialects like Māgadhī, sākārī, Chāņdālī etc. no such distinction is found, and the language remains the same whether in prose or verse. The oldest commentary on the Sāhityadarpana written by 3TIGA, the son of Viśvanātha, explains er here to mean farg which meaning the word Gātha has in the early literature. It is only later that Gāthā comes to mean a specific Prākrit metre. We can then understand Viśvanātha to mean that only the songs, which occur in the speech of the ladies who speak Saurasenī, should be composed in Māhārāstrī. The practice of early Sanskrit dramas shows that a song to be sung by a female character was used at the

Page #509

500
Amrita
appropriate occasion and Māhārāstrī was the most suitable Prākrit for it, because it was used for this purpose and owes its extreme attenuation to it.
The subordinate role which Māhārāstrī plays in Sanskrit drama can also be seen by ascertaining the amount of Prākrit stanzas in them. A quick check reveals that the number of stanzas in Prākrit in the early dramas was small. The three dramas of Bhasa प्रतिज्ञायौगन्धरायण, चारुदत्त and बालचरित show a total of 12 Prākrit stanzas and not a single one out of them is put in the mouth of a woman. In the dramas of Kālidāsa, pligorie has 8 stanzas out of which 7 are spoken by women but only one of them, put in the mouth of the tê in the prologue, can be called a filfa. In factors there is a single stanza uttered by the heroine herself and in fac ity, out of a total of 31, only 3 are put in the mouth of the ladies. Of the three dramas of Harsa, Hat has 8 out of which 4 are uttered by ladies and only the three faucieuss sung by the all, can be called Gīti. In MER, 2 are uttered by 37104 Ch, the heroine in disguise, only one of which was sung by her. The dramas of Bhavabhuti show no Prākrit stanza and on the contrary Hicidit sings two stanzas in Sanskrit. Both the Prākrit stanzas in a er are uttered by the राक्षस and राक्षसी. It will be seen that we have to make a distinction between Prākrit stanzas sung in the dramas and others used for other purposes. The songs are very few and can be considered as written in Māhārāstrī while all other stanzas will be in the language used by the dramatic characters, viz. Saurasenī, Māgadhi, Sākārī etc.
As a practical device which will help to avoid the pitfalls in this regard, the quotations taken from the dramas are always supplied with the name or status of the speaker. The literature available in other dialects like Paiśācī and Cūlikā Paiśācī or the numerous subdialects like Prācyā, sābarī, Śākārī etc. is so scanty that no problem should arise. The available information in their case will be found under words peculiar to them according to the statement of the Prākrit Grammarians. (vii) Apabhramsa
From 1918 onwards a large number of works in Apabhramśa have come to light due to the labours of many scholars and to-day we posses a sizable literature in it, mostly in the form of the lives of the Jain prophets or the Jain versions of the two epics, the Rāmāyana and the Mahābhārata, which form a part of the lives of the prophets Munisuvrata and Aristanemi, Some books preach Jain religion and others are didactic in nature. In comparison, the Apa. works written by non Jain writers are scanty. We possess the doubtful verses occurring in the fourth Act of Vikramorvašīya and

Page #510

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
501
a few stanzas in the works of Abhinavgupta dealing with Tantra philosophy. A few stanzas are also quoted in the works on Alankāraśāstra.
The most striking feature of the Apa. literature is the fact that the whole of it is written in verse and there hardly exists any prose in it, except a passage in the book called Kuvalayamālā, written by Uddyotanasūri. But the most valuable material of this language is found in a large number of stanzas quoted by Hemacandra in the fourth Päda of his Prākrit grammar. They show not only traces of their non-Jain origins but also contain purely secular matter and show traces of different dialects.
The traditional distribution of the Apa. language into the three dialects called Nāgara, Up nāgara and Vrācada is of no great importance. The retention of ļ and of ș in clusters which is given as a feature of the Vrācada dialect is actually found in the works like the Mahāpurāna of Puspadanta written in the south, while this dialect is assigned to the north. The only region where Apa. works were written without break is Gujarat and Rajasthan i. e. Western India. Here it is not possible to separate the Apa. works from those written in early Gujarāti and Western Rajasthānī. For this purpose we will have to use a metrical and a structural criterion by which works called Sandhis are assigned to Apa. while those which go by the name of Rāsu or Phāgu will belong to the earliest NIA. languages. A similar distinction is drawn between Apabhramsa and Avahatthā works.
The Apa. grammar is a greatly simplified form of the Prākrit. The final vowels are all shortened except in monosyllabic words. Both the declensions
and conjugations are greatly reduced. The multiplicity of form in many · grammatical categories is indicative of dialectal origins but they cannot be
now sorted out on this basis. The available literature is of a purely literary nature and hence we have treated all the Apa. works as forming one language for the purpose of this dictionary, which will give only the basic word-form. The language used by the Digambara writers appears to be fairly uniform and constitutes the bulk of Apa. literature. The few works of the Śvetāmbara poets, written in Western India, do show a few peculiar features in grammar but on that account their language need not be set up as a different dialect for the purpose of the lexicon. The use of Deśī words and roots shows a perceptible increase in this literaturę.
For this dictionary, two works Prākrtapaingala and Sandeśarāsaka, have been included in the Apabhramśa group, though linguistically speaking they belong to the Avahatthi stage.

