Collected Articles Of LA Schwarzschild On Indo Aryan 1953 1979

Author(s): Royce Wiles
Publisher: Australian National University
Catalog link: https://jainqq.org/explore/034460/1

Page #1
Collected articles of LA Schwarzschild on Indo-Aryan 1953-1979 compiled by Royce Wiles hArAvavAiTayAeM ikalAe ragaDAvataddAkA 26 tasmArahA akAsalidAstA pravidA aMtagaDa jgsii|| parikSAdAMtagaDaramA grADAvA

Page #2

Page #3
COLLECTED ARTICLES OF L.A. SCHWARZSCHILD ON INDO-ARYAN 1953-1979 2017-0940 Collected articles of LA Schwarzschild on Indo-Aryan PK1472 .H47 1991

Page #4
COLLECTED ARTICLES OF L. A. SCHWARZSCHILD ON INDO-ARYAN 1953-1979 compiled by Royce Wiles TIEN INSTITUTE LIBRARY Faculty of Asian Studies Monographs: New Series No. 17 Faculty of Asian Studies Australian National University 1991

Page #5
First published in Australia 1991 Set up and printed at The Australian National University. (c)1991 Faculty of Asian Studies, Australian National University Notes on the future system in Middle Indo-Aryan'; The possessive adjectives of late Prakrit'; 'Notes on the declension of feminine nouns in Middle Indo-Aryan; "Some aspects of the history of modern Hindi nahin "no", "not"; "Notes on some words meaning "immediately" in Middle Indo-Aryan': 'A study of some features of the imperative in Middle Indo-Aryan' 1953, 1954, 1956, 1959, 1961, 1965 Royal Asiatic Society. Reprinted by permission. Notes on the history of the infinitive in Middle Indo-Aryan'; 'Prakrit thakka, 'tired" 1955, 1958 Linguistic Society of India. Reprinted by permission. Quelques adverbes pronominaux du moyen indien [I]: 'Quelques adverbes pronominaux du moyen indien: Il les adverbes de temps'; 'Ghummira, gholira, "agite, branlani" 1956, 1957, 1962 Societe Asiatique. Reprinted by permission. Some forms of the absolutive in Middle IndoAryan': 'Notes on some Middle Indo-Aryan words in -II-': 'Some Indo-Aryan words meaning "all": "First, second" and "third" in Middle Indo-Aryan'; The Middle Indo-Aryan prefix vo- 'off and some phonological problems associated with it'; 'Some 'unusual' sound-changes in Prakrit'; 'Initial retroflex consonants in Middle Indo-Aryan' 1956, 1957, 1960, 1962, 1965, 1972, 1973 American Oriental Society. Reprinted by permission. 'Gleanings from the Vasudevahindi": "Notes on two postpositions of late Middle Indo-Aryan: tanaya and resi, resammi'; The indeclinable je in Middle Indo-Aryan' 1958, 1959, 1961 Bharatiya Vidya Bhavan. Reprinted by permission. Some sporadic changes of vowels in Middle Indo-Aryan' 1964 Mouton & Co., Publishers, The Hague, The Netherlands. Reprinted by permission of Kluwer Academic Publishers. Remarques sur quelques conjonctions du moyen indo-aryen 1968 De Boccard Edition-Diffusion. Reprinted by permission. 'Some interrogative particles in Prakrit 1968 Shri Mahavira Jaina Vidyalaya. Reprinted by permission. Variant forms of the locative in Middle Indo-Aryan' 1977 Bhagavan Mahavira 2500th Nirvana Mahotsava Samiti. "Distinction and confusion: a study of neuter plural endings in Middle Indo-Aryan (c) 1979 School of Oriental and African Studies, University of London. Reprinted by permission. This book is copyright. Apart from any fajn Beifing for the purpose of private study, research, criticism, or review, as permitted under the Copyright Asi, na part may be reproduced by any process without written permission of the copyright holder, National Library of Australia Cataloguing-in-Publication entry Hercus, L.A. (Luise Anna), 1926 Collected articles of L.A. Schwarzschild on Indo-Aryan 1953-1979. Includes index ISBN 07315 1220 0 1. Indo-Aryan languages. 1. Wiles, Royce, 1958-II. Australian National University. Faculty of Asian Studies. III Title. (Series: Faculty of Asian Studies monographs. New series; no. 17). 491.1 Distributed for the Faculty of Asian Studies by: Bibliotech, ANUTECH Pty Ltd, Australian National University. GPO Box 4, CANBERRA ACT 2601 AUSTRALIA Telephone (06) 249 2479 Fax (616/06) 257 5088 Preface CONTENTS Acknowledgements Publications by L.A. Schwarzschild on Indo-Aryan 1 Notes on the future system of Middle Indo-Aryan 2 The possessive adjectives of late Prakrit 3 Notes on the history of the infinitive in Middle Indo-Aryan 4 Quelques adverbes pronominaux du moyen indien [I] 5 Some forms of the absolutive in Middle Indo-Aryan 6 7 8 Notes on the declension of feminine nouns in Middle Indo-Aryan Notes on some Middle Indo-Aryan words in -II Quelques adverbes pronominaux du moyen indien: Il les adverbes de temps Prakrit thakka, 'tired' 9 10 Gleanings from the Vasudevahindi 11 Some aspects of the history of modern Hindi nahin 'no', 'not' 12 Notes on two postpositions of late Middle Indo-Aryan: tanaya and resi, resammi 13 Some Indo-Aryan words meaning 'all' 14 The indeclinable je in Middle Indo-Aryan 15 Notes on some words meaning 'immediately' in Middle Indo-Aryan 16 Ghummira, gholira, 'agite, branlant 17 First, second' and 'third' in Middle Indo-Aryan 18 Some sporadic changes of vowels in Middle Indo-Aryan 19 The Middle Indo-Aryan prefix vo- 'off and some phonological problems associated with it 20 A study of some features of the imperative in Middle Indo-Aryan 21 Remarques sur quelques conjonctions du moyen indo-aryen 22 Some interrogative particles in Prakrit 23 Some 'unusual' sound-changes in Prakrit 24 Initial retroflex consonants in Middle Indo-Aryan 25 Variant forms of the locative in Middle Indo-Aryan 26 Distinction and confusion: a study of neuter plural endings in Middle Indo-Aryan Indexes Grammatical index by Colin Mayrhofer Indexes of Old, Middle and New Indo-Aryan words by Royce Wiles Corrigenda vud 1228Ren 22 Wang De Dao Jian mm 8 N Nan Qu 8 159 Fen 19 15 16 V vi vii 37 42 52 57 69 77 82 89 99 104 111 117 128 134 141 146 153 164 191 199 223

Page #6
Preface On the occasion of Dr Luise Hercus's retirement in 1991 from the Readership in Sanskrit which she has held since 1973 in the Faculty of Asian Studies of the Australian National University, some of her colleagues, friends and former pupils have decided to reprint in book form the collection of 26 articles in the field of Middle Indo-Aryan studies which she published under the name of L.A. Schwarzschild between 1953 and 1979. Many of these articles will be well-known to scholars working in the field, but some of those published in commemorative volumes are less well-known and are not easy to come by. The majority deal with broad concepts, such as the future system, or the history of the infinitive in Middle Indo-Aryan, but even those which seem to be studies of individual words turn out, on cxaminiation, to be investigations of the fundamental principles upon which the words are based. Published together in this way these articles make clear the great value of the contribution which Dr Hercus has made to Middle Indo-Aryan studies, Their usefulness is increased by a Grammatical index and Indexes of Old, Middle and New Indo-Aryan words, compiled by Colin Mayrhofer and Royce Wiles respectively. K.R. Norman Cambridge

Page #7
PUBLICATIONS BY L.A. SCHWARZSCHILD ON MIDDLE INDO-ARYAN Acknowledgements Compiled by Elizabeth Kat Page no. 1951 Review of Jayakanta Mishra, A history of Maithili literature, vol. 1, Allahabad, 1949. Journal of the Royal Asiatic Society 1951:121-122. The credit for the idea to collect and reprint these articles with indexes belongs to Professor J.W. de Jong, without his interest and encouragement this compilation would not have been produced. Colin Mayrhofer generously agreed to compile the Grammatical index (pages 191-198) and so reduced the amount of work I needed to do. 1953 1 Notes on the future system in Middle Indo-Aryan. Journal of the Royal Asiatic Society 1953:42-52 1954 12 The possessive adjectives of late Prakrit. Journal of the Royal Asiatic Society 1954:127-136. The Faculty of Asian Studies of the Australian National University provided computer facilities for the indexing and also agreed to finance the publication of this volume as a tribute to Dr L.A. Hercus and the contribution she has made both to Indo-Aryan studies and the University. Without this help and the permission of the copyright holders this volume would not have been possible. Any faults remaining in the work are of course mine. 1955 22 Notes on the history of the infinitive in Middle Indo-Aryan. Journal of the Linguistic Society of India (Chatterji Jubilee Volume) 16:29-34. Madras. 1956 28 Quelques adverbes pronominaux du moyen indien (1). Journal Asiatique 244:265-273 Royce Wiles Canberra 37 Some forms of the absolutive in Middle Indo-Aryan, Journal of the American Oriental Society 76/2:111-115. 42 Notes on the declension of feminine nouns in Middle Indo-Aryan. Journal of the Royal Asiatic Society 1956:181-190. 1957 52 57 Notes on some Middle Indo-Aryan words in-II. Journal of the Royal Asiatic Society 77/3:203-207. Quelques adverbes pronominaux du moyen indien II: les adverbes de temps. Joumal Asiatique 245:241-252. Prakrit "hakka', 'tired'. Journal of the Linguistic Society of India (Turner Jubilee Volume) 19:311-318. 1958 69 77 Gleanings from the Vasudevahindi. Bharatiya Vidya 18/1:22-26. 1959 82 Some aspects of the history of modern Hindi nahin'no', 'not'. Journal of the Royal Asiatic Society 1959:44-50. A listing of all the publications of LA. SchwarzschildLA. Hercus to 1990 including the work on Australian Aboriginal languages (published under the name LA. Hercus) is found in Language and history: essays in honour of Luise A. Hercu edited by Peter Austin et al. Canberra. The Australian National University, 1990. (Pacific linguistics. Series C. 116). pages 13-20. Isobel White has contributed a biographical introduction to that volume, pages 1-11.

Page #8
Page no 89 Notes on two postpositions of late Middle Indo-Aryan: tanaya and resi, resammi, Bharatiya Vidya 19/1-4:77-86. Page n. 1964 Review of H.W. Bailey, ed. Indo-Scythian Studies: Khotanese Texts V. London, 1963. AUMLA Journal of the Australian Universities Modern Languages Association) 22:340-341. Review of Radhagovinda Basak, od. Asokan inscriptions, Calcutta, 1959. Journal of the American Oriental Society 79/4:290-291. Review of Robert Birwe, Der Ganapagha zu den Adhyayas IV und V der Grammatik Paninis: Versuch einer Rekonstruktion, Wiesbaden, 1961. Journal of the American Oriental Society 84/4,459-460. 1960 99 Some Indo-Aryan words meaning all'. Journal of the American Oriental Society 80/1:13-17 1965 134 Some sporadic changes of vowels in Middle Indo-Aryan. Indo-Iranian Journal 8:25-31. Review of Paul Hacker, Zur Funktion einiger Hilfsverben im modernen Hindi, Mainz, 1958. Journal of the American Oriental Society 80/2:167-168. 141 The Middle Indo-Aryan prefix voor and some phonological problems associated with it. Journal of the American Oriental Society 85/3:350-354. Review of Un oditto bilingue greco-aramaico di Asoka Testo, traduzione e note a cura di G. Pugliese Carratelli e di G. Levi Della Vida, con prefazione di G. Tucci e introduzione di U. Scerrato, Rome, 1958. Journal of the American Oriental Society 80/2:155-157. 146 A study of some features of the imperative in Middle Indo-Aryan, Journal of the Royal Asiatic Society 1965:92-98. 1966 1961 104 The indeclinable je in Middle Indo-Aryan. Munshi Indological Felicitation Volume, Bharatiya Vidya 20-21/1-4:211-217. Review of D.H.H. Ingalls, An anthology of Sanskrit court poetry, Harvard, 1965. AUMLA (Journal of the Australian Universities Modern Languages Association) 26:303-304. 1967 111 Notes on some words meaning 'immediately in Middle Indo-Aryan. Joumal of the Royal Asiatic Society 1961:39-44. Review of Vladimir Miltner, Early Hindi morphology and syntax, Prague, 1966. Archiv Orientalni 35:684. Review of H.W. Bailey, Indo-Scythian Studies: Khotanese Texts IV, London, 1961. AUMLA (Journal of the Australian Universities Modern Languages Association) 16:252-253. 1968 153 Remarques sur quelques conjonctions du moyen Indo-Aryan. In Melanges d'indianisme a la memoire de Louis Renou, 671-676. Publications de l'Institut de Civilisation Indienne 28. Paris: de Boccard. 1962 117 Ghummira, gholira, 'agite, branlant'. Journal Asiatique 250:65-75. 159 Some interrogative particles in Prakrit. In Shri Mahavira Jaina Vidyalaya Golden Jubilee Volume, 204-209. 128 'First', 'second and third in Middle Indo-Aryan. Journal of the American Oriental Society 82/4:517-522 1969 Review of C.G. Hartman, Emphasizing and connecting particles in the thirteen principal Upanishads, Helsinki, 1966. Indo-Iranian Journal 12/1:3435. Review of Siegfried Lienhard, Tempusgebrauch und Aktionsartenbildung in der modernen Hinds, Stockholm, 1961. Journal of the American Oriental Society 82/3:434-436. Review of Ludwig Alsdorf, Asokas Separatedikte von Dhauli und Jaugada, Mainz, 1962. Journal of the American Oriental Society 83/3:379-380. Review of R.S. McGregor, The language of Indrajit of Orcha, a study of early Braj Bhasa prose, London, 1968. Journal of the American Oriental Society 89/3:636. 1963 Review of J. Bloch, Application de la cartographie a l'histoire de l'IndoAryen, Paris, 1963. Journal Asiatique 251:119-120. Review of Debi Prasanna Pattanayak. A controlled historical reconstruction of Oriya, Assamese, Bengali and Hindi, The Hague, 1966, Indo-Iranian Journal 12/1:35-37. Vill

Page #9
1970 1971 Page no. Review of V. Trenckner et al. A critical Pali dictionary, vol.2, fasc.4, Copenhagen 1967. Journal of the American Oriental Society 89/3:666. 1974 Review of L. Alsdorf, Die Arya-Strophen des Pali-Kanons, metrisch hergestellt und textgeschichtlich untersucht, Mainz, 1968. Journal of the American Oriental Society 90:586-587. Review of P. Gaeffke, Untersuchungen zur Syntax des Hindi, The Hague, 1967. Journal of the American Oriental Society 90:359-360. Review of S.K. Chatterji. The origin and development of the Bengali language, vols 1 & 2 (1st published Calcutta UP, 1929), London, Allen and Unwin, 1970, AUMLA (Journal of the Australian Universities Modern Languages Association) 36:276-277. 1972 164 Some 'unusual' sound changes in Prakrit, Journal of the American Oriental Society 92/1:100-104. 1973 169 Initial retroflex consonants in Middle Indo-Aryan. Journal of the American Oriental Society 93:482-487. Review of Paul Thieme, Kleine Schriften, vols. 1 and 2. Indo-Iranian Journal 15/4:290-291. Review of A critical Pali dictionary, vol.2, fasc.6, Copenhagen 1970. Journal of the American Oriental Society 93/3:376-377. Review of Raimo Anttila, An introduction to historical and comparative linguistics, New York 1972. Journal of the American Oriental Society 94/2:258-259. Review of K.L. Janert, Abstande und Schlussvokalverzeichnungen in AsokaInschriften, Wiesbaden, 1972. Journal of the American Oriental Society 95/3:547-548. [and R.K.Barz], Review of D.P. Kapp, Das Verbum paraba in seiner Funktion als Simplex und Explikativum in Jayasis Padumavati, Wiesbaden, 1972. Journal of the American Oriental Society, 94:491-492. [with R.K.Barz], Review of M.K. Verma, The structure of the noun phrase in English and Hindi, Delhi, 1971. Journal of the American Oriental Society 94:492-494. 1975 1976 1977 1978 Page no. Review of Shashi Kant, The Hathigumpha inscription of Kharavela and the Bhabru Edict of Asoka - a critical study, Delhi, 1971. Journal of the American Oriental Society 95/2:333-334 Review of V. Trenckner et al., A critical Pali dictionary, vol.2, fasc.7, Copenhagen, 1971, Journal of the American Oriental Society 95:153-154. Review of Hermann Berger, Das Yasin-Burushaski. (Werchikwar): Grammatik, Texte, Worterbuch, Wiesbaden, 1974. Indo-Iranian Journal 18:122-123. Review of L. Cousins, A. Kunst, K.R. Norman, eds Buddhist studies in honour of I.B. Horner, Dordrecht, 1974. Indo-Iranian Journal 18:294-296. Review of A.M. Ghatage, R.N. Dandekar, M.A. Mehendale, eds Studies in historical Sanskrit lexicography. Poona, 1973. Indo-Iranian Journal 18:116117. Review of Manfred Mayrhofer, Die Arier im vorderen Orient - ein Mythos?, Wiesbaden, 1974. Indo-Iranian Journal 18:291-293. Review of V. Trenckner et al. A critical Pali dictionary, vol.2, fasc.8, Copenhagen, 1973. Journal of the American Oriental Society 96:460-461. [and R.K. Barz], Review of Peter Edwin Hook, The compound verb in Hindi, Ann Arbor, 1974. Journal of the American Oriental Society 97:560-562. Review of Wilhelm Rau, Metalle und Metallgerate im vedischen Indien, Mainz, 1974. Indo-Iranian Journal 19:152-153. [with R.K. Barz), Review of Nilrata Sen, Early Eastern NIA versification, Simla, 1973. Journal of the American Oriental Society 97:376-377. 175 Variant forms of the locative in Middle Indo-Aryan. In A.N. Upadhye et al. eds, Mahavira and his teachings, 77-87. (Bhagavan Mahavira 2500th Nirvana Mahotsava Samiti) Bombay. Review of Adelheid Mette, Pind'esana das Kapitel der Oha-nijjutti uber den Bettelgang, Wiesbaden, 1973. Indo-Iranian Journal 20:294-295. Review of G.H. Schokker, The Padataditaka of Syamilaka, vols. 1 and 2, The Hague-Dordrecht, 1966-1976. Indo-Iranian Journal 20:278-280. -xi