Page #511

502
Amrita
IV. ORTHOGRAPHIC CONVENTIONS AND THE PHONETIC FORMS OF THE PRĀKRIT WORDS
The books in Prākrit which are used for this dictionary show a number of writing conventions partly based on the manuscript tradition and partly on the views which the editors held about them. Others are based on the teaching of the Prākrit grammarians who teach phonetic changes which a Sanskrit word undergoes when it passes over to the Prākrits, which view the editors may or may not accept. These differences in the sounds (or letters) used, affect the alphabetical arrangement of the words a good deal and hence the lexicographer has to look into the problems arising out of them, and take suitable decisions to be followed in the arrangement of the dictionary: Some of the more important ones are discussed here and the practice followed in the dictionary is clearly indicated.
(1) The writing of 4 in the place of a vowel left over (udvrtta) with the loss of an intervocalic consonant is briefly called tula. The rules laid down by Hemacandra in this respect are neglected by the writers of the MSS. and also by modern scholars while editing the works in AMg. JM. JS. and Apa. It was only Pischel who tried to follow them in his edition of the Deśīnāmamālā by discarding the MSS. evidence on the ground that the explicit teaching of a grammarian must be given due weight against the chaotic writings of the scribes. Hem. (Gr.) 1. 180 lays down that whenever an intervocalic consonant is dropped, the remaining vowel will be pronounced as a form of 2 which is to be articulated very lightly provided the surviving vowel is 37 or 371 and is preceded by 37 or 377 and not otherwise. TERUGA! Gfc afa ga: 3Tauf: 3Taufa percecrafastafa By the general practice of the grammarians 37 stands for both 37 and 37 unless followed by 'an indicatory a. Among the eastern grammarians, it is only Mārkandeya who refers to his phenomenon (2. 2). This is regarded as a writing convention of the Jain scribes. In the purely non-Jain works like गाहासत्तसई, रावणवह, गउडवह, लीलावई and other later works like कंसवध, उषानिरुद्ध and some of the सट्टकs, no यश्रुति is used. In the case of 13506 when all the MSS. come from Jain Bhandaras and use I throughout the new edition of Prof. Suru drops this 2 in all cases while others like Prof. Patwardhan and Dr. Upadhye continue to use it even in nonJain works when the MSS. show it. The same holds good of a large number of Sanskrit dramas written by Jain and non-Jain writers. This practice is thus taken to be a writing method and is treated as such and no phonetic value is attached to it. Hence also the complete neglect of the restrictions imposed on its occurrence by Hemacandra.
The nature of this -y-, like a similar use of -v-, is given by tradition

Page #512

A. Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
503
S05
as having a very light articulation, as against the indistinct pronunciation of & referred to by Bharata in 17. 13. Speaking phonetically we may describe the first as a glide and the second as a slightly fricativised voiced stop in the intervocalic position. This will help us in understanding Hemacandra's rule precisely and ascertain its validity. Here a glide of a palatal colour is involved as occurring between two vowels which is due to the loss of single consonants in that position. Hemacandra's condition is two-fold, reference being made to the immediately preceding vowel and the vowel which follows it. The situation thus created consists of two vowels coming together but without forming crasis, and forming two distinct syllables. This will inevitably give rise to a glide sound between the two and yet will not have a full consonantal value. The possible sequences of this nature will involve the use of a front, a back or a central vowel at both the places. The i- like glide will thus assume three phonetic values one which is the same as a clearly articulated y- sound, the other between two front vowels becoming hardly audible and a third, a fleeting y- like sound lightly pronounced. It is only when the preceding and the following vowels are central ones, do we get a y- glide to mark the syllabic boundary and Hemacandra's rule correctly describes it as writ:. The other procedure to keep the syllabic boundary will be to use a glottal stop which Sanskrit does not recognise as a sound of the language. That a similar situation does not arise in case of vis due to the fact that its pronunciation has already shifted to a fricative sound. That the scribes did not take note of such a fine distinction in case of the written records is quite understandable and they used the same symbol for both. Thus a form like afa or 44 as found in a work like a cafect should not be judged as an archaic form but merely as a writing convention for 4 or .
Another writing convention is found in the MSS. of both the canonical works in Ardha-Māgadhi and the post-canonical works in Jain Māhārāstri. It consists in writing the lettera in place of many originally different consonants, and is called asia on the analogy of Zafa and a fa, but does not belong to that category. Most editors refer to it as a peculiarity of the MSS. they use, and being convinced that it has no phonetic validity, they do not take the trouble of recording it fully. This lack of evidence in the recorded readings makes it difficult to judge it correctly. This writing of a is fairly wellrepresented in the edition of a cies in both its Khandas and most Cūrnis on the canonical works use it extensively: Schubring has very carefully noted the writing of this in his edition of the work sfa i s and allows us to judge its nature to some extent. An analysis of its usage gives us the following result. Compared to Sanskrit this written a corresponds to different original

Page #513

504
Amrita
consonants like क as in कुप्पासत= कूसिक, चक्कवात= चक्रवाक, सावत= श्रावक, कालियकक्कलित कालिकोत्कालिक, in place of ग as in पओतप्रयोग, in place of च as in विमोतितो विमोचितः, वातिकवाचिक, अणुवीति= अनुवीचि, सूति सूचि, कदाति=कदाचित्, मोतेति= मोचयति; in place of ज' as in पारिताय= पारिजात, गत=गज, भारदात=भारद्धाज; in place
of द as in नारत= नारद, नती= नदी, जति= यदि, जणवत= जनपद. भेदाति= भेदादि. , अकोविता=अकोविदा. कोहातीणं क्रोधातीणं क्रोधादीनां. मेतिणी=मेदिनी; in place of य as in
आयरित आचाय, छाता= छाया; जिणाततण=जिनायतन, माति=मायिन, ताती-त्रायिन etc. Such changes are phonetically impossible and cannot be admitted in any language. In most of these cases, a corresponding form with a written 4 is also found, sometimes side by side, which makes it obvious that the scribe wrote it randomly and attached no importance to it. अकामए चरए तवं is followed by अकामते चरते तवं, संसारकंताराओ तातीति ताई, णेरइया णेरतिता and many similar cases are found in the work 4TS. This usage is so inconsistent and occurs so. unexpectedly that most of them must be cases due to negligence of the scribes.
To judge this mode of writing one must ascertain its source. No attempt is made so far in this direction except the one made by Dr. Upadhye, but in a halting manner. He suggests that this writing originated in the copying of the Chedasūtras and it spread to other works from it. It was done to keep these writings concealed from the common reader or make them less accessible, being meant for the initiated few. Such attempts are not rare in Indian literature and one that is often met with in the Tantra works is based on rearranging the letters in a complicated manner, so that one has to know this scheme before one is able to read them correctly. But these considerations do not apply to the Chedasūtras of the Jain Canon. The writing is not confined to any particular type of works but is found used in all types both old and new. Nor is its usage confined to words which need concealment but is found in all kinds of words and these words are written both with and without a and are not of the nature which are to be guarded from the common reader. Pt. Malvania's reference to the भाष्य stanza 6227 of the निशीथसूत्र and the commentary in the 3rd part of व्यवहारसूत्र do not prove that they are to be excluded from the common man. The चूणि on the निशीथभाष्य makes it clear that what is to be kept back are mainly the 37491846 and only incidentally the Chedasātra itself. Its wording is परवयणरहस्सं अववादपदं सव्वं वा छेदसतं. The व्यवहारभाष्य confirms the same view in the passage छेदसूत्राणि प्रकल्पव्यवहारादीनि तानि वसतौ अपरिणामको अतिपरिणामको वा शणुयात् तथा विद्यामन्त्रान् च वसतौ कस्यापि दीयमाना(?)न् अविगीता निर्धर्मा शृणुयात्. The same verse occurs in the बृहत्कल्पभाष्य 790. प्रवचनरहस्य is to be taught, taking into consideration, the place, the time, and the person to whom it is to be taught. In the last case three types of students are thought of. They are called परिणामक, अपरिणामक and