Page #10
1982 Page no. 1979 186 Distinction and confusion: a study of neuter plural endings in Middle IndoAryan, Bulletin of the School of Oriental and African Studies 42:329-333. 1985 [and R.K. Barz], Review of John D. Smith, The Visaladevarasa: a restoration of the text, Cambridge, 1976. Indo-Iranian Journal 20:295-297. Review of George Baumann, Drei Jaina-Gedichte in All-Gujarati, Wiesbaden, 1975. Indo-Iranian Journal 21:70-71. Review of Colin P. Masica, Defining a linguistic area: South Asia, Chicago and London, 1976. Indo-Iranian Journal 21:68-70. Review of Kiyoaki Okuda, Eine Digambara-Dogmatik: das funfte Kapitel von Vaakera's Mulacara, Wiesbaden, 1978. Indo-Iranian Journal 21:71-72. Review of R.C. Zachner, The Bhagavad-gita, Oxford, 1969. Indo-Iranian Journal 21:68. [and F.B.J. Kuiper, T. Rajapatirana, E.R. Skrzypczak, eds), Indological and Buddhist studies: volume in honour of Professor J.W. de Jong on his sixtieth birthday. Canberra: Faculty of Asian Studies/ANU Press. 692 pp. [and R.K. Barz], Review of M.C. Shapiro and H.F. Schiffmann, Language and society in South Asia, Delhi, 1981. Indo-Iranian Journal 28:295-296. - xii - Notes on the Future System in Middle Indo-Aryan By L. A. SCHWARZSCHILD formation of the future tense is often mentioned in discussions on the classification of the modern Indo-Aryan languages. An -8- type of future occurs in Gujarati, Lahnda, and Jaipuri, an -- type of future exists side by side with participial formations in Marwari, Hindi (Braj), and Bundeli, and in Bhojpuri and Awadhi (in the third persons only), as well as in Kashmiri, where it has assumed the meaning of a past conditional. Despite this Marwari 1 belongs to the -s- group, forms with -- being found in those dialects of Marwari that regularly have voiceless --<--. The explanation of the Kashmiri forms is similar (6> and -sy->-->-h- in Kashmiri). Thus -s-forms in principle are found over a large part of the North-West and West of India. Attempts have been made to trace back the modern conditions to earlier stages, and S. Sen claims that the two types of future go back to different Indo-European originals: "From early times there were dialectal forms with the base-afflix -ha, which became quite dominant in Apabhrama." (This is incorrect.) "The origin seems to have been the I.E. stem affix -so-, OIA. -sa-, occurring in the desiderative, the aorist, and as a root-determinative." The accepted opinion is that the two types of future have the same origin, the OIA. future suffix -i-gya-, and that differentiation did not take place till the Middle IndoAryan period. As has been shown conclusively by Turner and by Bloch, the future suffix of Sanskrit could have in MIA. a special development into h- instead of -sa- which is phonetically regular, because it was in a weak position, being a terminational element. The change of -as-to-h-seems to have started in verbs whose stem ended in a diphthong or long vowel, as is proved by the Pillar edicts of Asoka, which have dahanti, hohanti (Delhi Topra, Edict VII), while the Dhauli edict further has the second person plural chatha. 11 am indebted for help to Professor Burrow and Professor Turner, and to Dr. W. S. Allen for information on Rajasthani 8. Ben. "A comparative Grammar of Middle Indo-Aryan," Indian Linguistics, vol. xii, 1951, p. 110. H. Smith, in J.A., eaxl, pp. 160 ff., however, goes back to pre-Vedie, differen tiating between a proto-Prakrit --i-fi and the ordinary Indo-Aryan -synti, R. L. Turner. "The phonetic weakness of terminational elements in IndoAryan," JRAS., 1927. "The future stem in Aaoks," B808., vi, 1930-32. J. Bloch, Les Inscriptions d'Asoka (collection E. Senart), Paris, 1950. -1

Page #11
NOTES ON THE FUTURE SYSTEM IN MIDDLE INDO-ARYAN 43 There are no examples in the other Rock Inscriptions, which might be an indication that the change was specifically eastern. This is confirmed by later evidence. The change of s to -- took place via the intermediate stage which is attested in the oldest parts of the Jains canon in verbs whose roots end in a long vowel. The prose of the Ayarangasatta han disime and dame (Ayar., II, 5, 11-13, repeated II, 6, 8-9). side by side with dihime, and the Soyagadangnutta has a third plural orali (Say., 1, 11, 29) as well as chipt and escanti? Apart from these few intermediate forms the Inngange of both the prose and verse sections of the Jains onnon is very similar to the Pali of the Gathis as regards the distribution of the future endings The aflix is thus scarcely found in the first persons except in verbs ending in a long vowel. A first person plural bagimo occurs in the Bhagavatista VII and tundidmi and na m i in chapter xviii, which is probably one of the later sections of this heterogeneous text: affih (or jafikami ording to some MSS) curs in the Nirayavaliyao IX, which is also late, and there are other spordie examples in very late canonical texts such as the metrical Prakiras. In the second and third persons, however, the forma in und -w- alternate freely, e.. the MSS. of the Vivagasiya live with alternating with celissai in the same passage (1,31). The endings are favoured at first only in verbs that end in a vowel; they become more and more frequent in consonantal stems in the later part of the canon, and they are practically the rule in the second and third persons in the metrical Prakiras, the Anuyoga viimastes, and other late texta. The reason for the difference of treatment of the first person is probably this: the change of-s-to-, being peculiar to the weak terminational elements, takes place most easily in the most frequently used form, which are those of the third person singular. The 38 futres persist for some time in the third person plural, but much longer in the first persons. This greater conservatism of the first person is by no means isolated, as is shown by later developments It is also sul for the third person singular to show the two main effects of frequency of use on the one hand rapidity of phonetio evolution, on the other resistance to morphological analogy. The particular frequency of the third person of the future is connected The des of the Uttaridhyaya r , XVIII, 90, of Charpentier's edition make for present indstive pive of diri 14 NOTE ON THE FUTURYSTUM IN MIDDLE INDO-ARYAN with the fact that the first on eally can express both will and futurity, the thinlar n leser extent the second may expires only futurity without the more personal nuance of will or obligation, so first place, thin synt. This explains the future of modern Awali and Bhojpuri, where the participial forms have replaced the nel future only in the first and second persons A parallel is provide luy theme of sonary envlinge in the future (found in the prone of the Svetambara canon, and particularly frequent in Jaina Saursen). Gray has discussed the way in which the terminations of the injunctive and the norist influence the future tense, and has been able to trace any such influence in the third person singular. Classical Malarstri han composite future tense. The MSS. of Han agree in using in the first persons (one exception: bila, variants limi, krimi, vers 187), and in the other persons (again only one exception: T issi, second Telinga version, verse 962). As pointed out by Nitti Dolci there is a close agreement between Tlala and the in Vamruci. Vararuci allows both -88- and in the first person, but insists on -- everywhere else, except for Saurant, where is permitted throughout (Var., VII, 13). Hemaennem and Kramadisvara follow him, with the result that later texts also standardize the usage. The Grammarians thus fix what is only a transitory stage, and the change of -85-to-may therefore be said to belong to Mahiristai as much as to Jaina Prakrit. Jaina Saurasenlthe language of the Digambara Canon, on the whole forms the future with the affix -85, exclusively so in the major works of Kundakunda, the Dhavali (cf. Hiralal Jain's edition of the Satkh a , Ammoti, 1940), the Gommata , and the Tiloyupannatti, while in the less dogmatic works there are examples of the future. The-88- type of future is found in the Sauraseni passages of the Sanskrit dramas, and the views of the grammarians conour with this evidence. The inscriptions ennnot afford much help in localizing the Prakrit conditions still further, as the examples of the future L H. Gray, "Olmersto n Middle Indian Morphology." 2808 vil (1935-37). Der Sadek derle, A Weber, Leipale. 1531. L ole, Lexu a rit, Puis, 1938 Decke, Milfiger D e n Festle HJacob, Bonn, 1925 Melale, Historical f unctional Paris, Poona, 1948

Page #12
NOTES ON THE YUTURE SYSTEM IN MIDDLE INDO-ARYAN 45 terse occurring in them are very few indeed, but the exclusive use of aud -sy- forms in the Klarosti inscriptions from Chinese Turkestan proves that the development of -88-to-- did not feature in North-Western Prakrit. It seems clear therefore that in the first few centuries A.D. the -|- future was characteristic of Jain Prakrit and Maharisti, while the sibilant belongs to the more western dialects, particularly Saurseni. At this stage a further development takes place: -88- weakens to -3- in the popular language of those districts where the sibilant was still preserved in the future tense. The reduction of double consonants occurred generally at & much later date and is one of the distinguishing features between Apabhrama and the modern vernaculars. The simplification is therefore again a peculiarity of the weak terminational element. It is reflected in Apabhrama texts and in such Prakrit works as are influenced by popular speech; it curs in the Vajjalaggam of Jnyavallabha, which has hosgiebhansali. The main bulk of Apabhram a literature is written in two dialects, Digambarn and Svetinbam Apabhrams, the latter being the language of the Svetambara Jains of Gujarat. There are a number of differences between the two dialecta, such as the endings of the genitive singular of nouns, and the ending of the third person plural of verbs -ahi in Svet. Ap., ants in Dig. Ap.), and many others, listed by Alsdorf and Modi. None of these differences are absolute and there is a great deal of interchange of forms. The future originally has the affix -- in both dialects. What is probably the earliest Digamhara work, the Paramatmaprakiss of Joindu (sixth century ?) has only an -- future (but third pl. shikimin Joindu's Yogasara), and the same applies to the parts of the eighth century Paumacariu of Caturmukha and Tribhuvann Svayambhu that have been published so far. In the later texta -- forms are sometimes found. As -- was slow to appear in the first person even in the neighbouring -- languages which influenced Apabhrama, it is not surprising that it is very rare in that 46 NOTES ON THE FUTURE SYSTEM IN MIDDLK INDO-ARYAN person in Digambarn work. There is only hom (var, hohamm) Mahapurana, VII, 18, and kilikin and pickihime in the Jasaharacariu. In the third person singular resistance to the influence of the fornis was great. There is one example of lachi in the Harivamparina of Pandanta as opposed to twenty-one forms in ea, and the distribution is about the same in the other texts. Though the Nagakumirearstill has only-, inost of the texts already prefer the -- aflix of the future in the second person. In the third person plural the forms have practically become the rule. Only the Kamandacaru of Kanakamara, which is later than Puspadanta, whom it mentions, and which is more influenced by Gurjata Apaburmia (as is slown by the preference for the ending ki in the 3rd pl.) has the forms cadeam and Kareshiw. One can therefore conclude that Diganiban Apabhrama is basically an --dinlect, which has undergone the influence of -- forma particularly in the third person plural. The language of the Svetambar Jains became the literary medium of a vast region and was brought into contact with many spoken dialects and with the literary influence of Jains Prakrit. The -- forms of the future are therefore found as well as in most of the texts. The Sanntkumirncari of Haribhadra (1159) already has Karin, mariai, visite, and there is an even greater mixture in the Kumarapalacarita of lemacandra, the Kumarapalaprati bodha of Somaprabhu (1195) and Abdul Rahman's Sandesausaka (twelth century). Hemacandra in his grammar allows both forma (Ho, IV, 388 c). The confusion is still the same in the very late Prabandhucintamani which has belah side by side with such forms as avisi and jinisit. Svetambara A bhirama is thus only a very imperfect representa tion of the popular preech of the extreme west, which as is proved by the earliest texts in the vernacular did not have a future Old Gujarati, Okl Rajasthani). The confused state of western Vaidya, Mbuntu, Handy, 1937. Vnity, wh in, Kaja K. Alorf, Harmonlamurt. 13. Hall Jain, Ka r in, kann, 1934. Dr. Allen Mo tion Internation of the Old Tajnathani formi, on scent of the sounding wireless) in nome dialects of Marwari, the apeners of which writo - But since this voicedes sed, he pointa t erl a nt , this interpretation would still prove the point for the age of the worst in the A lhami period and how the recondary name the modern conditions in Rajsthani J. Laber, der das vajala do Jaywallas, Leipzig, 1919. L. Alsdorf, Apabbra Shidien, Hamburg, 1937. M. C. Modi, Apabranapithimi, Ahmadabad, 1936. M. C. Modi and Bhayani Introduction to the Pawmasiraris by D ila, edited In Singhi Jain Series, 1943. Upadhys, Jolandadena's Par mu r alla and Yogarine, Bombay, 1987, M. C. Bolt, amikha Bw a ne Tribh Susu . Bharatiya Vidy I. pl. 3, Bombay, August, 100.

Page #13
NOTES ON THE FUTURE SYSTEM IN MIDDLE INDO-ARYAN 47 Apabhrams cannot therefore be taken as a direct indication of modern conditions as has been suggested recently by Tagare! T CONNECTING VOWEL It is almost certain that the simplification of- to-, both in its passage to -- and later in Apabhramla brought about a compensa tory lengthening of the preceding vowel, which was shortened again on account of its weak position As shown by Turner, the Adolcan examples of the Delhi Topra diet cannot be adduced as evidence of the transitory longthened form. Neither does it appear in the standard worloes of the Jains canon, being only found in the loss regular texts. Thus the Marni samadhiprakima repeats mar (Ist sg. verse 940 onwards), and the Mahansiasutta contains cas (Ista) and Marikinti (3rd pl.). Non-canonical literature continues this tendency, and again the first persona sem to be the most conservative. The Paumacariya of Vimalasari,which is probably as early as the first century A., las cumbum (Istag, 10, 12), gamit (3rd g., 14, 13). karilan (Ist sg., 26, 67), cte, as well the more doubtful haiti (3rd pl., 118, 63). The Vasudevahindi' of the sixth century has bhanjian, ghai (Bhavnagar lition, I, 51, 92), pucoliamo (1, 89, 19). jimo (1, 91, 8), ete, and the Dhirtaklina of Haribhads of the eighth century Warham, Walime and tecilimo. By the time of the last work Join Mahirstri mast lave ceased to be a spoken medium, but the lisence of such forms from more polished work of Haribhadra, the Samarilocakahi, shows that the lengthened vowel was still felt as a vulgaris. It was a popular feature that survived into Apabhiramla and is found in the Vikramorvai: pakkimi, jikisi, karilini. Later the vowel was definitely shortened and even appears as an in some dialects (e.g. Chattisgarhi deblo). The normal Digambara Aabhramla future is in Forms in short -- are extremely rare, there is only bolli (Ist sg.) in the late text, the Karakandaonriu of Kanakamans (VII, 11, 18). The explanation of the usual forms in e , , et, is 48 NOTES ON THE YUTUBE SEM IN MIDDLE INDO-ARYAN not obvious Jacobi thought that the simply represented Jengthening: "So ist das Futurum sufi durch Ersatadchnung Rusiyati enstanden." Though the short vowels and are easily confused, there are only isolated examples of this happening with the long vowels, which were obviously more distinctive, and where there was nographie difficulty. The spelling with is so consistent in the future that it must be due to annlogy. Whint springa to mind is the influence of the causative forms in - from a which were so frequent in Prakrit. This is the explanation given by Tessitori and by Pigchel Aabhiramis vom ---Stamme tur (He., IV, 414, 4) wie J.M. (Av., 12, 12) vom e-Stamme su R." The prakritio type of conjugation in AS to be usual in Apabrania, where the ordinary thematie - has replaced it completely in conjugation, and one only finds remnants of it in passages under prakritie influence. It seems therefore unlikely that a type of conjugation which was already moribund should influence the ordinary future endling. (A parallel nae is that of the gerund, where has given , which alternates with shortened forms - -Pri, but note). It is probable that the futures with the connecting vowel -- were under the influence of those important root vers in which the -- was etymological cani, si, desai, lesai. This development, which he become general in Digambara Apabliramla is only spore in Curjana Apabhramla and in the Old Western Rajasthani text, the Dhol Maruni Doli, ete. It has given in to the forms of m e Labin, which have throughout except in the 1st pl. (land r es, mas, maresi, murid, mares, mares). In Svetambar Ayam the connecting vowel is - though is time found. The Bhivanandhi-pakaranam of Jaynilovanni, for instance, has only in the first person, and -isal, in the thind the Sanatlumacariu la larins, etc. (Ist ) in the third, while lemn candra quotes verses containing the first singulars p er andar. This is continued in the early stages of the modern maculars; the Vasanta Vilasa Phaga has a first g inn, fales, des, which is the usual mixture Tr i al Grammar Apobre s, 1945. *H. Jacobil P a ris, Bhavnagar, 1914 * Aldorf "The Vodeship of the Jain Mabini," 1908. vill, 1935-37. D e Harid , ed. N. Upithye, Singh Jain Berles, Bombay Job b , Introduction, P2, Munich, 1918 Pachel, Gramatik der Pra d e, 037, Str . 1000 Tel Noles on the gr of the Western i ni, A., 1914-16. M. Mo " The ananchipelag of Jayadevamoni, ADORI. xl,