Page #514

A Comprehensive and Critical dictionary of Präkrit Languages (Introduction)
505
3fa4f1|14. The second and the third type of students are to be excluded .from the teaching of the Chedasūtra. The reason is the attitude they show towards the rules and their exceptions. The fu1140 admits both the rules and the exceptions as legitimate, the 374f414 admits only the rules, but not the exceptions, while the sufrufo 140 admits only the exceptions and not the rules. Obviously in the field of a penal code the last two will draw wrong conclusions and hence they should be excluded from their study, particularly the atonements prescribed. No question of secrecy is involved.
What then is the origin of this habit of writing a ? It is clearly based on some kind of misunderstanding on the part of the scribes. The occasional lack of attention on the part of the copyist explains their random occurrence. The influence of the Sanskrit language at the basis of the Prākrit words may have affected him to some extent, and is probably the main cause, when the original Sanskrit contains a 7. Its use for other original sounds can only be explained on the basis of a habit to transcribe a Prākrit word by the substitution of its counterpart, a kind of Sanskritisation of the language. Once he develops the habit of changing a word like 44 to Ha, it is but natural that now and then a written 27, even if it comes from some other Sanskrit sound, say in a word like ture standing for me, will be written by him by using 7, i. e. To Such a process can only be unsystematic and sporadic and the agfa is exactly of this nature. Such writing based on a misunderstanding is at work in all the classical language MSS. and some of them may find a place as the correct usage in the language as well. Whether this is a m usic or a regularly pronounced phoneme of Sanskrit is immaterial for such mistakes and both will be treated in the same manner. This position is confirmed by the fact that in case of such sounds as are not normally dropped or written as 9, i. e. 2 and 3 and 9 or a, there will be no occasion for writing them as a, and this agrees with the known facts. The consonants which are written as त are क, ग, च, ज, त, द and य. From this a very important and valid inference can be drawn. The MSS. which were copied by the scribes had in them a written -y- for all the above consonants which thus indirectly proves the rule stated by Hemacandra in 1. 177 stating the loss of intervocalic stops and also proves its antiquity.
In view of the above discussion, it appears quite reasonable to take the following decisions in setting up the lemmata in the dictionary. The ya-śruti will be allowed in the case of the Prākrit dialects AMg., JM., JS. and some of the Apa. works, but without violating the evidence of the MSS. which use side by side the simple vowels as well. In the latter case of cross-reference to the form with the ya-śruti will be useful. If the word is found in the Nom.

Page #515

506
Amrita
sg. ending in -e or -0, the base in these dialects will be given with ya and not -a. The writing of syllables like , I , etc. which sporadically occur will be given as 3, 3, 1, 371 etc. In case of the dialects Māhārāstri and the dramatic Präkrits and some Apa. works, the ya-śruti will be avoided though the citations will show them with the ya-śruti if the edition uses it, particularly in a work like Līlāvaikahā, and stray verses cited in other works. The writing of the agfa will be avoided in setting up the lemmata but used where it occurs in the MSS. as an exclusive form. If both forms occur the form with -ta will be avoided as far as possible. The reader will locate all such words in the place proper for it, as if the writing of the 7 is absent. No form with the agfa will be given as lemmata unless no form without -ta occurs. The asta will be treated as occurring regularly and will be so noted in the lemmata and in the citations.
(2) Another writing convention, which has become very common in Prākrit, pertains to the writing of Anusvāra for all kinds of nasal sounds. While Sanskrit uses it before such sounds as the semivowels, sibilants and the voiced aspirate, Prākrit writing convention is to extend it to all the places, before the stops, and before pause, for all nasalised vowels and all the nasal consonants except before n and m, where it is assimilated to the following nasal consonant. But in all such cases its original phonetic value remains intact. When the earlier Präkrit works used the nasal consonants following the Sanskrit mode of writing and made a distinction between an Anusvāra and a nasalised vowel, no such distinction is made in Prākrit Devanagari orthography. Even at the end of a word or a metrical line, where Sanskrit insists upon the use of -m, the Prākrits will use an Anusvāra. The writing of the Anusvāra will be used in this dictionary whatever the mode of writing in various editions which are found to be inconsistent, and therefore call for some kind of normalisation.
(3) The distinction between an Anusvāra and a nasalised vowel will be made only where the metrical rules do not help us to decide the length of the syllable as either short or long, which otherwise decides whether the sound is a full Anusvāra making the syllable long or only a nasalization which keeps the value of the syllable unchanged. In practice nasalised vowels may be occasionally marked with the sign of nasalization (called Candrabindu) at the end of words in Apabhramsa where the length of the syllable is not determinable.
(4) The use of a dental n and a retroflex n in Prākrit has become a rallying point of different views and different practices. The Prākrit grammarians differ from each other and on some essential points the