Page #14
NOTES ON THE YUTURE SYSTEM IN MIDDLE INDO-ARYAN 49 of forms found in the earliest texta from Gujarat, such as, for instance, also the Bharatc var. Bahubalirasa The slightly later texts examined by Tessitori have-iso (Ist sg.). Tai (3rd sg.). - (3rd pl.), all this despite the danger of confusion with the <-iy, beside -- <-syys) of the passive. The long -7 has remained in modern Gujarati in the first person singular, while elsewhere the shortened fornis in prevailed, generally weakened further to --- Examples of this are already found in the OU Western Rajasthani texts, and are frequent in works such as the Nasaleta-ri Katha (O. W. Marwari), which has a first person karas , second lablasi, third sigas, and the Suvabaluttarikatha (probably Eastern Rajasthani), which has a first sg, karise and karasu, second and third karsi. The connecting vowel of the future tense thus shoves again that the final repartition of forms, lowever far back its origin, belongs to the period of the modern vernaculars. THE VOWEL OY THE ENDING It is found in the Jaina canon, as in Pali, that forms with the future aflix -- change the vowel of the ending to -1, while those that preserve the-88- also preserve the vowel ---, so that on the one hand Sanskrit at gives -ih, on the other isso. Apart from the first person and a few isolated forms such as the Pali kuhati, there are practically no exceptions to this rule. In fact in the Jainn canon the -Is-adfix is so much associated with the vowel -- that it influences the conjugation of the present indicative of verbs whose roots happen to end in-, so that one finds repeatedly analogical formations such as winti, lahinci, oto. (Paghuvayaranaim, 10th Anga). The fact that in the first persons only the stages and later -- (not yet -hi) are found in the Jaina caston shows that the change must have taken place vis-issadi > sali >thai > - > - It probably began in the third person when -83-was reduced to-h-, across which assimilation to the connecting vowel could take place ensily, and it was further helped by the presence of final - A future ending-hi can therefore he said to be curacteristic of the East Central region of India, where Jains Prakrit was formed, and it T'inchel, Materialien Kenntnis der Arbeit, Berlin, 1902 10. K e Naalarl Karold Hathanie, Lopai 1925. J. Hertel, therdan S u rf Ko , Festschritt E. Winelioli, Leipaig, 1014 CE H. Smith's article (loc. cit.) for different explanation. He postulates separate pre-Vodie origine for the and forms. RAS. APR. 1019. DO NOTES ON THE FUTURE HYTKM IN MIDDLE INDO-ARYAN is precisely in part of this region that one finds a continuation of the - type of future in the male vernaculars: Bhojpuri and Chattisgarhi (vokli, velik ). Wherever Digambon Ab o pter future forns from a neighbouring dialect, that is particularly in the third plural, they we had on the t h e lines in the works of lugdanta not only thin pl. Iniiwi, wiki, te, but also likiwei picchiwi (Jasharacariu): Whihisi, wikis (aunlogical longcomes from the third person): wrik, hol, te, Futures in--are extremely rare and contined to a few obvious western borrowings, especially in the Karakatdach riu. It was therefore a dialect of the East Central region which influenced Digambara Aabhrama. Slightly further towards the weat the -ha type of future prevails. It seems to have been climacteristic of the most central lialect, Maharastri, and to use spread from there. Hala (particularly MS. R) preserves a number of rases of -ha-despite the influence of the grim s : Trini, lakini, che, etc. The Lilavailala, which claims to have been written in pure Malarstei (about A.D. 80)) contains viniluns, dixhari, pleasi, side by side with lahtikisi, eto. The Jaina Saursent of the less dogmatic of the Digambara texts Boens to live adopted the landing quite freely. The Bhagavati Archana, for instance, has iniisi (y. 1008), Kohali (v. 338), Rohani (v. 1099): Vattenka's Malacara has ww. chasi (II, 90). vladi (II, 3, 32), eto. There is often even a metathesis of the --; shahidi (Blag. Arach., . 749), mucidi (v. 1919), etc., and these * wrong" forms are too numerous to be dismissed as ordinary mistakes. The borrowel antre of the -I-futures in Jaina Sauraseni, from Jaina lakrit on the one hanel (the Malacarn has been said to adapit a large section of the Aurapaocakkhana-Panna with only minor linguistic alterations), and from Maharastri on the other, might explain the hesitation between the-he-and-ho-future and the uncertainty as to where to put tho . As distinct from Digambar Apabhramia, the language of the Svetamburs of Gujarat was influenced by the spoken language of the centre, and one finde ici, -hi, etc., almost invariably in those cases where the future has been introduced. The preference for the vowel-c-in the ending proves that this was not a mere prakritiem, HAN. Upadhye, Lol. A romantic Katy i rana Prakrit of Kowhala, Singhi Jain Series, 1949.

Page #15
NOTES ON THE MIDDLE SYSTEM IN MIDDLE INDO-ARYAN 51 as is so often thought, since then one would have expected the Jains Prakrit ending - The influence of the Midland continued in the early period of the modern vernaculars and the type of future spread further and further from the region of Western Hindi to which it belongs. Thus in the sixteenth century Awadhi of Malik Muhammad Jaynsland of Tulsi Din the forms still prevail in the third persons, while elsewhere the future is found (Tulsi Dis: Ist a n, 2nd -hasi, 3rd karili, puiki, etc., 2nd pl. a , 3rd pl. karibi, kuria). In modern Awadhi (3rd sg. mariai, 3rd pl.w ikiai) the Hindi (Braj) forms have replaced the older type of conjugation. Futures in coeur even in the texts of the extreme east: #second pl. future imperative ai (probably basikal < iyatha) is found in the Dohas of Sarah, a work admittedly under Western influence; the few forms with i n the Old Bengali Caryus and the third persons in the hai, in the early Middle Bengali of the Srikroa-kirttann of Cand have the same explanation. The future is a western importation in the Maithili "Avahattha" of Vidyapati Thakura, where there is a distinct Braj influence. In purer types of Old Maithili one only finds participial futures (eg, the Parijataharana of Umapati Upadhyaya, perhaps enrlier than Vidyapati). These imported future forms, absent from the modern Maithili and Bengali languages, show the extraordinary vitality of the language of the centre in the early pluses of the modern vernoulars. There was also a variant development of the ending isyati, in which assimilation took place across the sibilant: issidi, which is frequent in the Saurseni portions of such dramas as were preserved in southern MSS. The change seen to have been confined to the third person. Examples already occur in the South Indian Inscriptions of Asoka radhis site is common to the minor Rock Edicts of L. Dhu, Podul linguistic dy of the ni c estry Audi (Hindi). London, S. 3. Sales , "The verb in the m an of Tule Din," Allahabad Unin. Sudio, 53 NOTES ON THE FUTUR SYSTEM IN MIDDLE INDO-ARYAN Ruprath, Balmagiri, awl Siddapur. The only northern manu script in which suche forseer is the Saurent section of Asvagho's dramas and the contracted Although forms obviously belong to the extreme south, the grammar and them aren, perhaps on account of the presence of he's theory that Jain Busent is meant by this las been disproved by the texts, where there is no sign of such futures apart from the one curious double for bhavissibili (Bhat. Aradh, v. 1488). It is difficult to establish the later history of this ending, because the most southern language, Marathi, has not preserved the sigmatie future (although Tulpule and others claim to have found traces of it in ORL Marathi), and because the tendency to assimilate the to the surrounding vowels continues, so that the forma in is, which one finals o often in the Old Western Rajasthani text and even in isolated in Svetambara Apabirusa) are more likely to be on the normal Svetimbar future in than on the southem forma reflected in the Asokan inscriptions. One can thomfort conclure on the evidence of the modern Verneurs and of the older text that there coexist in Middlo Indo-Aryan five main future and although they are not all attester at the same period (1) - Thi ART CENTIK (attested in Jina Pralcrit). CENTRE (in Milanytri). (3) si WEST CENTER (orought into contact with 1) in Borar, Digambarn Apabhramsa). WEST (Svetimbara Apabhirami). (5) - EXTREME SOUTH (Southern Saurseni manuscripta). This listribution was obscurely intricate interborrowing of forms and the in ce of the language of the centre. Lades, w e ndiwber chel, der with *N. . Tule, Yadar Mar , Berlin, 1911 33. , Bombay, 1912 * N. Chudhuri, Studies in the Aparate of the Debirga, Caletta, 1910. Shahidullah, Lerchants mystique de Kanha de Baraka Chatterji, Origin ander e Bengali la Caletta, 1996 Hallar"Vypt the Mathil pool and language". De Lell, Cole 1900. Sak , The language of the King Indian Linguistic, Gron, "The Pirita Haruna of Umipati Upadhyays W. Brand Oria Orientale Society, 1917) 10

Page #16
The Possessive Adjectives of Late Prakrit Br L. A. SCHWARZSCHILD Poss OSSESSIVE adjectives are not strictly essential parts of speech; their place can often be taken by the genitive of the personal pronouns. This has led to a certain lack of continuity in their development, which has, however, often been exaggerated. Apart from very isolated survivals like maia, the Sanskrit possessives had already died out in Prakrit. The Modern Indo-Aryan vernaculars have entirely new formations, the most widespread of which is that in . In the singular one can distinguish three main groups of possessive adjectives in r (1) Western Hindi, as characteristic of the first group has mero, tero for the possessive. (Oblique forms of the pronoun are mo, muj, mohi, muhi, and to, tuj, tohi, tuhi.) (2) Gujarati and most of the Rajasthani dialects except Mewati belong to a second group whose main characteristic is the vowel --: Gujarati mara, taro (oblique base ma, ta). (3) Eastern Hindi and the Magadhan languages on the whole have -o- as the vowel of the possessive, thus Avadhi mor, for (oblique mo, to). The absence of any similarity with the oblique in group I shows that the possessive adjectives are not simply based on the modern oblique forms of the personal pronouns, but are older. The modern possessive adjectives and their distribution are in fact already foreshadowed by the Apabhramea texts. The standard forms of Western Apabhramsa as given by Hemacandra are: 1st person: mahara pl. amhara 2nd person: tuhara These forms as well as a second person plural tumhara recur in the other Western Apabhramka texts, generally with the addition of svarthe-ka; e.g. Kumarapalapratibodha maharau, Sandesarasaka amhariya, etc. Possessives are rare in this dialect. Although there are no possessive adjectives of this kind in the earlier Digambara texts, the works of Joindu, and the few fragments of the Paumacaria of Svayambhu that have been published so far, there is a large number of forms of the possessive in the Hemacandes, iv, 434. R. Pischel, Grammatik der Prakritsprachen, p. 308. Strassburg, 1000, -12 128 THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT later Digambara Apabhramsa texts. The plural is amharaya, tumharaya as in Western Apabhramsa. In the singular on the whole the forms of the type meraya, meri; teraya, teri outnumber those of Western Apabhramsa, which are also current. This applies to the major works of Puspadanta, but in the later Karakandacariu of Kanakamara the distribution is equal. In the Bhavisattakala, as pointed out by Alsdorf,' meraya occurs only once, while there are examples of tuharaya, tuhari. This distinction between Digambara and Western Apabhramka shows that there has been a continuity of evolution, that forms of group I never belonged to Western Apabhramsa and the languages derived from it, while Digambara Apabhramea represents a mixture or groups I and II. The origin of the Apabhramsa forms is not clear. The theory of Bopp and Lassen that the types I and II are a continuation of the Sanskrit madiya has been completely discarded since the days of Beames. The main theories since Lassen are the following: Sen derives the Apabhramea forms from maha and tuka (mabh(y)a, tubh(y)a) with the affix -ra or -ara, cf. OIA. -ra (la), ala, madhura, bakula, srira, etc. Dave derives the sixteenth century Gujarati forms maharas and taharau from ma+haran and ta+harau.-harau itself he refers back to Sanskrit ghors which was also to have given the postpositions Arau, hraim, rahaim. Beames states that "there is no difficulty in connecting these adjectives with the older genitive of the noun formed with the affixes kera and kara... Popular Prakrit has such forms as mahakero, from which mera would naturally flow, and the rustic form of the same, namely maro, would be equally derived from mahakaro. Pischel explains the forms mahara, mahara as based on the genitive sg, maha kara, developed from karya without epenthesis. In his discussion with Hoernle he makes it clear that he considers wera, etc., derived from maha + kera. L. Alsdorf, Harivasapurina, p. 166. Hamburg, 1936. J. Beames, A Comparative Grammar of the Modern Aryan Languages of India. London, 1876, pt. ii, p. 311. 8. Sen. Comparative Grammar of Middle Indo-Aryan Indian Linguistics, vol. xii, 47. T. N. Davo, & Study of the Gujarati Language in the Sixteenth Century. London, 1935, p. 59. R. Pischel, 14. 1873, p. 121. -13

Page #17
129 THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT This theory is supported by Chatterji and Tessitori among others, though the latter thinks that merau and moral, occurring in the Old Western Rajasthani texts" bear an analogy to the Braj and Bundeli oblique forms me and mo". Tagare derives mahara from maha+kara, mera from ma +kera, tuhara from tuha + kara; tera from tea> ta+kera. Of these explanations, that of Sen is improbable on account of the rarity of the suffix -re, fira, a variant of the frequent -ala of Apabhramsa, OIA. -ala cf. rasala, racala in the Nayakumaracariu. Forms of the possessives in -I- never appear in Apabhramsa or even in Eastern texts like the Caryagitis, although this is the more usual variant of the adjectival suffix. Dave's theory seems to account very well for the sixteenth century Gujarati forms, which might have been analysed as ma+harau subconsciously by the speakers of this period, when the original formation of the possessives had fallen into oblivion. The theory leaves the Apabhramsa maharau, etc., quite unaccounted for, and assumes them to be unrelated, as they cannot be based on ma+harau because of the vowel lengths and because harau did not exist as a postposition in Apabhramsa. Further, the series of postpositions that Dave quotes as cognate is not with certainty related to ghara. (For a different, but also unconvincing explanation of rahain, etc. cf. Tessitori). There remain the explanations which allow for some continuity in the evolution of the possessives and which involve the addition of a derivative of ky to the personal pronoun. Pischel and Tagare agree in postulating two forms "kara and kera, both from karya. As is well known, -kera is frequently used to indicate possession, and is often found with nouns and even more with pronouns in both Prakrit and Apabhramsa. In the case of pronouns it is almost invariably the genitive that is used before kera, with nouns it is sometimes the stem. (This renders Tagare's ma-kera, etc., improbable.) Thus Hala has maha-kera (v, 17); Sauraseni mamakera, lassakera, amhakera, etc., are found in the dramas from Bhasa onwards. Kera is fully recognized by the grammarians. Hemacandra, ii, 47, writes" idamarthasya kerah" (cf. the list of examples 18. K. Chatterji, Origin and Development of the Bengali Language. Calcutta, 1926, p. 813. L. P. Temitori, "Notes on the Grammar of the Old W. Rajasthani," IA. 1914-16, para. 83. Tagare, Historical Grammar of Apadhraman, Poons, 1948. -14 130 THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT given by Pischel in his commentary), and the form continues in Apabhramsa, (cf. Paumacariya of Svayambhu II, 118), and then in the works of Puspadanta (ef. Harivamsapurina, glossary: keraya), and it is used in the early stages of the modern vernaculars e.g. Old Guj. Bharatesvara-rasaka 105: jai risahesara kera puta. It still occurs in the sixteenth century Awadhi of Tulsidas and Jayasi: Bandon padasaraj sab kere (Tulsidas) and hauns panditan kera pacchalaga (Jayasi), etc., quoted by Ayodhyasinh Upadhyaya,1 kerau often also occurs in its unshortened form in the OWR. texts of Tessitori and it survives into modern Gujarati (cf. Grierson *), E. Hindi, and Bihari. The distribution of kera, -era as a genitival postposition seems to have no relationship with the distribution of the type mera, so that it is difficult to argue in favour of a protoHindi dialect which favoured epenthesis and where kera was generalized rather than the hypothetical kara. Thus both OWR. and Bengali have the postpositional affix (kerau, but the posses sive pronouns belong to groups II and III. A purely phonetic explanation would also make it difficult to account for the complete absence of hamera and tumhera in the plural. The development must therefore be to some extent analogical, and the analogy is most likely to have come from the genitive of the personal pronoun. A link with the personal pronouns can be seen at various stages in the development of the possessive adjectives, for instance in the Eastern forms mor and for, and in occasional rarer forms such as mujjhara (var. mujjhare) found in the Prabandhacintamani (p. 11, line 8). In Maharastri the intervocalic consonants were lost early and the affix kera, when in composition with a pronoun, therefore lost its initial consonant. The fact that the possessive pronoun was often used in an unstressed position in the sentence led to further reduction, mahaera> *mahera, and under the influence of me, te there were formed the possessives mera and tera. This influence of me must belong to the pre-Apabhramas stage, as me and te, though widespread in all the Prakrits as unaccented forms and recognized by the grammarians, became more and more rare in Apabhramsa (only one example of me, for instance, in the whole of the Harivamsa purana). The line of development of group I of possessive pronouns was therefore the following: Ayody sinh Upadhyaya, Hindi Bhaga aur make sahitya ka vikas, pp. 83 ff. G. A. Grierson, Kuhn's Zeilschrift, 38, 1905, pp. 913 ff. -15

Page #18
THE POSSESSIVE ADJECTIVES 09 LATE PRAKRIT 131 mahakera(ka) > mahaeraa) > erata) > meras > meri. thakera(ka) > thacrata) > tuberala) te > terac > teni. The early contraction of the vowels we in hiatus may be put down to the Inck of accentuation of the word, while the loss of al is due to the analogy of the pronoun. In the plural there was no doubt a similar influence of the Prakrit form of the genitive plural amhuinam, tumlinam, which was replaced by the Arabhrama period by amhala, wali, and anha, who 18 would normally go to a pointed out by Alsdorf: "Es ist auch unglaubhaft dass aus der Kontraktion von ac oder i jemals etwas anderes entstehen sollte als e." Therefore an analogical influence is also responsible for the plural forms: ambakera(ka) > macrofa) amhang > amhiva. Inhalera(ka) > merala) tumhanaux > kuhara. There was thus not a variation of postposition within the same dialect, but the analogy of the personal pronouns was responsible for the difference of development. 132 THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT and that kura does not occur until very late. On account of this late occurrence and the fact that kurz is restricted to Bengali, Hoernle thought it was a double genitive formed very late from ku + na (Gaudian Grammar, p. 236, footnote). Though this explanation is no doubt incorrect, it is quite probable that lara is only a late deformation of Zara itself in Bengali. It is certainly almost impossible that such a late form, restricted as it is to Bengali and not even common to the other Magadhan languages, should have given rise to the whole of the possessive formation of Western Apabhram and the languages derived from it. The reason for the adoption of this hypothetical kvira by Pischel and others seems to have been a phonetic one: waha and tha + kvira would easily give makar and tuhara, whereas the form vara, given first of all by Beames, presents certain phonetic difficulties in view of the question of the yarudi aka ay seems to have had a double development according to whether the y resulting from the l was pronounced or not in late Prakrit, so that the group sometimes gives and sometimes. There seems to have been a certain amount of hesitation and dialectal and chronological variation in the development of this group, which is further obscured by interborrowing and the tendency to substitute the suffix -ers for what would have been the correct ending of the word. It may be possible to explain a certain number of the inconsistencies by such a substitution of suffix rather than phonologically. Thus it appears that in Apabhrama in the case of some words, and in others in the early stages of the modern vernaculars the ending era was favoured. But it is not at all clear whether in those words akin has been replaced by akara which has developed to era via ayara, or whether there has been substitution of the suffix era tura and gambiry > gawbhira, a development which is unusual with the vowela, norya > acera, etc., not acara. Pischel insists that "Nobody can doubt that dra of Bengali originally was mura and that kira and era are only modifications of the same word, vix. Mirya." But Chatterji la shown that the postposition responsible for the creation of the Bengali genitive and plural forms W era originally R ichel, op. cit. 1 Heimer Smith, HSL, 31. p. 116. - 16

Page #19
THE TOSSESSIVE ADJECTIVES OP LATE TRAKRIT 133 by Professor Turner. There are even some examples in which the suffix e seems to have replaced an entirely different ending : e.g. Digambara Apabhrama revaya (Bhavisatta kaha, Harivamin purina, etc.) for xita. Although none of these cases are itbsolutely conclusive, they all seem to point towards a preference for the endinger in Digambara Apabhraya and not so much in the Western group of languages, where the postposition kera was not treated as a suffix o soon. A derivation from bara of the Western Apabhramsa possessives can therefore not be excluded on phonetic grounds, as all the case in which *alara gives erat might equally well be substitution of suffis. kara is attested in the earlier stages of the modern versculars, being used side by side with kera to form possessive constructions; thus kara occurs in a really old Gujarati text like the Buddhirasa, sixteenth century Awadhi has both are and kera, e.g. Tulsi Das katra jati Kara rosa, ete, further East it appears already in the Kirtilata of Vidyapati and is generally well attested. The origin of this Kara has been suggested by Baburam Saksena to be the elongated Prakrit form of the past participle of vler, lurila > kario; yet this seems unlikely as there is no sign of the in the oldest forms of this postposition that are found. It is therefore to be separated from the Rajasthani instrumental Kari. It is much more likely to be a weakened form of kera itself, so that one might assume the following line of development for the Western Aphabramsa mahiran : maha + Kerala) > maha + karaka > mahalarau > maharaw, and an exactly parallel development takes place in the second person. It seems therefore that in the West thek was not always slurred quite so early and the vowel was weakened first when this postposition was used onclitically, with the result that it disappears completely in OWR mai clara anha (a)ra amhanam > whara or whether it was brought about by the slight remnant of a vowel that may have remained in the enclitic kara before it was reduced 134 THE POSSESSIVE ADJECTIVES OV LATE PRAKRIT completely to you. The evidence of the nouns-the short vowel of the masculine nouns before the postposition, e.g. Nasaketa-ti Kathi proves nothing as the possessives were compounded at a much earlier date when the vowel -- before rau might still have been sounded. A possessive with a long vowel which might also have similar explanation is the isolated tajanau which occurs in the early Apabhrms of the Kuvalayamala. According to Master this is based on the genitive bujha with the addition of the postposition Wais, cf. Old Gujarati nau, which occurs in the texts studied by Tessitori. The use of medial cerebral in the Apabhtasa form, however, renders doubtful that identification with naw (cognate with wafPage #20
THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT 135 had a considerable extension in the early period of the modern vernaculars. There are sporadic examples of morau, etc., in the OWR. texts and in later works such as the Dhola Maruru Duha, where these forms alternate with the also presumably imported Hindi type mero and the Western wharau; the Bisaldev Raso, which is claimed to be the earliest Braj text, but is in fact written in a Rajasthani dialect 1 (as is proved by many other linguistic features, such as the presence of -s- futures), contains the o form i, 78, side by side with the ordinary Rajasthani tharo, etc. As usual the greatest power of extension in the early period is that of the Hindi forms, which have also penetrated into works in Rajasthani (OWR. texts, Dhola Marara Duha, and modern Mewati). In the sixteenth century Awadhi of Jayasi, the Hindi forms are still competing with the Braj mora and tora. The Eastern type of possessive is based, as shown, on a Western mahara, tuhara, which is found in Western Apabhramsa and to some extent in Digambara Apabhramsa, and it has been altered by the analogy of the genitive. The 3rd person possessive adjective which occurs in the Dohakosas, tahara, is to be explained on a similar basis, as it is influenced by the genitive ta taha. This development is exactly parallel to what is found at a slightly later stage in Gujarati, where maharau, etc., become maharau, and the change of length is due to the influence of the oblique ma. It would appear therefore that the modern possessive adjectives in -- are not based on a variety of postpositions, but all derive from the Middle Indo-Aryan kera; and the modern differences of form are due to the influence of the genitive of the personal pronouns which took place earliest in the Hindi type of possessive. The possessive adjectives as given by the Grammarians of Apabhramsa raise among other questions that of the use of the plural for the singular. This occurs already in one of the examples given by Hemacandra and it is categorically stated by the earliest of the "Eastern" grammarians that tumbhara and ambhara correspond to teadiya and madiya (Pu. v, 30). This is not simply a mistake arising from the use of the plural of respect; there is a phonetic reason for the use of amha, etc., in the singular. As pointed out by Pischel in his grammar long ago, and substantiated by the examples found by Alsdorf in the Vasudevahindi (there are further examples in the Dhirendra Varms, La langue Braj, Paris, 1035. -20 136 THE POSSESSIVE ADJECTIVES OF LATE PRAKRIT more popular Prakrit texts in Jain Maharastri, e.g. the Dhartakhyana), the verb to be, asmi sometimes tended to replace the personal pronoun, with the result that the form of the first person looked very much like a plural: asmi I am > amhi, amhe <*asme=we. A confusion arises in the second person also where the abl. sg. tusmat gave tumham and the nom. pl. was tumhe. This leads to the use of forms like tumha, tumbha in the genitive which are found in Ramatarkavagisa (the chapter on pronouns in Purusottamadeva himself has lacunae). So the possessives ambhara and tubhara have their explanation clearly in that old confusion, which was not generalized in Apabhramsa, as it was not conducive to the clarity of a literary language. It must have lingered in the popular speech of the Apabhiramsa period, and some of the modern uses of the plural for the singular may have their roots in this early confusion. The forms given by some of the later Eastern grammarians no longer reflect Apabhramsa at all, but are versions of the possessives used in the vernaculars. Ramatarkavaglia has meram, teram, and moharam and toharam, clearly the early vernacular Eastern forms, and in one of his stanzas (v. 7) he has the curious example meradu, the du being an affix (cf. Grierson, IA. 51). As often the case the evidence of the grammarians is again not in accordance with the texts; meradu is not found in any of the Apa. texts hitherto published. -21

Page #21
NOTES ON THE HISTORY OF THE INFINITIVE IN MIDDLE INDO-ARYAN BY L. A. SCHWARZSCHULD, M.A. (Oxon), Melbourne The major work on this subject, as on so many other problems of Indian philology, has been done by Professor S. K, CHATTERJI in his monumental "Origin and Development of the Bengali Language". This article is an attempt to elucidate some minor details. As pointed out by A. MULLIT, no precise general definition of any part of speech is possible, and the infinitive is particularly varied. It is closely allied to the nouns of action on the one hand, and to the gerund or absolutive on the other, and in some ways also to the past and future participles. The history of the infinitive is therefore complicated by constant intermingling and borrowing from these allied parts of speech. The gerund and infinitive in particular are often interchangeable. The basic distinction between the infinitive and the noun of action is generally held to be twofold: (1) syntactically the noun of action has a nominal, the infinitive a verbal function; (2) morphologically the noun of action is a normal noun with a full declension while the infinitive is more closely linked with the verbal system and does not decline. The development of these parts of speech on the whole is according to the following pattern: L. A. SCHWARZSCHILD In the Prakrits the forms are again more varied. The most usual ending of the infinitive in Ardha-Magadhi is ttae. There has been some argument as to the origin of this form. A. Wo derives it from the Vedic absolutives in t ye, E MULLA from the Vedie Infinitives in ove, and he is followed in this by G. V. TAGARES R. PISCH suggests that it was derived from the Vedie infinitives in tavci, which are doubly accented and might therefore have a curious consonantal developmentS. Ser agrees partially with War and suggests Vedie tydyai. The infinitive in tae cannot be separated from that in de which is formed by some verbs ending in a long din Ardha-Magadhi. The endings ttae and ye seem to have been kept clearly distinct from the absolutives in dya, de, and this in itself is sufficient to discredit the theory of Weber. Thus Ardha-Magadhi has the absolutives samdyde, aanlehde, samuthde (Ayarangasutta); and dyde still occurs in the popular Jain Maharistel of the Vasudevahindi. The infinitive however always has a short in the ending and we find for instance in the Nayadhammakahia 28 : a kappal..bhottee i payae od="it is not meet to eat or drink." The dative of the noun of action in a was kept apart in a similar way: the infinitive has a short -- while the dative of the noun of action ends in e, as in the frequent formula: paharettha gomende (eg, Bhagavatisatra XV 43) "he set out to go". The brevity of the vowel in Ardha-Magadhi proves that the infinitive in ttae is a fixed survival and is to be separated in derivation from the normal dative of nouns and the absolutive. An infinitive in titve exists in Pall and in the Asokan inscriptions where it is widespread except in the North-West, e.g. Girnir chamitave, Dhauli and Jaugada Khamitave ete. It seems most probable that the theory of E. MULL is correct and that the infinitives in trae of Ardha-Magadhi are the direct continuation of the Vedie infinitive in tave. There is no real need c Page #22
HISTORY OF THE INFINITIVE IN MIDDLE INDO-ARYAN 31 to postulate with PISCHEL a derivation from -tapai, which has little support, particularly as Professor EDGERTON has recently explained the Buddhist Hybrid Sanskrit infinitive in tavai not as a derivative of Vedic -tavai, but as a hypersanskritism for tave. The doubling of the -- can however scarcely be due to a simple phonetic evolution; it is probably due to the influence of the absolutive where -to->-t- in Ardha-Magadhi: tod, to sam sta, staram. This doubling of is exactly parallel to what is found in -ttu, -u a development from tum, cf. PISCHE, op. cit. paragraph 576). which was used as an absolutive in Ardha-Magadhi (change C). The tendeney for doubling the -- was further increased by the development of some consonantal roots where the doubling was phonetic, eg, bholetave > bhottae. The loss of the --- in the derivation of tree from tave presents no real difficulty. An intervocalie is quite often dropped in Ardha-Magadhf e.g. jiva > ja asrare > anhana, praurtta > payatta etc. The weakness of the pronunciation of -- intervocalie is further attested by the confusion between ----> -- and wintervocalie in a number of words in the Jain canon, to and this feature is continued later by Jain Maharastri and Maharastri, e.g. in the Lilavaikaha, and by Apabhrapsa. The loss of the -- is thus not unusual and owing to the phonetic weakness of terminational elementsil it became the rule in the ending -tave of the infinitive in Ardha-Magadhi The Sanskrit infinitive in tum > - , -um, is occasionally found in the Ardha-Magadhi of the canon, although it is by no means as frequent as the infinitive in Itae. It seems to be favoured only in some fixed locutions as for instance with payatta. This is illustrated by examples taken from a few paragraphs of the Nayadhammakahko: 77. tam icchamo..parivasium 78. citteum payatra. 87. anupariyatteun payatta 90. vesaggeum payatta. 90. na jujjasi.. jjhium. 99. tam seyam. Ihanavettae. 99. iechami.khanavettae. 99. khanaveur payatte ydi hottha. 130.uttarit payatte.hottha L. A. SCHWARZSCHILD In some cases where the infinitive in tam is used in preference to that in ttae it could very easily be replaced by an ordinary noun e.g. Nayadhammakahao 25 : dalayanti palomam ddum palmam bhot palamam paribhaeum, "they give him as much as he likes to give away, to eat, to drink and to distribute." The only other case in which the infinitive in tam is favoured is in composition, in fixed locutions inherited from Sanskrit, e.g. Nayadhammakahao 141: ghakima, palama. Apart from these locutions the infinitive in tum is not very much alive in Ardha-Magadhi, even if one includes cases where it has taken on an absolutival meaning (change C). It becomes slightly more prominent in the later canonical texts and in Jain Maharastri. Here the infinitive in tom sometimes appears strengthened by the addition of the enelitie particle -je, hence the forms gilim-je and gishenje etc. found in the Panhavagaranaim and in popular texts like the Mahnisihasutta and the Maharastri Dhurtakhyana. Dr. Sexs writes: "The accusative infinitive in tum was at best a dialectal feature in M.I.A.". It belonged however to the Midland dialect, Maharlestri, and owing to the great literary importance of the Midland it spread and it has become the normal form of the infinitive in dramatie Magadhi. It is also the usual form in Sauraseni, where it was almost certainly indigenous, and in Jain Sauraseni. It still occurs in Svetambara Apabhramsa, where it is not only confined to passages under Prakrit influence. It is frequent in Sanatkumaracarita, generally with the loss of final-m, e.g. Joahit, taste, tihin, etc., and the Kumarapalapratibodha" has jampiu, hariu, laddhu and kahium-pi, which are the only infinitives in that text apart from one example of an absolutive used as an infinitive. There is no evidence of the infinitive having survived any later in this form. Kramadisvara allows an infinitive in-aim, which is not mentioned by any other grammarian. The most usual form of the infinitive in Digambara Apabhramba however is that in th. It is the most frequent type of infinitive in the works of Puspadanta and in the Karakandacariu of Kanakamara It occurs for the first time in the Paramatmaprakasa of the Digambara Joindu possibly of the sixth century A.D. It has probably survived into modern Marathi," in the infinitive in . The infinitive in-ahu is very rare indeed in 12. S. Sex, A critical introduction to the Past Wurzburg, 1906, p. 13. 13. W. SCHINO, Das Mahinshasutta, Berlin 1918, p. 32 14. A N. UPARK, Dhah , Sinchi Jain Series XIX, Bombay 1941. Inte 9. F. EGSTON, Buddhist Hybrid Grammar, New Haven 1953, paragraph 36/14 10. R. PISCHE, op. cit., paragraph 231. 11. R. L. TU , The phonetle weakness of terminational elements in Indo-Aryan, JRAS 1921. 15. S. Sex, Comparative Grammar of Middle Indo-Ary, Indian Linguistics XII, 16. LADY, Der Kumpileprotibodhe, Hamburg 1998, Introduction, p. 64. 17. J. BLOCK, La formation de la langue Marthe, Paris 1520 .24 -

Page #23
HISTORY OF THE INFINITIVE IN MIDDLE INDO-ARY AN 33 Svetambara Apabhramsa and has left no remnant in modern Gujarati. The formation of the infinitive thus represents one of the major differences between Svetambara and Digambara Apabhramia. The origin of the ending -ahu has not been clearly explained. At first sight it would seem likely that we have here a simple case of the so-called ha-kruti, the -h- being inserted before the old infinitive ending in -um. But PISCHEL warned already long ago: "-h- fallt weder aus, noch wird es zur Vermeidung des Hiatus eingeschoben," and he proves that all cases that had up till then been presumed to be examples of the use of the ha-sruti are really based on false etymologies. Such a categorical denial is now no longer possible. Old Western Rajasthani for instance occasionally shows insertion of -h- e.g., suhanau<*suanauPage #24
206 ANNEE 1056 l'influence les uns sur les autres; nul n'a evolue isolement, ce qui l'aurait conduit hors du systeme pronominal. Le systeme du sanskrit, tres complet, comporte pourtant quelques faiblesses qui ont ete en partie responsables de l'evolution ulterieure : QUELQUES ADVERBES PRONOMINAUX DU ADVERBES BORTIS WANIE MOYEN INDIEN D ATIF.. yadd, garli Lada, kuri PAR INTERNATIF... tru e gatha at itham, Bain ha D L. A. SCHWARZSCHILD SSTATI DE PRIMITE DEMONSTRATIT ADJECTIFS QUANTITE DEALITE yarat Igral tida) yadro lido) DEMONSTRATIF.. ENTE INTERROGATI DENONCRETE PROXIMITE.++ DEMONSTRATIE yat wi Hall Dans plusieurs langoes indo-europeennes de l'epoque moderne, les adverbes pronominaux tendent a sortir du systime pronominal. C'est un exemple du passage continuel du motive a l'arbitraires (Saussure). Il faut dire philologue pour se rendre compte des liens historiques qui rattachent ici et la aux pronoms demonstratifs en francais moderne. Les langues neo-aryennes de l'Inde cependant, bien qu'elles soient pour la plupart tout aussi avancers vers l'etat analytique que le sont les langues romanes, ont pourtant conserve des systemes pronominaus beaucoup plus complets. Ainsi en hindi moderne walvin, yahin, win, alin. Jahru, adverbes de lieu, run. yon, kyon, Lyon, jyon, adverbes de maniere, correspondent aux pronoms (cas oblique) s, is, kis, tis, jis. C'est l'aboutissement d'une evolution millenaire : il y a eu toule une serie de refontes et de changements analogiques en moyen indien, et l'on n'a jamais perdu l'idee d'un systeme complet tel qu'il se trouvait en sanskrit et en indoeuropeen. Le desordre qui nous frappe tant dans les textes du moyen indien, surtout en prakrit et en apabhramsa et moins en pali, n'est qu'apparent: les nombreuses variantes qu'on y trouve representent souvent l'ebauche de systemes nouveaux. Les adverbes ont eu de ADVEHES DE LIED. Les adverbes de lieu sont intimement lies aux adverbes de direction, pour le sens aussi bien que pour la forme. On a donc cree dans la langue parlee "it selon l'analogie d'ilo (tatas: tatra, il : itra); et qui signifiaitici>> comme atre. Il n'y a pas de doute sur l'exisLenee dite qui a laisse nombre de derives en prakrit. On les trouve deja dans les inscriptions du roi Asoka: era (Shahbazgarhi), la (Girnar, Dhauli) held (Khalsi, Dhauli, etc.). CCXLI.3.