Page #516

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
507
interpretations of the rules of grammarians have become a matter of doubt for some scholars. Different normalisations have been suggested and used by scholars and even the antiquity of a work is made to depend on it. Dialectal differences play a prominent part in it. Hence some amount of clarification must be attempted before a scheme is suggested and followed in this dictionary.
. Some amount of basic clarification can be achieved by trying to use the phonemic principle for this purpose. This will help to separate the writing conventions from the phonetic values of the sounds involved. The generally accepted writing mode for a given Prākrit dialect may also be used for this purpose. Pāli, and Paisacī with its sub-variety Cūlikā Paisacī use only 7 to represent this sound and it will be retained here as it is. Like the absence of ya-śruti in non-Jaina works and MSS., 7 is mostly absent and only u and to are found. The Jaina works except those in the dialect called JŚ. mostly give both 7 and u as optional sounds and are so written in the initial position. In the intervocalic position is preferred, and in case of the clusters resulting in a geminated nasal, both 7 and u are used like other consonants. No final nasal sound occurs and the final -r in the original Sanskrit words is respresented by an 3769r. Thus there remain two positions in the phonology of Prākrit words where both the sounds are used in writing. These are the initial and intervocalic positions where both n and n and nn and nn are used with no discernible principles.
In the initial position both n and n are used at random or one or the other is used exclusively in a given Prākrit dialect and hence there results no contrast betwen the two. In the medial position an ungeminated sound is mostly used as U and not 7, giving rise to no contrast as such. But in the intervocalic position the geminated sounds and uut are found, but their distribution in a given dialect does not coincide with the distribution of an initial 7 or u. In this position both 5 and vel contrast with n. Hence one may regard the relation between initial n/n as in complementation with nn and nn in the medial position and consider a medial single 17 as a different phoneme. This view is in agreement with the development of these sounds in the New Indo-Aryan languages. Hence the writing of the initial n and n may be based on the writing conventions of a given dialect, and setting aside the writing of both and a (or vu) as two distinct phones in some dialects like Māgadhi, the use of 7 and vu may be taken as a matter of free variation in the other Prākrit dialects as is reflected in the MSS. tradition, which runs parallel to the use of a 9 and the corresponding pure vowel. The etymologically valid distinction between the presence or absence of a n sound in the

Page #517

508
Amrita
corresponding Sanskrit words may not be allowed to decide the writing system of MIA., when they have completely merged together. Of course if the dialect shows a phonemic difference between the geminated ñña and geminated nna or nna, it will have to be kept intact.
(5) The remaining writing conventions may be merely stated. The writing for which is found in the southern MSS. and editions is uniformly followed. The representation of a double aspirated consonant by using two aspirates together, which is found in some of the earlier editions of Präkrit works will be represented by the more phonetically accurate method of using the unaspirated consonant as the first member of the cluster. The लघुप्रयत्नतर य or या will not be distinguished from the normal य or या, as was done by Pischel in his Prākrit grammar because it is not found necessary. The difference between the short and long 5 and 311 sounds can be ascertained from their environment and hence not always necessary to be indicated. This may be necessary only in the case of the final syllable in Apabhramśa, where the principle of anceps is not valid. The MS. tradition of writing and 3 in place of short T and 311 is often retained without normalisation, because it cannot be precisely decided to what extent it is phonetically different in view of the tradition current in Indo-Aryan languages to consider 5 and 3 as shortened forms of T and 371. Thus in the environment of a cluster which follows, no distinction of a phonemic nature is found between [i] [e] or [e:], and the sounds may be really different in each case. The citations will follow the writing found in the editions used. .
V. TRANSLITERATION
The transliteration used is the usual one for Indo-Aryan , languages. Particularly to be noted are 3 car as , as ś, 7 as ș and fast as ḥ, the nasalisation of vowels shown as on Devanāgarī letters as ~, Accents are not marked. Taking into consideration all the Prākrit dialects together, the alphabetical order will be as follows : 37, 371, 3, 8, 3, 3, #, # ], TT371, 311 - :), a, a, 7, 9, (G), , 39, 37, F1, (37), 2, 3, 3, Z, UI, 7, 4, 2, 87, 7, 9, 9, 7, 97, 9, 7, , 3, 7, 1, 4, . In following this alphabetical order, an attempt is made to keep it as close to Sanskrit as possible, because nearly all words will be followed by their Sanskrit equivalents. For this purpose what is written as an 34aR is treated as a separate phoneme before the semi-vowels, fricatives and the aspirate h, as a para-savarna before the stops, and as a 4 in the final position. The distinction between short and long and 377 and the mark of nasalisation of vowels will be treated as mere diacritical marks and will not affect the alphabetical order.

Page #518

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
509
VI. ABBREVIATIONS
Various types of abbreviations have to be used in a comprehensive dictionary in order to make the work as compact as possible. These abbreviations fall into a number of groups : (i) abbreviations of grammatical terms and terms of general nature used in textual and literary criticism. These have been mostly standardised and they will be easily known. (ii) There are symbols which will be used to make clear some aspects of linguistics which, otherwise, will take a lot of space if explained in so many words. (iii) Then the names of various languages and dialects will have to be indicated by abbreviations and also the names of ancient writers and modern scholars will be abbreviated when reference is made to them. There will be abbreviations of the names of the series of publications and other bibliographical information to be given. Above all, a dictionary which is giving quotations from various books on an extensive scale must make use of abbreviations of the names of books which form the corpus of the dictionary. As these names of books are expected to occur thousands of times, it is essential to make the abbreviations self-explanatory or at least easily recallable, even at the cost of making them a little longer, when once they have been noted.
Abbreviations which occur in books which are to be read continuously and which deal with a specific subject or topic will not be found suitable in a dictionary which is to be constantly referred to and in which the context extends over one small entry and gives no help in guessing the source. The practice followed in the lexicons of the classical languages of Europe viz. to use the name of the author first and then indicate his work, will have to be reversed in the case of the classical languages of India. The primary reference will be to the book and only in case of homophonous titles of books author's name will be given to distinguish between them. In the Indian literature as a whole, the names of books are found to be longer than those of the authors and homophony is prevalent in both the types of names. The method used is to combine the abbreviated name of the book, followed by the abbreviated name of the author in brackets. As the number of books to be referred to goes on increasing, the abbreviations also become longer to avoid overlap.
The self-explanatory nature of the abbreviations of books will depend upon the system used in forming them. The syllabic writing used in Indian literature makes it necessary to abbreviate the names to the first syllable by using the principle of acronomy. It may also require us to split the name into its constituent elements and use this principle for each one of its parts and indicate the division by using the capital letter for the first syllable of each such part. Most of the names will consist of two or three such units and very