Page #25
TELJES ERRES PRIVAUX DE MEN INDIEN 267 hans les inscriptions plus tardives, comme dans tous les dialectes noven indien sauf le sanskrit des textes bouddhiques, il va eu le placement complet de la serie , etc.. qui auraient du donner batter, etc.. selon les lois phonetiques (cl. Asoka la Girnar. Khalsi, etc.). Il a cone deus problemes a resoudre: A. L'origine du tr e dans les adverbes de lieu: B. Lorigine de la voyelle au lieu d' dans les derives d'in. A. On a tente plusieurs explications de ce premier phenomene. 1. R. Pischel croit que tatra, kwire, etc. ont subi l'influence de l'adverbe vedique (cf. aussi . Broch, Laugar marthe, P. 906). II. S. K. Chatterji suggere que le verbe arhiturnit pu exercer une certaine influence et que yah lah seraient peut-etre les formes primitives proposition de avance pour les derives modernes par kaulo, Grammar of the Hindi Language. 8 638). Cette theorie est notee par J. Block, L'Indo-Aryen, p. 91. III. S. GOLDSCHMIDT (Prakritica, p. 9) et aussi W. GEIGER et S. S croient que c'est un changement phonetique et que-rest devenu -tih-dons certains cas. ANNEE 1956 ment entre les deux categories en moyen indien. Les adverbes de maniere kathaw, fathi, yathi sont devenus gulierement en prikrit lenha, ahas, jahon. full (intha) fait exception. Celui-ci, par son developpement phonetique regulier est devenu identique en prikrit aux adverbes de lieu thi thi. La necessite de separer les adverbes de lieu des adverbes de maniere a ete sentie si bien en moyen indien qu'on a etabli une distinction artificielle : itthon est employe comme adverbe de maniere en prikrit, pendant qu'etha, itthaes employo exclusivement comme adverbe de lieu. IL L'explication de M. Chatterji presente une seule dificulte : les mots composes en -tha, skr. the sont des adjectifs et des noms et non pas des adverbes ni des conjonctions. III. L'explication phonetique de M. Geiger et de M. Goldschmidt se heurte a des difficultes serieuses. Pischel croit que les exemples donnes pour le developpement phonetique trt reposent tous sur des erreurs de lecture. L'attitude generale envers de telles erreurs >> a bien change depuis Pischel: on a constate que les erreurs reelent souvent des artistes precieux, ou des irregularites inte ressantes. Tout de meme, il faut admettre avec Pischel que les mots ciles : mali pour wirewalla, mati (Merchakatilvi), puiki pochi et non pas pitrie chez Holo, et d'autres, ne sont pas du toul probants. Cette critique s'applique aussi bien sus exemplis ciles plus tard par Geiger : pali hiyo ireliye mane, mot qui a cerLainement subi l'influence de pali, pkr. otticasi el de achiya w ie (en dernier mot est aussi le nom d'un personnage dans les legendes bouddhiques). Les autres cas cites par Geiger se laissent expliquer d'une facon toute semblable : ce sont des contaminations, ex. pali, pkr. bridi Khadil, sous l'influence de biexelati, hindi moderne bheli pali ploai-arroser, vient de v o ucher, 1. L'explication de Pischel n'est pas tres vraisemblable: il est difficile d'admettre qu'itth. itthon, adverbe de maniere en wadique et en sanskrit soit devenu adverbe de lieu. On semble avoir distingue nette Sulon les les phoniques devrait donner toujours en piliet dans les dialectes prakrits. Mais dans la langue plus moderne, en apab i la pourtant quelques antes de l' enskrit apo ne qui a smile en prakrit. C'est probablement trait Wilderw e rto de Nord-Ouest, qui n's peste wetente que par la forma qu'on trouve che d e stul de mime ne forme artificielle, semble-t-il. R. Peca, Gran der Pralitarache, Strasbourg, 1900, 393. 8. K. CHAPTERI, Origin and Development of the Bengali Language Codes, 1986 1998, p. 85. W. Gues, i Lil l Spruche, Strasbourg. 1911, M 8.5a, Gamprate of Middle Indo-Ary, but I. p. 36, Un tel changement est pourtant possible dans d'autres etapes de l'inderen Ainsi le prati moderan utilise les adverbes de maniere , etc., de l'apabila pour en fin des adverbes de lieu tihen, etc. (ou l'influence d'ila?), la place de hocam adverbes de mare etant prise par les derived INR P L ., 06. Nepali Timary, Londres , 17.

Page #26
QUELQUES ADVERBES PRONOMINAUX DU MOYEN INDIEN 269 non pas de prus, cf. l'emploi dans les textes juridiques du verbe skr. Vapre au sens de se laver (Manavadharmaiistra, etc.), Quant a pkr. phisuya, l'etymologie de ce mot a fait couler beaucoup d'enere. on a propose prasuka, prisuka, qui n'est autre chose que le mot prakrit sanskritise, et l'etymologie la plus generalement recue est apar inka (el. sparsaphasa) ce qui est purs, ce qu'on peut touchers. Il semble done bien que toutes ces aspirees du moyen indien n'ont rien a faire avec une e primitive, elles sont toutes dues a des sibilantes sanskrites, et le changement tr>th ne saurait guere representer une evolation phonetique. Il faut done chercher une autre explication. L'adverbe prakrit et pali hetthaadhas au-dessous, dont l'etymologie est douteuse. pourrait peut-etre nous aider a expliquer les adverbes de lieu. Pischel a demontre que heltha ne vient pas d'adhestha comme le croyait Johansson, mais d'adhestat pour adhastat, selon le simple ahe< adhas. Il y a des cas analogues: puratthapurastat, qui est assez frequent en ardha-magadhi (Nayadhammakahao, Uuarajjhayanasuita, etc.). En sanskrit, il y a paratra dans l'autre mondes, et parastit a l'ave nir, au loins. Ce dernier a donne regulierement parattha en ardhamagadhi et dans les autres dialectes litteraires du moyen indien, et il a peu a peu assimile paratra qui devait donner paralla. En sanskrit la desinence -tit etait deja en train de perdre sa signification primitive, qui etait celle de l'ablatif: elle parait souvent tout simplement comme locatif, parastit signifie done au loins et non pas seulement de loin, il en est de meme pour parastit, bakistat (vedique). De la s'est peu a peu degage l'idee d'une desinence ->ttha pour les adverbes de lieu. Or l'evolution phonetique des adverbes de lieu derives de pronoms aurait occasionne de nombreuses rencontres homonymiques: atraatta (forme tres rare qui ne se trouve que chez Asoka et dans des textes assez artificiels a une epoque plus tardive et dans le compose attabhava dont les parties n'ont pas ete reconnues; ce mot est frequent dans la sauraseni des drames); mais il y a aussi en moyen indien des mots atla derives d'artta, atta et apta, et a l'Est c'est aussi Jonassos, Indogermanische Forschungen, III. p. 418. 1 R. Pisces, op. cit., SS 107. -32 270 ANNEE 1956 un derive d'atman. Tattatatra etait tout semblable a tatla>, illae, ettae, kutra devenait kutta, interrogatif, mais kutta signifiait aussi en prakrit chien (sanskrit kurkura?) et e contrat, mot desi selon Hemacandra. On a donc favorise la desinence -tha-tat qui se trouvait deja dans purattha, parattha; et attha, jattha, tattha, etc., ont pris la place des anciens adverbes de lieu en tre dans la plus grande partie du Nord de l'Inde, comme le montrent les textes. B. Le deuxieme probleme est celui de la voyelle -e- dans l'adverbe etthaitra, qui tendait a remplacer atra comme adverbe de proximite. On trouve dans les inscriptions du roi Asoka les adverbes cites plus haut, etra, heta et etaPage #27
QUELQUES ADVERBES PRONOMINAUX DU MOSEX INDIEN 271 logie est a rejeter : les mots d'origine de lettiya. etc., sont skr. biyat et igal elargis par des sullives. Il est tres probable que c'est l'influence du pronome qui a ete responsable du developpement phonetique ulterieur de ce groupe de mots. Elad est devenu dile-avec redoublement excessif de la consonnet. Etarhi devient donc ethe, takietamil devient lawha de la: il y avait donc une nouvelle base demonstralive et qui serait une certaine influence sur les derives pronominau. Il y avait en prikrit des adverbes ille avec la variante lo que se trouve aussi en pili), kattotete, adverbes de direction derives du sanskritilo, las lutas. Selon les regles phonetiques, tax, etc.. auraient de devenir owwe, kvo, qu'on trouve en marriet en apabhrama. Mais letto, jo, etc., qui representent d'ordinaire cette serie en prikrit ont te refaits sur les pronoms a l'aide da suflive- , -0 kad-108, yaltas, yad u lto, jatto o d'oli par analogie illo, a llo. El parait avoir ete cree sur le meme modele sous l'influence de la nouvelle base pronominale ele. Si l'on a donc favorise ette au lieu d'ithe, adverbe de lieu, cela s'explique par la force du systeme : 272 ANNEE 1956 Sous l'influence d'estha, ittha et d'ello, ille, on a parfois change la voyelle des adverbes de lieu et de direction derives des autres pronoms. On trouve donc parfois en prakrit titte, jille et liha, etc., bien que les formes etymologiques faithe, fatto, etc., soient beaucoup plus fruentes. En apabhrama lologie est plus complete, tattha, kat the, etc., sont devenus plus rares que tetthu, jetthu, kettu, et ce sont probablement ces derniers qui ont donne ethan. johan, tethex, lehen var. thais, etc.) en Bhojpuri , les formes correspondantes en views marathe, le magadhi jenim, le bengali jeti, heibi, etc. Il est difficile de trouver une difference quelconque entre les deux grande dialectes litteraires de l'apabhramsa, celui des Digambora et celui des Svetambara, quant aux adverbes de lieu. Les adverbes de lieu s'emploient souvent dans les deux dialectes comme adverbes de direction. Il y a aussi une nouvelle serie d'adverbes de lieu, fetike, take jettahe. C'est la serie tarki, garbi, karli, darbi; prikrit tale, jahe, kahe. eithe, qui a donne ettahe, et de la fertithe, etc., en apablama tahe. d'abord adverbe de temps, est frequent dans les textes prakrits (mahirijainn) un peu populaires comme adverbe de licu, par fois avec une nuance temporelle. Quelquefois on a interprete la finale - he comme desinence ensuelle, et l'on a substituee a che desinence qui n'etait frequente qu'au feminin), et l'on a ecrit done betalo, qui s'emploie comme adverbe de direction en apabhrama: du reste on trouve deja fattohimo, etc., en prikrit, c'est katto adverbe de direction, avec l'addition de la terminaison double to. Une autre variante a ete creee en apabhrama par l'influence de la serie etale, ele, sur les autres adverbes de lieu et l'on a donc jethali Bhavisayattakaha) et Whale (Harirampurvina). La serie ettahe, etc., a laisse des derives en maithili, tatay tate)jalay, katay, tay et olay, i en hindi occidetal (braj) tila, jita, itai, tai ette. Les formes de la maithiliseren contrent aussi en brajbuli. La repartition des adverbes de lieu est Pour une autre explication, c. 6. m, Blouri Wildaur siya, Paton 1966. p. . 09 Chatterji propose de care le comme etymologie de totale, mais la o la deja l , estantemente en print, et les formes qu'on trouve par pabria, represent plattele, et adres de dire the d ate dans le.. IV, baby Lalela, este de siled Singh Jin Series, Bombay, 19 . En wyen indien les formes citees sont celles de la muharistijaina, mais elles appartiennent aussi aux autres dialectes du prakrit, avec quelques legeres variantes) : DVERES ADJECTIF GENTE e be) ale to Iathiye elbette jestlige tjetliye atthew DEWATI eliye El come priest het toka, il i per la premie fois

Page #28
SOME POEMS OF THE ABSOLUTIVE IN MIDDLE INDO-ARYAN LA SCHWARTHILD QUELQUES ADVERRES PRONOMINAUX DU MOYEN INDIEN 273 done posterieure a l'epoque de l'apabhramsa, mais il faut noter qu'on y trouve les premiers commencements des systemes modernes. Malgre les tendances archaisantes des textes litteraires, on peut donc degager quelques principes d'evolution chronologique en moyen indien: 1" Avant le temps d'Alokaire est cree sous l'influence d'itas; Avant le temps d'Asoka : tendance d'une a devenir era sous l'influence d'etal>eta: 3. La voyelle i de l'ephemere *ira a penetre dans le reste de la serie, donc sitha, etc.; & Un peu plus tard : Inila ate remplace par tha <-il dans les adverbes de lieu: 5* Tendances populaires des premiers siicles ap. J.-C, et qui se retrouvent en apabhrama: a la voyelle e d'atha itre s'est introduite dans le reste de la sorietarla, cic) confusion complete entre les adverbes de lieu et les alverbes de direction: dahelari devient adverbe de lieu 6 En apabhirama, on a croelettahe, etc., selon tale. THE GEREN, O d c ine or of the Andha M , they do not seem to ative participle frosty is the n e restricted to say i t. They are Aryan Tengages in all their stages of develop Ardemlight and Jain Sani faca has gives the modem Hindi back and tricot, the number of examples of this type of pala idential are with the derivatives of Sanskrit alation quoted by Pichel and the based on his work is very limited, and though most There are pe 14. P. 597 CE J. Bech, Pur d e te Tuth, Pula (Caleta, Paris, 1915). 1 Peder Pro RL , Dory of Language . 100. 0 L 10. 30

Page #29
112 SCHWARZSCHILD: Some Forms of the Absolutive in Middle Indo-Argan pratha back which occurs in Apabhrania as pitki (Sanatkumaracarita 695.6). In Sanskrit preku could have the meaning of "surface," and confusion between pitthijkeajajaja jhaya. In its development into the modern languages this word has been combined with danda stick, pole, hence Hindi jhanda etc. Apabhramis jhuni noise," is derived by Hemarandra in his grammar 1. 68 and IV. 439 from dhrani, Jhups occurs only in very late Jaina Prakrit and Apabhrampan texta. Hemacandra's explanation is not inevitably correct: dhani has normally given dhupi in the Prakrits and this may have survived into the modern languages as dhuni (Hindi: Nepali dhuni etc.), although the dental in P. 8. G. M. renders this difficult (Turner, loc. rit. under dhuni). In Apabhraman one finds a verb jkunai which is equated with jugups by the grammarians and generally translated to scold. The phonetic development from jugupe- to jhunis not satisfactory, and it is much more probable that both this verb and the noun jauni are of a similar Desi derivation. Whether Hindi wards of the type jhunjhuna rattle are connected remains very doubtful. In any case words of this whole group, including jhusi, which are perhape onomatopoeie, cannot be adduced as important examples of a sound-change. rijjamce is evidently impossible. Doubts on this subjest are also expressed by F. Edgerton, Buddhist Hybrid Sanskrit fremmar (New Hares, 1950), 36: 39. 38 SCHWARZ CHILD: Some Forms of the Absolutice in Middle Indo-Aryan 113 The change of te to tu by Samprasarana was only to some extent regional; it was also a learned feature, and is found particularly in words borrowed from Sanskrit right up to the modern 2. The most frequent type of absolutive in the period. V tear to hasten, appears already in Southern Asokan with a s, filandiya and fulige, Prakrits ends in (1) ga. It is the normal form found in the Dhauli and Jangads Edicts respecin Maharastri. It is found already in the Bhabru tively; Apabhramia has furai; the Hindi furant edict of Asoka (Hultzsch's reading: abhinddet is associated with this. The second person pronam); it is found in Ardhamagadbl as well as in noun is similar in development to the absolutive: Maharasti, also in Jain Sauraseni se-dina. dans team> tuman. Here the w forms are spread even even occurs in some of the Saurssent passages of further than in the case of the absolutive and are the Sanskrit and Prakrit plays, though of the by no means characteristic of the south alone, grammarians only Hemacandra admits this ending This is due to the influence of the Vedic team for Sauraseni. Thus the Karpuramanjari of Ra and other declensional forms of the second person, jasekhara has ghettina (1.12.2 in Konow's edi-e. g. Sanskrit fubkyam, which have the radical fa tion) side by side with the more usual gephia (IV. The forms developed according to sound change 19. 18 etc.). -tana further occurs in Gatha Pali. (2) were particularly favoured in this case because Professor Gray claims that the ending (1) sound change (3) led to confusion with tags of the represents a contamination of the Vedio -fedna (m) third person, while sound change (4) led to con with the infinitive in tam and that the existence fusion with the derivatives of atman. Although of the suffix -gles helped in this evolution. He there is little doubt that on the whole the change concludes: JS vaftidana Page #30
SCHWARRIO 8ome Morma of the Absolutie in Middle Indo-Aryan SCHWARZCHILD: Bonte Forms of the Absolute in Middle Indo-Aryan 15 Tare most closely associated with Digambon and Man in the Sanders. This seems to Svetimas Apabrami the form is found to be have caused the loss of the labial type of absolutive, moribund. It cours for instance in resin but there is a possibility that it survived for a time in the East Centre and left remontin erly in the Old Western Rajasthani texts Awadhi 16th century Awadhi has absolutives in amised by Tosator, while in the later texts of but also infinitive forms, mainly used in the the same region . the Dhol Mar Doh, oblique, indi (Padumiwa of Jayas: calai, there is no sign of it. There are only the forms inte). Dr. Salon s in us to the origin - und -t. These are from and is now of the verbal noun ini, but they might well despite the views of generally secepted to o back to the Apabramba absolutive with less of the intervale. Apart from these uncertain The Labial type of absolutive had been further enants, the labial type of absolutive, so popular weakened by the time of the Inter Apabhraman in Apabhru , has died out completely. texte by the tendency to drop the intervocali -- - This leads to confusing form such a man for CF. er and Introduction . 16 of the edition by Vijay Munia . Binita the Stoghi Jain 1 stor Notes on the Grammar of the older (Roby, 1943). Western Raja A XLIV (1016). mgraph 131. S e war in the Run of Tulsa, 8. K. Chat Organd Deropesent of the Be a ta Diversity will (1998). paraph , a Longus ( C outta, 1906). pp. 1011 note 2. tilly Batern and that it does not appear in early Easter Inngong like the Ardhamadil of Apabran e the theory of its survival in the cenon: bedoe; ben Paphav, 397) and G Sarathi unlikely (Uttar. 961) dey, JR Prakrit h ubbhe Varurul and the called "Eastern School of Art and the Maharlari of H ipotere Grammarlansive formdan sanding of onion has bra for dra from dana (Hals A. 35. the absolutive in M adh. Hemand does not aloo Hemscand . 79), well its papai, give this form, and it does not appear to exist in which may come from prod i lly the time the acadhi of the Nevertheless such of Aabhad this and become more form might well ha existed in spoken Migadi e rally accepted, and it is found in both Sum and it would represent an extraordinary weaken. brand Digambon Apahami. xmples are ing of the morphological termination: Al the endings of the solutive, the flix - do development parallel to that of of butt touns (tu o ), the form of the the Western ). The final - is no doubt becond per promon, which side by side substituted for the under the influence of the with the Eastern for tal, and has only as Ahsolutive in which is an attributed to s peared in the modern tanpages of the West on dh by the m arins, though it does not become out of its similarity to the derivatives of man general until Aabhram . This ouble absolutive > The Abeolutive suff e nd A probably never Yery widely used and disp u tnam by this soundchange became and Dennd entirely in favour of the simple which is pain. With extraordinary weakening is there the basis of the modern absolutive. epdinge gave , lui changed sometimes 4. The fire indications of the sound change for further to . The short forms -evi and are >pp are found in the team in crisis the most frequentin Apabhrama from the begin Girne Edicts have the absolutive for debito nine, and they appear in Pruit testa influenced (13.) Laseyti (IV.). pricajid X. 4.). by popular speech where the Apaburra endings of the abolutive have penetrated. Absolutives in alecofpd (XIV. 6). Apart from the one finds e i n the found in the Pascaris of either in the text or in the world of the am Vimalasiri, in the M y further derivatives of blatives in iniste text of the Dhar mars nam with the lablal type of similation khya (ex. pindel III. 18), while the LII matha hamna One only finds derivatives of the Vedic H Altheagh -pingid (- (Gr. only), which seem to ha p penedoubtless stages in the evolution to characteristic of the N. w. district is shown by it, they were only intory. They are still the abaolutives in found in the con Edicts found fairly frequently in the early Southern Apabra h i and Manari and in the Inter o Catumukli Tribhuva of Shib Khresthi inscriptions, i Symbu, but in the Inter Ciganland Sve t by soundchange tambar testa forma in (3), which seems to live penetrated to the North po are extremely rare, West). The endings and finem) seem to One berally finds only - - - and in have replaced t o fill in the crossing numbers of solutives in i Farmaco and they coexisted with the absolatives in taining till occur in late text wuch as the A m ersted in the Prabanthacinta which are the farms in Sanseni. mani ( From the evidence of the Prakrite it would 84.40 of the edition in the Singhi Jain DDT s if the labialising tendeney were not con- Series i s p.23, v. 48 fulani, var, lut Still the ha form of the beliative is distinctly fied to the Western districts it will probably Apabra vulgar feature that took a long time to become ly the time of the modem vem established in the literary Innga n d it had its centre of radiation in the West Modem inter A Weber, Desplat de Leipzig borrowing of words has obsessed the situation still Portentative spite of these JRAS 1952 further. One finde example of labilisation in an TR. L. Turner lot, JRR 1 WE Clarke Mal Alam ,"JAOR p ed. A. N y Singhi Jain XLIV (1994), pp. 81 Serm . 100. HR Timer, The pain wakaf termin A. A N. padhye, S hi Jaan Series e 31 Bommailement in IndoAre JRAR 1027. bay, 1949). V. 750.

Page #31
NOTES ON THE DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN BY L. A. SCHWARZSCHILD "THE TENDENCIES OF Apahhrama in phonetics and grammar help to bridge the gap between typical Prakrit and the modern languages." A. Woolner expressed this view in 1928 when comparatively few Apabhramsa texts had been published and the remarkable continuity of the Indo-Aryan languages was accepted as a fact. Since then their general pattern of development, which appeared with simple clarity to Woolner, has often been obscured by the wealth of linguistic material of the Middle Indo-Aryan period that has come to light. There is known to be a number of cases where Apabhramea, as revealed in the texts, does not provide a link between Prakrit and the modern languages. There have been two alternative views on how such inconsistencies are to be explained. Any feature of modern Indo-Aryan, Apabhramsa or Prakrit, that could not be linked with the language immediately preceding it, is regarded by some as a survival from a much earlier phase of IndoAryan, Le. Vedio, or even Indo-Iranian and Indo-European. For instance, the old pronoun of the third person, ana- is already moribund in the Rigveda, where it occurs only in the genitivelocative dual, avos; it does not occur in Sanskrit, Prakrit, or in Southern or Western Apabhramsa and yet it is regarded by some as the basis of the modern remote demonstrative pronoun: Braj seo, wuh, wah; Panjabi oh, uh, etc. This type of explanation dismisses the main literary languages, Sanskrit, the Prakrits, and Apabhramsa as artificial and seeks for the true continuity in the spoken language in as far as it is unattested by the written documents. Carried to extremes this view would make us believe in "a mysterious living language without written texts, as is usually claimed by philologists, without much documentary or philological evidence" a complaint sometimes made in the field of Romance philology. A. Woolner, Introduction to Prakrit, 2nd edition, Lahore, 1928, p. 6. 8. K. Chatterji, Origin and Development of the Bengali Language, Calcutta, 1926, p. 837. J. Bloch, L'Indo-Aryen, Paris, 1934, p. 198. H. F. Muller, & Chronology of Falgar Latin, Halle, 1929. JRAS, OCTOBER 1956. -42 13 182 DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN Some linguists, on the other hand, have laid stress on the connection between the spoken and the written languages throughout the history of Indo-Aryan. Incongruencies are therefore explained as independent creations: thus Baburam Saksena suggests that the pronoun seo, sesh, or seah of Eastern Hindi is not a survival of the Vedic ava-, but a new creation on the analogy of the proximate demonstrative pronoun yah. This type of explanation leads us to see a constant process of recreation in the development of the IndoAryan languages, rather than a continuous evolution. I Some of the inconsistencies between the consecutive phases of Indo-Aryan may be only apparent and this seems to be the case in the declension of feminines in Prakrit. The most striking feature of the declension of feminine nouns in Prakrit is extreme simplicity; the four oblique forms of Sanskrit have been virtually reduced to one. Sanskrit. Pali. malaya Prakrit. malae Feminine Instrumental Dative # malay malayal Ablative, Genitive malayah Locative malayam 30 Masculine Dative devaya devaya The declension of feminines ending in -i and - is parallel to the - declension. 41 " -43 2 devie There are, however, variants in some of the Prakrits, particularly in the inscriptions; e.g. Ardha-Magadhi, Jain Maharastri, and Maharastri have an ablative maldo, which is borrowed from the masculine ablative, and this is also found occasionally in the Sauraseni of the dramas. Further, Maharastri and Jain Maharastri have the variants -da and di in the instrumental, genitive, and locative and according to the grammarians also in the ablative. Pischel explained the usual Prakrit oblique in -e from the dative in-yai which replaced the other terminations in the spoken language. In support of this theory he quoted the fact that the dative ending -yai is used for the genitive-ablative ending -yah once in the Atharvaveda and frequently in the Brahmanas. There is a similar Baburam Saksena, The Evolution of Aushdi, Allahabad, 1937, p. 180. *R. Pischel, Grammatik der Prakritsprachen, Strassburg, 1900, p. 250.

Page #32
DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN 183 substitution of the dative for the genitive-ablative in the younger Avesta, and one is therefore tempted to think that Prakrit usage continued an older pre-Sanskrit type of inflection as, for instance, in the instrumental plural masculine: Prakrit devehi, Sanskrit devaih, Vedic decebhih. This theory has been generally accepted in works of Prakrit grammar. It leaves the ending of the instrumental unaccounted for, as there is no substitution of -yai for the instrumental in the Brahmanas. It also does not explain the terminations of Pali and Buddhist Hybrid Sanskrit. Geiger, in his discussion of the Pali oblique forms in -dya, supposes on the other hand that the ablative-genitive ending has replaced the instrumental, dative, and partially the locative ending in Pali. Instead of accepting these contradictory theories of substitution of cases it may be possible, as suggested by Professor Alsdorf, to explain both the Pali and the Prakrit endings by regular phonetic changes, which were helped by a certain amount of syntactic liberty. Prakrit Endings. It seems probable that the instrumental forms in -ayd changed to-dyd under the influence of the remaining declensional forms of the feminine noun, both singular and plural, where the long vowel at the end of the base was characteristic. The bulk of the Asokan inscriptions, with some exceptions in the north and north-west, have a final dyd in the instrumental, e.g. pujaya (Girnar). There are also some rare cases of such forms in Buddhist Hybrid Sanskrit, quoted by Professor Edgerton, e.g. Wharyay sardham (Mahavastu, ii, 443, 8). The final long - tended to be shortened as it was preceded by an accented syllable and we there fore also find -dya in the Asokan inscriptions, e.g. isaya (Dhauli and Jaugada inscriptions). This instrumental in -dya has survived as the usual form in Pali, it occurs in Buddhist Hybrid Sanskrit and in popular Jain Maharastri (e.g. in the Vasudevahindi). According to Pischel the final -dya can only develop to -da, which is in fact found as a variant in Maharastri. In the Prakrita ya is sometimes replaced by -i- by samprasarana and this accounts for the other Maharastri variant, mali. 1 W. Geiger, Pali Literatur und Sprache, Strassburg, 1916. L. Alsdorf, "The Vanaderahindi, a specimen of archaic Jain Maharasti," BSOS., 1935. F. Edgerton, Buddhist Hybrid Sanskrit Grammar, New Haven, 1963, p. 64. *For a further possible contraction of the final to -a, ef. A. Meillet, "Le datif vedique artrate," BSL, xxi. 184 DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN The syllable ya may also change to ye or e in Prakrit. Examples for this are early: ye occurs instead of you e is one of the developments common to the north-west and to the east of India. It was adopted, on the whole, by the literary Prakrits in the case of the oblique endings of the feminine, but the language of the Midland tended to keep the old forms (Maharastri-da, popular Jain Maharastri -aya). This point of similarity between the north-west and the east of India is by no means isolated in Middle Indo-Aryan: a parallel case, for instance, is the use of the masculine nominative singular ending of ArdhaMagadhi and Magadhi in the north-west and sometimes in Niya Prakrit. On account of the sound-change yo> ye we therefore find -ye as the instrumental ending of feminine nouns already in the Asokan inscriptions of the north and north-west and it recurs in a later period in inscriptions from other regions, e.g. the Mathura Jain inscription of the second century B.c., bhayaye. With the loss of the -y-of-ye there developed the ending that is usual in Prakrit, e.g. malde. This same sound-change of ya> ye> e accounts also for the dative singular of masculine nouns: Sanskrit devaya, Prakrit devie. This explanation is supported by the existence of the variants in -da found in Maharastri and also in verses in Magadhi and -aya in Ardhamagadhi verses and in popular Jain Maharastri. The dative was extremely rare in Middle Indo-Aryan and was therefore open to the influence of other more frequent oblique cases, This, as shown by Geiger, helps to explain the Pali form. The ending M. A. Mehendale, Historical Grammar of Inscriptional Prakrits, Poona, 1948. T. Barrow, The Language of the Kharostki Documents from Chinese Turkestan, Cambridge, 1937. Sir G. A. Grierson explains similarities between the modern languages of the East and West by his theory of Inner and Outer languages, ef. his work " On the Modern Indo-Aryan Vernaculars". Indian Antiquary, supplement, 1931. The reason for the similarity in this case lies, however, more probably in the religious contacts between the two regions. -45

Page #33
186 DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN DECLENSION OF PEMININE SOUNS IN MIDDLE INDO-ARYAN 185 yui could only give-ye in Pali andys - in Prakerity is the role in the Asokan inscriptions (except for thiya in the Shahbazgarhi Edicts). As the final -Gou of the other oblique cases became-dyeby phonetic changes, these cases tended to be confused with the dative and one therefore finds ya used for the dative in Prakrit: ee, the Nanaghit inscription of Western India as Nayanikiya and the Bhaj inscription has the datives jaydyd, Bhadharyd. The ablative genitive and the locative endings are like the instrumental, derived from the corresponding Sanskrit ending by regular phonetic changes. The genitive and ablative forms in-flik lost the final visarga which rendered them identical with the of the instrumental - is found occasionally in the Prakrit inscriptions, thus the Bhart inscription of Central India has Pulid (ablative) and Ndgedy (genitive), and with the commal shortening of the final-d we also find-dya. The ending the develops in exactly the same way as that of the instrumental and this nccounts for the inscriptional ye, the standard Prakrite, and the Mail variant-da. By the rule of shortening of the final vowel after an accented syllable the ending of the locative gave-ym, which is found in the inscriptions of Asoka except in the north, north-west, and east; thus the Gimar inscription has ganandyan. This ending survives into later Prakrit inscriptions and into popular Jain Maharlas, for instance, in the Vasudevahindi. Finalmusvira is extremely weak and tends to disappear, is the case for instance, with the final nasal of the genitive plural ending in Maharastri. There is also a tendency for any ra to disappear with lengthening of the preceding vowel; examples of this have been noted in popular works in Jain Mabinistri, in the Paumacary of Vimalasari and the Vasudevahindi. The loss of the nasal made the locative identical with the other oblique cases and its development was the same. It is thus evident that the confusion between the cases of the oblique of feminine nouns was due to phonetic causes in the literary Prakrits there was no substitution of cases and in the inscriptions and in Pali and Buddhist Hybrid Sanskrit and popular Jain Mahirit such a substitution took place only in the dative Phonetic change alone cannot account for the inconsistencies between Prakrit and Apabhramla in the declension of feminine nouns. In Apabhrama the ending is found only in the instrumental, generally with shortening of the preceding vowel (mile). while the dative, genitive and ablative, and locative have a new ending-headlake) and more rarely-hi-hi. It has been shown by Professor Alsdorf and by Dr. H. C. Bhayani that he was the original ending of all the oblique cases of the feminine except the instrumental, and that the variants in hin, which occur occasionally, are a late and secondary feature in Apabhrama There are a number of explanations for this ending-he. A. (i) Pinchel thought that he came from yil, which is found in the genitive ablative singular of the feminine of the pronominal declension in Sanskrit and which has survived in the Pali tassi Page #34
DECLENSION OF TEMININE NOUNS IN MIDDLE INDO-ARYAN 187 borrowed only in those cases where the masculine had an -- ending. This excludes the locative, where the tastal termination of the masculine in Apabhrasa was a simple-t. It is also significant that the oblique of the feminine of the demonstrative pronoun in Apabhrama is always take, tahe, while the masculine is still usually lana, law, etc. It is therefore very probable that Pischel's method of derivation is correct here and that the pronominal declension provides the link between Sanskrit and Prakrit and Apabhrama. The declensional systems of nouns and pronouns have constantly influenced each other and this is evident already from the pronouns in Prakrit. Declension of the feminine of the demonstrative pronoun sa :Sanskrit. Pali. Prakrit. Instrumental. Bagi ya M. tie, ta, AMgtie, toe, 8. Ide Dative tasyai fassa or tissa Ablative tasyik tays AMg also to Genitive - tassi ortid M. tiesa, tie, tia, AMG tise, Ilie, tie, Stae Locative tasyam tasys or layam, but also M.AMg isa and TM the Here the link with the Apabhramsa declension is clear the instrumental is the only case of the feminine noun in Apabhramsa that does not have an -- ending; it is also the only case of the feminine pronoun in Prakrit and Sanskrit that does not take an -- ending. Of the Prakrit forms de represents the phonetic development of the instrumental layi> tayo tayaliyet > Ide and of the other oblique cases when they were influenced by nominal declension (e.g. ablative ag, layih > Pali taya and Prakritte). There was in Prakrit also a new stem of the pronoun tf., which was based on the frequent final-i of feminine nouns and, in particular, the suffix ki. This new stem seems to have been usual mainly in the Midland, it never penetrated into Sauraseni and did not oust the original stem completely in the Eastern dialects. The instrumental of ti, tiyer gave tia in Maharastri and with the change of ye to ye, it gave tic. 188 DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN In Maharastrt this new stem has a genitive-ablative and locative tissa < iyak and tisyim ( taayih and tasyim). In the regions where a after y sometimes changed to e there was also a tendency for all the ending to be confused and isyai, isyah, and Lleyin gave tise in Ardha-Magndhi. This form is interesting is it shows the intermediate stage of the sound-change sy> :> (as a parallel, cf. the rare form of the future desimo in the Ayirangautta). The change of ha in the genitive singular of masculine nouns in Apabhramsa) and in words frequently used in an unstressed position such as numerals. The oblique tise shows that this sound change also tended to affect the feminine of pronouns. An even more advanced form of the sound change is found in the Ardha-Magadhi, Maharlistri, and Jain Maharastri "locative" dhe, derived from *tase, which is parallel to tise. One would expect such a form the, if it penetrated into the literary language at all, to be used as a general oblique (except instrumental) like tise; but this is not the case. The Page #35
DECLENSION OF FEMININE NOUNS IN MIDDLE INDO-ARYAN 189 iii, 65. The more popular texts in Prakrit, such as the Paumacariya of Vimalasuri, the Dhurtakhyana and the Lalavaikaha all adhere to the standard practice of literary Prakrit in using the as an adverb and the declension of the oblique of nouns and pronouns remains unaffected. There is therefore no evidence of a direct link between this adverbial form and the new ending -he of Apabhramea. The influence of adverbs on declension is widespread, as can be seen from the ablatives in o of Prakrit based on -tas, found in tatas itas, etc.; or the use of tatra where, as the locative singular of the masculine pronoun in Niya Prakrit. It still remains unlikely, however, that a form as definitely temporal in meaning as the should have influenced the whole of the feminine declension in Apabhramsa. There is, however, an indirect link between the Prakrit tahe and the Apabhramsa ending-he. The forms of the pronoun based on ti-, although so frequent in Maharastri have not been continued in either Digambara Apabhramea or Svetambara Apabhramda, both of which use only the t4-stem, as had been the case in Sauraseni. Tie occurs as an instrumental in Apabhramsa, but only in comparatively few instances. The presence of tie on rare occasions in Digambara Apabhraipfa can be explained by the influence of the literary language of the Midland, and in Svetambara Apabhrama by the influence of the language of the Jain canon. The regular form of the instrumental in both Apabhramaa dialects is the. The demonstrative pronoun feminine, like the interrogative and the relative, was now treated as an -d stem. The place of tise is therefore occupied by thePage #36
304 SCHWARESCHILD: Notes on Middle Indo-Aryan Words in 11 forward to move is given in the Dhitridess ebly because of a misinterpretation of Hemacandrs of the very late grammarians Tamalaman Tarka. (VIII. 39 of the Desinimamila): vagli and Markandey, and is close enough in this sulla a to throw mi meaning to k to direct to given as equivalent by Hemand, who also uses famo the word with the latter meaning in the Kum the word solle was wrongly understood to mean tapilacarits (VI. 97)Bolla (1) and (9) On mest and was derived from ala. This mean therefore easily be regarded as one verb. ing is unknown to Prakrit and Apabhram texts Attempts have been made to derive the remain- and one on therefore dismiss from discossion ing the words solla from da spit, vin R a > olla (3) wit is "ghost word." olla, cooked on a spit. A variant explanation There remain then an adjective olla 'cooked presuppos & Bankrit original Saule cooked on and verb soil. "to cook, which from the Prskrit a spit. There are no phonetie difficulties in the evidence ooted above are not derived from fala explanations, but the development of the meaningspit. As already sented by Piachel these two is very doubtful. We find for instance in the Vive words are connected with the Sanskrit Vad to suya, 11th Ang of the Swetambar Jain canon prepare, to cook. According to Pincel the de a description of the wrongful deed of eating e v elopment of the Prekrit solla from Vedisan Here the verb solli : Vivigasuya 13 hab. instance of the tendency ford to change to 1 vind jalayanathalmate hearing andame parallel J. M. piliei prad e te fan ya korallien ya bhajjanda ingin te explains the adjectival form solle fromadna, talenti, hajjanti, sollen . They bent, frysadk for dia. The comparative rarity of the change of a to s well as the doubline of the make it in saucepans, cooking renses, frying pans, and on unlikely that we are here in the presence of a cool the of many mother creatures that roam straightforward sound change, and it appears that in the water, on land and in the sky. Here the the formation is a lot to some extent analogical mening of 'rosting o spit' is out of the ques There esist in Prakrit certain amber of tion, because of the other utensils that are men and adjectives that contain - where Santioned and because it would be imposible anyway skrithdr These forms are no doubt based on with Pischel quotes in gala-kolligo in O - earlier has been shown by Professor Vis 24cooked on coal, which according to him Turner! Thus Sanskrit Bhadre corresponds to a csendes the idea of a spit. From the way J.Mbhalla, pudra to culla, chine to chille (also Volis sed, and from the evidence of the verbal-chir u chilla) Allied to theme is J. M. parallel with it in these description, it is clear challikin, hide from chad to cover. Simi that soll simply means to cook. It is very oftentlarly Sanscrit dans moist, aber in Am sed of cooking meat, as for instance in the Vir J.M. and an alla. The same Prakrit dinlects Rays L.2. where there is a description of the also have an adjective olla, mening moist. This cooking of sundry cuts of meat. Also for instance is however not based on andre, but on udna, dla in the Niraya aliyo I: Ban Nyara (hypothetical adjectives), which are derived from alimemshin sollshi wa alikiye bajjiaki... bade ud'to moisten we becomes olla with the dahalom in the a ge their loi by mal development of too before a double con sent alla means of the Besh... of King Srenika, cooked confused with alla because of the hosted and fried. Because of the Association of identity of meaning, and this socounts for the kolla "cooked with o r t, and also prob. Prakrit grammation Associating both words with Pabel, on at *GAGrerom, The Prakrit Dh a Menon R Turner, Dictionary of the Nepali Language of the Asiati Binty of B el vol. VII, Cal London, 1931), Bhalo, P471.Pichel however uta, 1904), P. 77.170. show that the change wasir o p, et pareg. By A Hoornle, e giado Calcutta, grap 204. 18 , glossary, V alo by P. Valdys, Oudes Some of the words were later borrowed into dando (Pons, 1930), s . Sanskrit: alle and chaiN are found in the works of R. Piachel, Grammatik der Prakritsprache (SET Lexicographers, though their popular origin was recog burg, 100). ITL tiada ty Visa) NOTES ON SOME MIDDLE INDO-ARYAN WORDS IN LL LA BORWARD In The Rail Language Professor Burrow with the addition of Though uncogniad has tently given a clear exposition of the so by the grammars, the bond between nominal cepted view that some verbal roots in Sanskrit Coun d vera m till part of the living - sin suffixes that also coeur in nouns: "These ele guage in the Middle Indo-Aryan period, and in mentare carily recognisable, either through the attempt is made here to prove this by an examin coexistence of impler form of the root, or by the tion of some Prakrit and Apabrams develop existence of synonymous roots which differ only in ments. the final element. The identity of these elements One of the Middle Indo-Aryan words that have with the waite m ented in dealing with the need some dimension is rolla. It appear in the formation of non in evident, and it is in on Pindama s tive different words dance with the fact that nominal il verbal stems (1) verb to throw principles. are formed fundamentally on the (3) verb to drive forward The Sanskrit m a n , laying all the stress on (8) nom at the derivation of nouns from verbal roots, lost (6) & verb 'to cook aight of this fundamental similarity between Down and verb; the only verba they (5) an adjective- ally recognised as cooked being ocisted in formation with nouns were the Of these five wurds (1) and (9) obviously repre denominative verbs, obviously formed from nouns went one and the same verb: the meaning to drive T. Harrow, The Monkrit Lange T on, 1916), D. 1. Seth, Puladd (Caletta 2012 1985), sv. wolle