Page #519

510
Amrita
rarely four or more will be required. Further abbreviations may be used in the case of words which designate the nature of the book or its category. Such words are usually found at the end, and sometimes at the beginning of the name.
Names of books in Präkrit involve a problem which must be taken note of. It is but natural to expect that the names of Prākrit works should be in Prākrit. But in course of time, with the increased importance of Sanskrit, aided by the tendency to write the commentaries in Sanskrit, the Sanskrit versions of the Prākrit titles come into constant use. The later writers usually used them in preference to the Prākrit names. Hence by the mere name of a work it is not always clear whether it is in Prākrit or in Sanskrit. A number of Prākrit works are known only by their Sanskrit names and these cannot be replaced by newly coined Prakrit names. The name used by a long-standing tradition must be given due importance in this regard.
This dictionary uses Prākrit names of Prākrit works as far as possible. As a rule the names of the books of the Ardha-Māgadhi canon are given in their Prākrit form being well-attested in the canon itself. In the case of later works the Prākrit title is given and if it cannot be easily traced or is not found in the work itself, the Sanskrit name is given with an indication that it is in Prākrit by adding into parenthesis (Pkt.). If the work is in mixed Sanskrit and Prākrit, as in case of the dramas or the proto-canonical works of the Digambara sect, the Sanskrit name is preferred. In the case of the 410 and for the Präkrit name is followed by the Prākrit forms 4 and four and in case of a Sanskrit name you and for for the sake of consistency Considerations like being better known decide the choice of सेतुबंध for रावणवह, 964141414fsate for futufae, and 0-14IMI for it. The names of HEAS are given in Prākrit. The names of Prākrit Grammars are replaced by the names of their authors with the addition of Gr. added in parenthesis. From among the alternative names of a book, as far as possible, the older name is selected as in the case of Viyāhapannatti for Bhagavatīsūtra. The names of works dealing with Prākrit metres are given in Sanskrit as they also deal with Sanskrit metres. With a complete listing of the names of the books with their abbreviations supplied, no difficulty should be met with in this regard. VII. THE PLAN OF THE DICTIONARY
This dictionary includes in its scope a group of languages and makes use of a literature of varying types. It has perforce to use a large number of works which are not critically edited, and those which can be called carefully edited are based on different principles for different editors. The books

Page #520

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
511
themselves often show a shorter and a longer version, particularly in the case of the Ardha-Māgadhi canon. The divisions of a given text are done differently by each editor which makes it well-nigh impossible by the given reference to use the commentary on a given text, when the editors of the two are different. Indices are prepared by the editors in such a way that the original divisions of the book are neglected and references cannot be easily traced. All these considerations have made it necessary to adopt a complex system for the dictionary which needs some detailed explanation to facilitate its use.
A typical entry in the dictionary is made up of the following elements, in a definite order :
1. The lemma (in Devanāgari).
2. Transliteration of the Prākrit word, and its derivation or etymology given in the form of transliterated words and enclosed in parenthesis.
3. Alternative forms of the given word in Devanāgarī in square brackets. The quotations from the Kośas and grammatical works dealing with the word are included here.
4. The Parts of Speech in Italics. 5. The meanings in English.
6. After each meaning there follow the quotations from Präkrit works, arranged in a specific order.
7. Cross references where necessary. • ' 1. Lemma
Besides the regular phonetic changes which the Prākrit grammarians describe as being applicable to the Sanskrit words, they also prescribe some changes as being peculiar and applicable to the Sanskrit base of the words (Hem. (Gr.) 1. 15-22). Thus the consonantal endings of the basic words are either dropped or thematised by adding the vowel a to them and in the case of feminine nouns the suffix is naturally -ā or -1. In the case of Sanskrit nouns ending in vowels like #, some changes occur before the case-endings (Hem. (Gr.) 3. 44 foll.). All such changes are due to the attempt to get the word to conform to the phonological structure of the Prākrit languages. Hence the problem of setting up the lemmata of words in a Prākrit dictionary. The practice of modern scholars as seen in the glossaries prepared by them for the books which they have edited differs from scholar to scholar. Various methods have been used by them, keeping the Sanskrit ending as it is, to produce new

Page #521

512
Amrita
bases with no observable general principle. It seems best to set up the lemma in the Prātipadika form when it conforms to the rules of word-endings current in Prākrit. In case this is not possible, we set the lemma in the form in which it occurs before the inflectional terminations. More specifically, if a noun ends in the suffix -in (both krt and taddhita) it will be given without the final consonant. If the word ends in -an, the final consonant will be dropped, or a vowel like -a or -ā will be added according to the gender of the word. For words originally ending in the base will end in -u or -ara – āra as the case may be. The present active participle will be given as ending in -anta. The distinction between words ending in short vowels like -i and -u and long vowels like -ī and -ū will be preserved even in the case of feminine words. Words ending in -vas will be given as u-ending and those in -tr will be either shown as ending in -u or in -tā (> yā) and the ending -jña will be shown as -nnu or -nnu.
The adjectives will be given in their mas. or neu. forms and their feminine form will be shown as (f. -ā] or [f. -ī] as the case may be. If only the feminine form is available then after the designation of the part of speech as adj. the symbol (f.) will be added.
Pronouns will be given as they actually occur in their Nom. (?) Sg. form, and pronominal adjectives will be treated as ordinary nouns, so also the numerals.
The adverbs will be given in their actually occurring forms and in the case of adverbial use of the cases of some nouns and adjectives, they will be included under the corresponding nouns and adjectives. .
In case of the verbs, the practice of the Prākrit scholars so far has been indecisive and varying. Some use the Sanskrit root itself, while ot give the Prākrit form of the Sanskrit root, either using thematization or some other changes. Some give the 3. p. sg. Present form of the root, even when that form may not occur. Following a suggestion of H. Jacobi and confirmed by Edgerton, the verbs will be given here as ending either in -a or -e or - ā, thus forming three types of conjugation. The base will be decided by cutting off the inflections as given by the Prākrit grammarians. Secondary bases of the verbs will be treated in the same way.
As regards the orthography of the lemma which will be given in Devanāgarī throughout, an Anusvāra will be used for all nasal consonants, which do not occur in the Prākrit phonology. Earlier scholars have used both the para-savarna and Anusvāra indiscriminately.
Verbs will be distinguished from the other words by adding a hyphen

Page #522

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
513
at the end, even in the Devanāgarī form. 2. Transliteration, Derivation and Etymology
Transliteration into Roman is given after the Devanāgarī word into brackets so that those who are unacquainted with the Devanāgarī script can also make use of the dictionary. But in the same brackets are included some other words given only in transliteration. These are intended to explain the derivation and to some extent the etymology of the lemma. In the majority of the cases the etymology of the Prākrit word is to be found in a Sanskrit word. Only in rare cases we will have to search for origins in the pre-classical stages of language or even in the Indo-European stage. A brief indication of this will be given wherever necessary by quoting the relevant cognates. The derivation of the Prākrit words is a task beset with great difficulties. No attempt is made by the Prākrit grammarians to explain the derivation of the nouns, derivative verbs, adjectives, adverbs etc. as they occur in Prākrit, because they have laid down the rules of phonetic changes which relate a Sanskrit word with the word in Prākrit, sound for sound. A large number of words which in Sanskrit consist of two or more morphemes, which are based on the derivative history of the word, are all taken up as one unit and are used in Prākrit with the necessary phonetic changes. The Prākrit grammarians confine themselves only to such suffixes which are not found in Sanskrit. Hence the derivation of a Prākrit word can only be understood by looking into the corresponding Sanskrit word which is amenable to morphemic analysis, and this analysis is taken as the basis of the Prākrit word. To make such an attempt requires a good knowledge of the derivation of the Sanskrit word itself. Hence to make the process of derivation clear the transliteration of the Präkrit word and the transliterated form of the Sanskrit word which immediately follows are divided into their morphemic elements which are indicated by hyphens. This procedure works fairly well when the structure of the two words is nearly the same. But in many cases this is not so, because the way in which Prākrit uses its grammatical elements is not the same as is done by Sanskrit. In such cases two methods are used. One is to reconstruct the Sanskrit shape of the Prākrit morphemes sound by sound, and indicate their differences by using italics for the different elements. This often gives rise to a form which looks like Sanskrit only in its phonology, but not in its morphemic make up. The two languages may use different morphemes, or even if they use the same morphemes, their allomorphs will be differently distributed, and this must be made clear by using italics for either the
Şanskrit or Prākrit allomorphs. As this procedure gives rise to Sanskrit-looking · words which are not correct in Sanskrit, they are also given in their correct

Page #523

514
Amrita
form in transliteration. Hence immediately following the Prākrit used in Devanāgarī, there will be at the most three transliterated words in the brackets. In case the Sanskrit and the Prākrit word is the same in all respects, a single transliteration is given for both the Prākrit and the Sanskrit word. If there are phonetic and morphological differences, there will occur two transliterated words, the second being the Sanskrit form, provided it has the same grammatical and morphological structure. If it differs from the first, a third transliterated word will be added which will be a correct Sanskrit form of the Prākrit word as far as the meaning is concerned. This will also be the case if the meaning of the Prākrit word differs from that of Sanskrit, which will also indicate the traditional rendering of the Prākrit word into Sanskrit by the commentators. This arrangement will correspond to the classification of Prākrit words into tatsama and tadbhava types. The third type called Deśī or Desya will have a different treatment as far as its etymology is concerned and,' if known, it will be stated clearly in so many words. 3. Sound Variations
The close relationship between the various Prākrit languages was well established from the very beginning of the tradition of the Prākrit grammarians. This is reflected in the fact that they take Sanskrit as the basis for their description of the Prākrits. In addition they take one Prākrit as a basis for another Prākrit as well. By common consent Māhārāstrī is taken as the basis for a detailed description and this is taken as a new basis for the description of the other Prakrits like Saurasenī, Māgadhī, Paisacī etc. In addition, Māgadhi is described with reference to Sauraseni by noting the differing features of that language. This procedure is quite sufficient to conclude that it is a mere procedural relationship and is in no way connected with the independent nature of the Prākrit languages. That is why Rājasekhara uses the word sadbhāṣākavi to describe himself and Laksmīdhara gives the name sadbhāṣā-candrikā to his Prākrit grammar. Earlier, Rudrata enumerates the six languages as Prākrit, Sanskrit, Māgadhī, Paisacī, Sauraseni and Apabhraṁsa (2. 12). Later writers dropped Sanskrit and added CūlikāPaisacī as an additional Prākrit. Still later the number of Prākrit languages or dialects went on increasing with the result that the features by which one Prākrit was differentiated from the other became fewer and fewer and of less significance.
Even with the major Prākrit languages, in which a fair amount of literature is available, the form of a word shows a varying phonetic shape, which will make it necessary for the dictionary to give each form a separate place in the alphabetical arrangement. These differences, being phonetic in