Page #37
SCHWARZSCHILD: Notes on some Middle Indo-Aryan Words in -11 andra. There was thus one case of a root in -d, Vud (with a number of derivatives uda (ka) etc.) forming a new adjective, ella. On this was based a new verb, ollei 'to moisten," whose past participle passive, olliga, is frequent in J. M. and M. It was probably on this pattern that Vaud (also a verb with numerous derivatives, nuda etc.) to push formed a new verb nollei 'to push. Similarly the adjective nolla and the verb sollei were formed from Vaid, "to settle, to prepare, to cook' (deriva tives of Vand being sada etc.). This derivation explains the fact that there are such similar adjective and verb forms, and it also explains the double meaning of the Prakrit verb 'to direct (olla (1) and (2)) and to cook,' and the words zolla (1-5) are thus accounted for.. As ollei followed olla, "chillei "to cut, to tear followed chilla 'torn, a hole,' and we find in J. M. a past participle of this verb, chillia torn. There are a number of parallel formations in Prakrit: on the pattern Vchid chillei, Vbhid to split formed "bhillei, with the past participle bhillia; as Vchad 'to cover formed the derivative challi 'a skin,' vi-krdd 'to make a noise' formed rihallai, and ni-kad to cut up' nicchallai. The group of verbe in -- which was already swollen in Prakrit by the appearance of all the forms cited above-ultimately connected with dladjectives and nouns-was further increased from some quite different sources. Thus Sanskrit preryate gives pellai 'to send forward, to throw in J. M. etc., with the sound-change -ry-> that is found for instance in pallanka bhusai) to fall,' and Apabhramda bhullai "to fall. It may well be a derivation of the type discussed here, and one argument in favor of this is the existence of an Apabhramia adjective bhulla fallen,' which is the same in its relation to bhullai as solla is to sollei, ella to ollei ete. Uncertain also is the derivation of Apabhramda rellai 'to flood,' relli a flood. This is probably formed from Vri, Prakrit riyai by analogy with the numerous group chid- etc. discussed above. There are one or two tempting etymologies of equal uncertainty, among these one must note especially the verb vellai. This is recorded with several meanings in Prakrit: (1) vellai 'to shake (2) sellai'to play," 'to amuse oneself' (3) compound verbs wevellai, sevillai' to shake, to move forward,' but also-udvestate, just as samcellai is used as equivalent of samvestate (note also nievellai, with similar meaning, quoted by Pischel). There are two main theories to explain all those forms: (a) from a Sanskrit root vell, found in epic and classical Sanskrit, and (b) Pischel's explanation which derives this group of words, as well as the nouns velli 'a climbing plant, vella hair, distraction, amusement," from vina, past participle of a verb sil-, or from rilesti, vilyati, based on the same root. It seems much more likely that here we have an example of the type of really widely extended analogy discussed above for the case of the Apa bhramia mellai, rellai. It is comparatively early, because the words of the group sell- are found in Maharistri and in the Prakrita of the drama, not however in the earlier language of the Jain canon. They have thence been borrowed into Sanskrit which accounts for the epic and classical verb. vellati This Prakritic verb sell- belongs to a group of verbs that have been extended by means of suffixes already in Sanskrit: vij 'to be agitated" and tip 'to tremble," vyath to be unsteady." Which form of the verbs of this group was first of all responsible for bringing in the influence of the -- verba remains uncertain. Vij and vip, sep- seem to have fallen together in J.M. and presumably also in the Prakrit of other regions (perhaps excluding Sauraseni), as the normal derivative of both in J. M. is via-, vea. This form in probably the basis of the analogy. The change of rea-via- to vell- would thus appear to be parallel. to the change of riyai to rellas noted above," The example of these -- verbe shows how the verbal stem went hand in hand with the nominal stems: we find the maximum number of -- verbe in Apahhramia, precisely when the illa, ulla, alla type of nominal and adjectival suffix was at its maximum extension. It is not a question of cause and effect, as only some of the -- verbs are directly derived from nouns or adjectives. It is simply that the suffixes in -, both nominal and verbal, were popular in the spoken language particularly after the beginning of the Christian era, and were felt to be characteristic of Prakrit and For references to the earliest occurrences of this verb in Sanskrit ef. Bahtlingk and Roth's St. Petersburg Dictionary, a. v. sell Some of the other words in this group, velli 'a creeper and erilite curly occuring in Pall, Buddhist Hybrid Sanskrit and Sanskrit may represent a contamination between the Prakritie root rell- and Sanskrit valli 'creeper.' 55

Page #38
S WANSCHILD: Notes on some Middle Indo-Aryan Words in Il 307 even more so of Apabhrams. They were there bara Apabhraman alike. This characteristic Inter for extended more and more by analogical form- berame checured be further phonete development tion especially in the use of roots who simple but the yerbe in have still left their mark on form was phonetically extremely weak. This ap plies for instance in tid, me the modern language on account of the many and number of the forms cited above. The verbs of this group that have survived, forming for frequency of all the formation in -12- is one of instance many of the verbe in od in modern the striking features of Digambar and Svetim. Hindi QUELQUES ADVERBES PRONOMINAUX DU MOYEN INDIEN L. A. SCHWARZSCHILD II. LES ADVERBES DE TEMPS Rien de plus simple et de mieux ordonne, semble-t-il, que les adverbes de temps jab, tab, lab quando, tralors, quand?s, de l'hindi moderne; pourtant l'evolution de ces mots est assez compliquee. Comme on l'a vu ailleurs, le moyen indien est beaucoup moins clair que le neo-aryen : il y a la une diversite de formes et d'usages qui s explique en partie par la force du systeme pronominal, et par le jeu des influences analogiques entre les adverbes de temps et les autres derives pronominaux : adverbes de quantite - deja usites en sanskrit pour exprimer l'idee de duree - adrerbes de direction, etc. Malgre toutes ces possibilites d'explication, il reste des problemes. Nous allons essayer ci-dessous d'en resoudre quelques-uns. * JA, 1956, 3, p. 165. - 57 -

Page #39
242 ANNEE 1957 Les derives des adverbes temporels du sanskrit se presentent en moyen indien sous la forme suivante : SURASENT SANSKRIT tada yela kada idantu AMG, M. JAINA te, tay jad, ju Lada, Laya igi tea jada kada MAHARASTAI aly, teily) jmy, jaily Tay, y) ehi 58 APABRAMS tai, taiyahum jai, jaiyahu kaiyah rekin, eltie Le vocalisme des variantes y, etc., de la maharastri (dont il y a meme quelques exemples isoles dans les dialectes jainas) est assez surprenant. Ces variantes sont tres frequentes dans les textes quelque peu populaires, dans la poesie narrative (e. g. Lalaraikaha) et dans la maharastri des gatha lyriques, aussi sont-elles bien connues des grammairiens prakrits. Vararuci, qui decrit surtout la langue des gatha lyriques dans la partie ancienne et authentique de son oeuvre, comme l'a montre M Nitti-Dolei, dit dans sa regle Prakrtaprakasa. VI.8: ahe ia kale. Il permet done tai, jota et krii comme adverbes de temps. Cette regle a ete retenue par les grammairiens plus tardifs, e. g. Hemacandra (III. 65 de sa grammaire prakrite, Siddhahemacandram, adhyaya VIII) dit: ier dahe dala ia kale. Il est done d'accord avec Vararuci. Les adverbes de temps taily), etc., ne s'expliquent pas trop bien par les regles phonetiques: tada, taly), etc., representent l'evolution reguliere de tada, etc., en maharastri aussi bien que dans les autres dialectes du moyen indien. Pourtant les mots taily), etc., ne sont pas isoles quant a leur vocalisme. Il y a d'autres adverbes de temps qui n'ont rien a faire avec le systeme pronominal proprement dit, et qui se terminent par da en sanskrit: ekada une fois, sarrada <>. Tai, jai et kai alors, quand>> et <> paraissent dans les textes longtemps avant jai <>, <>> (qui vient de kada cit parfois, et peut-etre de kada pi a quelque temps que ce soit, mais non de kada ca). M. C. Modi ne dit rien du cote semantique de son etymologie. L'evolution du sens: parfois, a quelque temps que ce soit quand?>> n'est pas evidente, et ces considerations achevent de discrediter la theorie de M. Modi. 3 R. Pischel croit que taily)a, jaily)a et kaily) alors >>, << quand et quand ? viennent de "tayada, "yayada, "kayada (meme sens) par 59

Page #40
244 ANNEE 1957 voie de "tayida, "yayida et "kayida. Les adverbes vediques kaya comment et ayd ainsi auraient servi de modele a cette formation. Tayada tayida et ses congeneres seraient done semblables aux adjectifs pronominaux de qualite pour lesquels Pischel propose une base kaya+dri, etc. Il faut retenir un fait important de l'explication de Pischel: il distingue nettement entre le cas de sai sada toujours et celui du groupe de mots taily) alors, etc. L'evolution sai sai, Pischel croit que c'est la un changement sporadique d'a post-tonique en -i. Les preuves qu'un tel changement ait vraiment eu lieu manquent en moyen indien. On pourrait a la rigueur y voir un exemple de samprasarana 10, mais c'est plus probablement une simple assimilation: yd> ae, lequel s'est alors differencie en di, trait caracteristique de la maharastri qui se retrouve surtout dans la declinaison des noms feminins en -d: maharastri malaimalayaPage #41
246 ANNEE 1957 krit, et ce qui restait en moyen indien des nutres parties de l'ancienne flexion du pronom demonstratif de l'objet rapproche, e. le locatif pluriel es moharistri jaina). esim (mabarasti), l'instrumental du singulier (mubaras) ne ressemblait que trop aux formes correspoudantes de l'ancienne flexion detail : loc. pl. cete, instr.sg. masc. ce mahnisi). Tout isolees qu'elles etaient du reste de la flexion de ce pronom, i.e. ) ees formes e, , ete, etaient bien aples a donner l'idee d'un pronom eles derives d'ad etaient portes a faire la meme impression. Il y avait en outre l'adverbe tres frequent sanskrit mom > prakrit ) ainsi, qui laissait deviner un meme theme pronominal probablement par l'analyse inconsciente de la part des parlants: theme to jane theme ja, ... ) theme Quelques-uns des mots qu'on essayait d'eclaircir jusqu'ici par le vediquenya (selon Pischol) 10 ou par la syncope de la deuxieme syllabe du mot raw) ainsi - selon loerulo) 07 s'espliquent avec plus de vraisemblance comme des derives de renouveau pronom Plusieurs d'entre ces mots paraissent deja dans les parties les plus archiques da canon jain : vilaya. fem. m ye d'une telle ifrandeur, www si fins, mahliyo si richos, c si grand>> (-radila est ramene a sensorit t he ce qui a ero, ce qui est grand dans tous les livres qui font autorite, e. g. Boca, 'Tudo-dryer. p. 04). Une forme primitive en syi-semble impossible pour ces mots qui ont l'air tres moderne. Le deuxieme element est evidemment de for mation prakritique dans le cas d'erdele et d'auhalya (sullise prakritique), re qui rend tres probable que le premier element est lui aussi d'origine moyen indienne et que c'est donc le theme pronominalequi s'est si largement etendu en moyen indien, plutot que l'ancien ayi-, moribond des le ligveda. Les adjertils pronominaux teada, jeres grands combien grand relatif) dont se sert Hemacandra, www.palarie VIII, 56, "lo-rallhaye combien grand qui devait donner le marathe lovelt, de meme qu'une foule d'autres qu'on trouve deja en maharas!jnina : kemahilaya de quelle grandeurs, forwahihiya combien riches, etc., ont ete crees sous l'influence analogique des mots rorrespondants composes a l'aide du theme La theorie de l'ischel sur pkl. crime de tel aspects, el kerise de QUELQUES ADVERBES PRONOMINAUX DU MOYEN INDIEN 247 quel aspects est elle aussi absolument invraisemblable, comme vient de l'indiquer M. Edgerton ( Pischel 13 has a highly implausible explanation of Geiger (101 $ 11 a less implausible one. Au lieu de supposer des formes primitives ayadri, "kadrs comme le fait Pischel, il vaudrait mieux prendre pour base le sanskrit drie de tel aspect, kadree de quel aspects comme l'avait deja fait Geiger. Idrs a donne Tri (qu'on trouve dans la sauraseni des drames), mais dans la plupart des parlers moyen-indiens on emploie des formes en -:edila Asoka: Shabbisgarhi, Mansehra); prakrit de Niya, atrisa; ardhamagadbi, maarastrijaina, maharastri, parfois meme sauraseni eris, keris. Le changement de l' d'idria en e l est probablement du a l'influence de ce meme theme pronominal -, et alors brins a pris la place de krise, etc., sous l'influence analogique du moteris. Par la force du systeme cette serie de mots s'est etendue en moyen indien, et c'est donc la qu'il faut chercher l'origine des adjectifs teddaha, daha de tel aspect, etc., cites par Vararuci IV. 15 (le consonantisme de ces mots reste difficile a expliquer). Faire remonter au vedique des formes d'aspect aussi moderne qu'eddahe semble aussi impossible que dans le cas d'emahalaya, etc., cite plus baut. Il faut probablement ranger sous la meme rubrique des mots comme maiya sfant (ardha-magadhi, qui vient d'iyati yarika * pronominal eterm> avriya), d'ou le compose ardha-mangadhi al-khutto 22) tlant de fois, et kenalyn combiens, peut-etre faudrait-il y ajouter lorecien et keranciram depuis quand , bien que ces derniers soient plus douleur. L'histoire de l'extension de l'edu pronom demonstratif est assez longue : les langues modernes en sont encore temoin: e. g. awadhi moderne eto, jefen, etc., adjectifs de quantite, bengali moderne eta, eto, etc. Cette extension de l'e- s'explique peut-elre par le fait que le theme pronominals de l'objet rapproche, e, appuye qu'il etait par l'adverbe era(w) ainsis s'employait avec un sens beaucoup plus energique que les pronoms qui exprimaient l'eloignement, l'interrogation, etc. En fin de compte il est clair que la theorie vedique de Pischel est a rejeter pour tous les mots qu'on vient de passer en revue. Il serait donc vain de chercher des origines lointaines pour les variantes taiga alors, joyi, elc., de la mabarastri. Comme la - 63 -