Page #524

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
515
hiting in and yā, wrinkosen as t
nature, affect a large number of words. In all cases of this type the meaning and the usage remain the same. To give recognition to all such variants as independent entries will lead to endless repetition without any advantage. To add to this undesirable situation, we find that a word assumes different shapes in the same language and thus increases the number of entries. Most of such changes are optional and both alternatives occur in literature. The failure on the part of the modern editors to distinguish between real phonological difference and orthographical variation has led to a situation which is intolerable and the lexicographer has to make a determined effort to reduce the number of entries so produced. This situation is similar to one, caused by the spelling variations, found in Old and Middle English. The treatment which is given to these spelling variations in the lexicons of these languages is adopted for the Prākrit words as well. In this dictionary, all such variations which occur in the same language or even a number of languages are given as variant forms at one place under the word chosen as the lemma. Thus writing of vowels like a and à as ya and yā, writing a ta in place of the consonants which are lost, writing i and u for the short values of e and o, and many others will be treated as spelling variants of a given vocable. This will help reduce the entries to a considerable extent and help the lexicographer to quote passages at one place even when they contain the given word in different phonetic shapes. 4. Parts of Speech
The grammatical status of the words in the dictionary will be indicated by using the theory of the parts of speech. The system of Sanskrit which classifies all the words (Padas) into two classes, nouns (called subanta) and verbs (called tiñanta) and includes the other words among nouns by the device of adding terminations and then dropping them, is not much useful and cannot be followed for the Prākrit languages. Nor is it economical because the classification of the subantas will have to be divided further into many subdivisions to explain their meaning and use in the language. The European grammatical tradition began with Aristotle by setting up three classes called nouns, verbs, and a third heterogenous group called syndesmoi (= connectors) which included the article, pronouns, conjunctions and possibly prepositions as well. Thrax codified for. Greek the eight parts of speech as nouns, verbs, participles, article, pronouns, prepositions, adverbs and conjunctions. As Latin lacked an article. Priscian dropped it in his classification and added interjections in its place. These eight parts of speech thus became current upto the modern times and were used for all types of languages. In spite of the fact that this theory is based on both formal and

Page #525

516
Amrita
semantic criteria, its usefulness cannot be denied and it can be used for the Prākrits with some modifications.
For this dictionary the following parts of speech will be used : 1. Nouns along with the indication of their gender as m. f. n.
2. Adjectives (adj.) also including all types of participles which will be indicated by using their abbreviations in parenthesis : adj (ppp.) etc. When the adjective is available only in its feminine form, it will be indicated by writing (f.) after it. If all forms are found, the feminine form will be shown as ending in [f. ā or -1].
3. Adverbs (adv.). This category will include Acc. Sg. of adjectives used as adverbs and Inst., Abl. and Loc. cases of nouns similarly used. In addition it will include words used as adverbs like ajja, ahunā, divi etc. and nouns with the adverbial suffix Skt. -tas (Prākrit -o).
4. Pronouns (pro.) including personal, demonstrative, reflexive and other pronominal derivatives.
5. Numerals (nu.) include words used in the sense of a number (saṁkhyā) or numeral qualifying other objects (saṁkh
6. Verbs (v.). Here all stems which are conjugated in Prākrit will be included even when they are having different conjugational signs (vikaranas) or use derivative suffixes for passive and causative and even noun-bases (denominatives). These will be classified into three categories, those ending in a, those which usee (from Skt. aya) and those ending in a long vowel. Passive bases except those formed with the regular passive suffix -ijja or -ta will be given as bases with their passive or active meanings. A comprehensive statement or the formation of the verbal bases in the Prākrits can only be made when a detailed grammar of these languages is made available..
7-8. Two adverbial derivatives from the verbal bases are given an independent status here, the gerund (Ger.) and the infinitive (Inf.)
9. Under the designations Adnominal (adn.) (coined after adverbial) are included particles like a, a, a, 7, 7, 10, af, etc. postpositions governing various cases of nouns and adverbs of time, place, cause etc., if they do not come under the scope of No. 3. The criterion used for this purpose is the primary relationship of these with another nouns and not with the verb. When these words are used in groups or in correlative pairs they function as the traditional class of conjuction. This term conjunction is avoided because disjunctive relations between sentences and words are not

Page #526

A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
517
covered by it.
10. Interjections (intj.) which semantically are expressive of emotions and feelings and syntactically form independent sentences like the Vocative Case. 5. Meaning-analysis, Homophony and Polysemy
Meaning analysis is the very core of a dictionary. But no general principles can be laid down for it, because it differs from language to language and even in the same language from word to word. Both historical
gical considerations play an important role in the relations which subsist between different meanings of a word. Similarly no hard and fast line can be drawn between what may be called the overall meaning and a specific meaning. In the field of semantics, each word has its own history and it must be explained on the basis of its use in a given language. Semantic universals are hard to find and hence can be of limited use.
The apparent multiplicity of homonyms in Prākrit need not be indicated by the use of index numbers because in the majority of cases its source in Sanskrit, which is given throughout, makes it obvious. Only in rare cases where the Sanskrit equivalent itself is homophonous, can we use index numbers for this purpose. . The inherent ability of a language to make distinctions in meaning to any degree of nicety makes it necessary to analyse it on the evidence of the actual use of the word in literature. This is true especially in the case of classical languages where an appeal to the intuitive judgement of the speaker cannot be made to decide its precise meaning. In this connection, two things must be constantly and clearly kept in mind. The semantic spectrum of the target language and that of the language of explanation may not be the same, and in fact is usually different. Secondly, only such distinctions in the meaning can be made which are caused by the influence of the linguistic environment only.
In this dictionary the different meanings of a word are indicated by the use of numbers in serial order and no complicated system is used to show close and distant relationships of the meanings. When one meaning directly leads to another this is indicated by the use of the word hence. In the case of technical terms of different branches of knowledge, great care is taken to explain them fully, but no attempt is made to coin new words for this purpose.
In this part of the entry syntactic usage is given full representation and

Page #527

518
Amrita
the linguistic environment is stated in the form of classifications based on common sense, with no pretence of making them scientifically or philosophically very accurate. The labelling of meanings as literal, metonymical or figurative, which is generally followed in the dictionaries of the classical languages of Europe, is not followed here. But all the necessary evidence in the form of quotations is given in ample measure, and the reader is free to draw his own conclusions based on them. So also the classification of meanings as expressive, indicative and suggestive which is done by the writers of rhetorics in Sanskrit is not explicitly stated. Stylistic differences and usage labels are mostly confined to the passages taken from the dramatic works, whether in Sanskrit and Prākrit or wholly in Prākrit (the Sattakas). Here the name of the speaker or an indication of his status is always given at the beginning of each quotation from the dramas.
6. Arrangement of Quotations and Mode of Reference
The justification of the meanings assigned to a word lies in the quotations from the Prākrit literature supplied in the entry after each meaning. To keep the bulk of the work as small as possible, only selected quotations are given, which are necessary to determine the meanings. When the word occurs in a long passage as can be seen from the varņakas of the AMg. canon, only such portion of it is quoted as is essential to understand it. The portion which is dropped is indicated by using three dots (...) in its place. As the dictionary is using all the Prākrit languages, it is necessary to represent all of them, at least with one quotation from each Prākrit if the word occurs in it. This will mean that if a Prākrit language is not represented by any quotation, one can conclude that the word does not occur in that language and if no passage is quoted under a particular meaning it means that that meaning is not found in it. Very often a given word with a given sense may occur in different phonetic shapes in different languages. In such a case it will occupy different places in the Dictionary. In view of these and similar considerations a purely chronological arrangement for quotations from all the languages taken together is not possible.
It is found convenient to put the seven Prakrits which have a sizable literature into three groups based on their linguistic and semantic affinities. A word having a religious, a philosophical or a technical meaning is likely to occur in AMg, JM and JŚ. but may not occur in M., s., Mg. and vice versa. Moreover a word may be found in any two languages within a group. There are a large number of words which are peculiar to Apa. having formal and semantic features not found in other Prākrits. Hence for the arrangement of the quotations these languages are divided into three groups : (1) AMg., JM.,

Page #528

A Comprehensive and Critical dictionary of Präkrit Languages (Introduction)
519
JŚ. (2) M., S., Mg. (3) Apa. The above sequence of languages within a group will be strictly followed, so that quotations from them will occur in that order. For example, if a word does not occur in AMg. then quotations from JM. and Jś will follow. The same arrangement will be valid for the second group, first M. and then s. and/or Mg. The Apa. quotations will come last, which is as it should be in view of the place of that language in the Middle Indo-Aryan group.
The sequence between the first two groups cannot be pre-decided. According to the evidence found either of them may get precedence over the other. Considerations like the phonetic shape of the word, grammatical peculiarities and meaning will be taken into account in this regard. The chronology of all the books used for this dictionary cannot be definitely settled vis à vis all the other books. Therefore a convenient sequence which will be followed in giving quotations is given in the list called 'Language and subject-wise classification with approximate chronology of the Prākrit works'. In the case of AMg. the traditional divisions of the canon are followed though no pretence is made of a strict chronological sequence. In the canon there are sufficient indications to believe that its parts do not belong to the same period and are often separated from each other by a long period of time. But for the convenience of reference, a book will be treated as one unit and all quotations from any of its part will come together. It should be particularly noted that even the demarcation of languages like AMg., JM. or JŚ. is not always clear and often remains doubtful. Traditions are divided as regards the number of the Prakīrņaka works and many of them are attributed to writers who lived long after the finalisation of the canon. For the dictionary of the present type such inaccuracies are inevitable. All that we can do in this regard is to label a word as older or younger.
For giving reference it may be noted that more than one edition of a book is required to be used for citations. This is because they differ in readings, orthography and divisions of the text. In a few cases a book has two versions, a shorter one and a longer one. In such a case a word may occur in one but not in the other. To meet this difficulty, references to more than one edition are given after the quotation, the second reference being in parenthesis. When we quote from a Cūrņi or a Țīkā we give reference to the folio, side (a or b) and line because without this help the explanation of a word in a given text cannot be easily traced in the commentaries. The system of numbering various parts of the text, except for the largest divisions is to say the least, chaotic and there is no agreement between different editions of the same text. The original divisions into śrutaskandha, Adhyāya, Uddeśaka

Page #529

520
Amrita
and Sūtra is completely neglected while preparing indices for such works, making it impossible to use them. The number of verses in a text is not the same in different editions. This is also true of Sūtra numbers, because the Sūtras are differently divided by the different editors. All these facts make it well-nigh impossible to trace passages from one edition to the other. This explains to some extent the length of the references, and the reader is requested to try all of them if he wants to verify the quotations. 7. Cross-reference
The last element of an entry in the dictionary is a cross-reference to another entry which is related to it in different ways. It may be a mere variant reading of the same lemma or an earlier or a later form of the same word occurring at a given place of the text. The way in which the work of the dictionary is planned and carried out, it is inevitable to have more frequent references to the words which come later than those which have already been given. To remedy this situation, recourse is taken to give the reference in both ways. However, great care is taken to choose the lemma which must be based on its usefulness for a detailed treatment. Another way adopted to help the reader is to put together all minor variants of a word in one place. Citations are also kept at the minimum by avoiding repetition at different places unless absolutely necessary. . 8. Material from other sources
In order to achieve maximum comprehensiveness, words which are found in the earlier dictionaries are included even when we are not in a position to verify them along with the quotation given there. Naturally such entries are few and are given in square brackets with the indication of the work from which they are taken. The same procedure is followed when we find that a meaning is assigned to a word which cannot be verified.
OOO
A Comprehensive and Critical dictionary of Prākrit Languages (Introduction)
BORI 1993

Page #530
दादयश्रयः माणिक्याचतिदितरोवरयावकाता Nanananalo