Page #42
248 ANNEE 1957 plupart des derives pronominaus elles s'eclaircissent mieux a la lumiere des refections du systeme pronominal en moyen indien. Il semble bien que ce soient les adverbes de nombre kaikati, lai >: taiyahine dans le drapilapratibhoda de Somaprabha. On s'est servi aussi des formes courtes ( janya, etc., en apabhramsa, mais parfois elles ont perdu la voyelle finale sous l'influence de l'analogie de jai >, qu'on trouve des les textes anciens du kasmiri (Lalld-rakyani), et peut etre aussi les adverbes du nepali, kaile on kailhe, taile, etc., mais cette etymologie reste douteuse, car la presence d'une (cf. la forme ancienne takile) donne de l'appui a l'elymologie de Turner 16): taile>; S. R. Chatterji suggire pourtant une autre explication pour ces mots : skr.yadithi, tada + . Les mots marathes faim, kaim, jain alors quand , quands, remontent sans doute a uya, ele. Somme toute les langues neo-aryennes semblent avoir prefere CORLY-3 .65 -

Page #43
2501 ANNEE 1957 d'autres expressions pour exprimer le temps: on se sert largement des derives de tarat, d'abord adverbe de duree temporelle. Parfois ce sont les adverbes de maniere qui ont pris une nuance temporelle: c'est ainsi qu'on peut rendre compte du marathe erham apabhram a emrahimerahim (-him vient de takim, comme dans le cas de taiyaki. vide supra). Ce mot apabhramsa paralt deja chez Svayambhudeva au sens de maintenant. Meme la ou les adverbes de maniere n'ont pas survecu tels quels comme adverbes de temps, ils ont exerce une influence profonde sur le developpement des adverbes de temps derives de tarat, etc. C'est ainsi qu'on explique couramment (2) le phonetisme difficile de tab, etc., de l'hindi moderne, et des adverbes en be du bengali moderne. Ces refections-elles ont toutes pour base les themes pronominaux ont resserre les liens qui unissaient le systeme pronominal: il y reste des debris du systeme maharastri et apabhramsa (e. g. bengali jai), temoins de la continuite de l'evolution indo-aryenne. Mais c'est le systeme meme des derives pronominaux, refait et perfectionne en neo-aryen, qui est le temoin le plus fidele de cette continuite. Note aur taval. On rencontre assez souvent dans les etudes sur le moyen indien et surtout dans les glossaires des textes une theorie selon laquelle les t mols larat, yarat, adverbes de duree temporelle en sanskrit, auraient donne deux series de derives dans les dialectes prakrits: 1" en posi tion accentuee ces mots seraient restes comme jara, tara, mais a en position inaccentuee ils seraient devenus ja, ta. C'est Vararuci qui semble avoir ete le promoteur de cette theorie (Prakrtaprakasa IV.5): Faradadigu vasya il peut y avoir elision de ea dans les mots yarat, etc. Pischel op. cit., $ 425 a deja vu qu'on avait fait fausse route en acceptant sans hesitation une telle hypothese, et pourtant l'erreur persiste. Pischel propose de nouveau une etymologie vedique: ta et ja seraient des derives de tit ainsi et de yat pendant que, d'autant que (cf. aussi l'adverbe sanskrit af alors>>). Bien que tat el yat ne se presentent plus tels quels en sanskrit, ils etaient a la base de nombreux mots composes, et on avait encore le sentiment que c'etaient des ablatifs de pronoms (ya- et ta-). En prakrit on a cree de nouveau des -66 QUELQUES ADVERBES PRONOMINAUX DU MOYEN INDIEN 251 adverbes ja, ta, qui correspondaient aux ablatifs prakrits du type putta, nom. sg. pulla. Ja et ta comme adverbes de l'ablatif ont presque toujours une signification consecutive: le sens ne correspond pas a celui des adverbes vediques que cite Pischel, sens fige qui n'a pu vivre en sanskrit et moins encore en prakrit, et il ne correspond pas non plus au sens des adverbes yasat et tarat. Ta se presente tres souvent au commencement d'un discours dans les drames e.g. Malarikaguimitra III: ta java pamada-vaya-paliane... ayyezami, je vais donc suivre la gardienne du jardin... Dans ces cas-la, ta est souvent suivi de jara, qui donne le sens temporel, et les << chayas>> rendent ta par l'ablatif sanskrit tasmat. Meme la ou ja et ta se presentent seuls, toute nuance temporelle qu'on pourrait y sentir est celle d'un ablatif et non d'un adverbe de duree: ta peut donc signifier apres cela >>. Il ressort du temoignage des textes moyen-indiens que ta et ja ne sont pas des derives de yarat, tacat, adverbes de duree, ni strictement des adverbes vediques yat et tat. Ils representent plutot des formations nouvelles qui ont pour base une forme prakritique de l'ablatif pronominal tat, yat (au lieu du sanskrit tasmat, yasmat) d'apres l'analogie de la declinaison des noms masculins en -a. NOTES 03 L. Nem-Duca, Les Grammairiens prakrite, Paris, 1938. L. ALSDORF, Haritaminpuriya, Hamburg, 1934. Glossaire, s. v. tai. B. K. THAKORE, M. D. Desai and M. C. Moss, Gurjareninirali, Gaekwad's Orien tal Series, No. CXVIII, Baroda, 1956. Glossaire, s. v. kaiya. La metathese n'a joue un role que dans l'evolution de mots tres differents de kaiya, cela ressort de toutes les etudes sur la metathese, depuis J. BRANES, Comparative Grammar of the Arian Languages, 1, Londres, 1875, p. 376, jusqu'a P. THIEME Merkwurdige indische Worten, in KZ, 67, 194s, p. 185. Cf. aussi le temoignage de J. Bioc qui dit En dehors de l'aspiration et des continues, la metathese semble tout a fait rares (L'Indo-argen, p. 167.) R. Pisca, Grammatik der Prakritsprachen, Strasbourg, 1900, $ 113 et 191 Le probleme du changement de semi-voyelle + voyelle en voyelle simple, dit samprasiraga, va etre etudie en detail dans un article de M. N. K. Norman de Cambridge, qui bien voulu me communiquer quelques-unes de ses idees. of JRAS, 1956, p. 183. 00 H. Jacoat, Bharinaattakaha des Dhanapala, Munich, 1918. L'article de M. N. S. Ghoshal, la derniere en date des etudes nombreuses sur ce probleme ne m'a pas encore ete accessible. M. N. S. GuosuAL, Apabhrama and 17. -67

Page #44
Reprinted from Indian Linguistics - Turner Jubilee Volume I, 1958 PRAKRIT THAKKA, TIRED 252 ANNEE 1957 post-apabhrama features in the early prekrits, is wie lateral arterly. XXX S. K. CHATTERI, Origin and the Bengali Language Calcutta, 1986, p.858. Pour les d es de liv /, 1956, 3. W I Indo rices, in KZ. 6. 19 . p. 156. SS, Comparative Groma of Middle Indo-Aryan, is die Linguis , Vol. II. Caletta, 13. CEDA, 1956, 3. pour les adverbes de lien derives de cetime. HC entre outros F. En dit lllid Samarit G , New llaves, 1953, 91, . J. Budrimestre (. . . ) w a t anterieur de indien le theme catenere in vivo, quien l'extension de la forme a feminine kapal la th is une influence sur les pronoms relatifsche , .. Vec a Orillis, II, 1953. p. 3 L. P L ( ni). g. 1 . F. H e , Unidade, Calcutta, 1894. Glossaire... m alaya. F. E Buddhist llybrid Sanat D ery, New Haven, 1953, .. L. A. ScuWARZSCHILD, Melbourne IN W. , Pali Literatur vid Sprache, Strasbourg, 1916. I) Pour la theorie la plus recente sur I'r. c. II. B et, Zavi Probleme der wille indischen Landele, Munich, 1955, p. 43. IR L'explication phonetique proposte par G yer pour crise est rejeter comme elle n'est pas d'accord avec les tendances phoniques du moyen indien ilong ne peut at devenir 19 C. la formule te ur m wit m ediertel est le nombre de mes peche, j'ai peche tant de loin, Vavalu , I. Ona bucoup derit sur cette wirie de mus Nu , NZ, 35, p. 51 GRIGER, op. cit., . P l us Jacow. AZ, 58, p. 57 EDGERTON op. cit., 8, 10, II. , . . . . CMI On serait tente de proposer le thorie has los que lestrie tolly, etc., sort el pour quelque chose dans l'evolution des adjectifs de qualite, faina, etc, qu'on trouve en apable de surtout dans les rues diddotann o les adjectifs de qualite de l'hindi modem internete. L'explication par les formes dilciles de la saunaseni, Wide pour trin i t, n'est pas tout fit convaincante Il se peut que elarbes d e nieul d'abord esprim des Bundes de sens, ayam depuis quand ele, que ces ralli le de ses se sont peu a peu obscuri, come il arrive souvent aber les adverbes fredda le+ de in NHL TENER, Nepali Landen , wie Cr, bote supplantire sur la OM S. K. Chu , Origin and therefore the li langage Calcutta, 1996, 5600 TORSER, p.cit., Y. a . There is a certain exuberance which will not rest contented with traditional expressions, but finds amusement in the creation and propagation of new words and in attaching new meanings to old words', states Professor JESPERSEN in discussion of vulgarisms and slang. From the reading of some philological works it would appear as if the speakers of Prakrit and Apabhramsa had been singularly devoid of this natural creative exuberance, Middle Indo-Aryan first became known to modern philologists in the form of the more stereotyped literary dialects, particularly those of the drama, and there has been a resultant tendency to derive the entire vocabulary either from exact Sanskrit or Vedio prototypes (attested or hypothetical), or as a borrowing from the non-Aryan languages of India. It is mainly the work of J. BLock and R. L TURNER that has shown how many new and vital words were coined in the Middle Indo-Aryan period and have survived into the modern Aryan languages of India. The etymology of Prakrit and Apabhramba thakka tired', 'stationary' and thaklai 'to be tired', 'to stop' is a case in point. The older type of explanation is presented by R. PISCH, who derives thallen from a hypothetical Sanskrit sthakyati. He is followed in this by TYSITORT, Mopri and others. G. V. TAGAR gives a similar etymology. TEstac-ne as the origin of the adjective thakken tired. On the other hand, J. Brock and Professor TURNERT do not reconstruct a hypothetical Sanskrit or earlier original, but suggest that the word was formed in Prakrit itself from the root athd-'to stand', by means of a suffix 1. O. JESSEN, Language, Its Nature Development and Origin. London 1922. p. 2 R. Pesem Grammatik der Prakritsprachen. Strasburg 1900. 488 3. L. P. TESSOR, "Notes on the Grammar of the Old Western Rajasthani". Indies Antiquary 1914-1916. $72. 4 M. C. Mour, Gurjar all, Baroda 1956. Glossary sv. that. 5. G. V. TARE, Historical Grammar of Apabhradie Poona 1968. Glossary .. thaki. 6. J. Broch, La Formation de la Langue Marathe. Paris 1920.994. 7. RLTUR , Dictionary of the Nepali Language. London 1991. s.v. than. -69

Page #45
L. A. SCHWARZSCHILD containing -kk-. This article is an attempt to justify the latter view by a comparison of some of the other words in -kka found in Prakrit, of which there are a considerable number; past participles, verbs, nouns and adjec tives. 312 1. Past Participles in -kka. The past participle in Sanskrit ends in (a) -ta, (b) ita (c) -na (the latter form is mainly peculiar to roots with a final i,, d and j), but the verb pac- 'to cook' has the isolated past participle pakva 'cooked'. The first group (a), though the most important in Sanskrit has become less in Prakrit. The regular phonetic development, notably the assimilation of the final consonant of the root to the t- of the ending, has obscured the connection of these past participles with the rest of the verb. In Prakrit therefore it became more and more usual to form the past participle as in (b) by the addition of -i(t) a to the verb, generally to the present stem. PISCHEL thinks that the third group (c) of past participles was also extended in the popular language even before the formation of the Prakrit dialects, and that there was for instance a Sanskrit past participle mukena from the root muc- 'to release', side by side with the normal Sanskrit mukta 'released'. This, PrsCHEL believes, accounts for the Prakrit past participle mukka 'released', which is more frequent than mutta mukta. There are some clear extensions of the use of the ending -na, such as dinna 'given' for datta from the root de- 'to give', but the bulk of the examples given by PISCHEL fall into two categories: 1) past participles in -lla, which are due to an extended use of the Prakrit suffix -lla, and are not derived, as Pischel supposed, from popular Sanskrit forms in -dna and -Ina; (2) past participles in -kka, which are mainly associated with roots in -e and -e, and for which PISCHEL presumes a Sanskrit final -Jena; here a derivation parallel to that of the -lla past participles is more probable, and it seems unlikely for instance that there existed before the Prakrit period a hypothetical form *mukena from muc- 'to release'. In some of these cases it is almost impossible to reconstruct a plausible Sanskrit prototype: dams- 'to bite' could only have given a past participle dafna if we were really dealing with a mere extension of the -na ending of the past participle, and yet PISCHEL is forced to reconstruct a form dakna bitten', to explain the Prakrit dakka. Here even those who believe in a general extension of the ending -na admit instead an introduction of the suffix -kla in Prakrit.10 8. R. PICHEL, op. cit. SS508. 9. JAOS 1957. 10. P. TIME, "Merkwurdige Indische Worte", K.Z. 67. p. 155, Gottingen 1942. -70 PRAKRIT THAKKA, "TIRED' 313 There are some past participles in -kka in Prakrit which show the normal phonetic development of the corresponding Sanskrit forms, without the necessity for assuming an extension of the -na ending. Pac- 'to cook' has in Sanskrit the past participle pakva, which becomes in Prakrit pakka; Sanskrit vrase- 'to cut' has a past participle vrkna which becomes Prakrit vukka (there is a doubtful occurrence of this word in Hala v. 465, but it is definitely attested in Jaina Maharastri)." The Sanskrit adjective suska 'dry' generally became sukka in the Prakrit dialects. This was associated with/sus-'to dry' and was similar in appearance to the above-cited past participles. Another adjective that came to be regarded as a past participle was eakya 'possible', Prakrit sakka. Sakka took over some of the functions of satta < sakta 'capable', past participle of the verb sak- 'to be able, and appears in Jaina Maharastri in the meaning of 'capable'. Sakka is in fact listed as equivalent to the past participle satta by Hemacandra VIII.II.2. From the use of these forms in -kka there arose gradually the notion that -kka was an ending of the past participle and it was therefore liable to analogical extension in Prakrit. The analogy is most likely to have started among verbs whose roots ended in -e. These verbs remained as a fairly compact group: in cases where the c was intervocalic in the present indicative and should therefore have fallen, it was generally preserved e.g. pacai < pacati, and in Apabhramsa, it was even doubled under the influence of the fourth conjugational class and of the passive e.g. russai 'to please'. The extension of the -kka past participle was probably facilitated by the fact that these verbs already had a velar sound in the future and also had derivatives formed with a -k-. Mukka from\/muc'to release' is one of the most widespread of the past participles in -kka and was probably favoured by the confusions that beset the correctly developed mutta mukta (mutta < mutra and murta). Mukka is formed from muncai (var. muai) by analogy with pakka from pacai. Similarly luncai formed a past participle lukka, sincai formed sikka, and rincai rikka, while the participle lukka from the verb mluc-12 to go to rest' is somewhat more doubtful. These verbs maintained their association with the verb pac- 'to cook', past participle pakka, which had influenced them: there is in Prakrit a variant past participle pikka 'cooked from pakka + sikka and rikka. Like the Sanskrit verbs in -c those in Sanskrit & and -3 remained as a compact group in Prakrit: they too formed their future with a velar con 11. J. J. MEYER, Hindu Tales. London 1909, p. 68. 12. J. J. MEYER, op. elt. p. 259. note 2. lukka, 40 -71

Page #46
314 L. A. SCHWARZSCHILD sonant. Possibly under the combined influence of the verbs in -c and of bus- 'to dry' (which had the apparent past participle Pkt. sukka Prakrit dankhai, past participle dakka. More doubtful examples of this development arey/ bhras 'to fall', Prakrit bhamsai, past participle bhukka, andy/slis- 'to grasp past participle thikka 'hidden' in Prakrit. Some of the other examples that one might quote for the analogical extension of the -kka past participle are much more uncertain. Vokka, which appears in Apabhramsa as the second member of compounds with the meaning of 'speaking' might well come from vac- 'to speak' which has in Prakrit a future vokkhai. Sakka 'moved' is attributed by PISCHEL to a prototype evaskna, while the Dhatvadesas render it by V sarp- 'to creep' with which it might be associated. Still more dubious in its derivation is the word cikka 'touched'. Even if one leaves aside these doubtful cases the past participle in -kka appears as a well-established element in Prakrit morphology. II. Verbs in -kka The past participles in -kka are remarkably similar in their formation to those in -lla. Both these types of past participles often gave rise to new verbs, and hence even new past participles in -i(t) a. This latter development is more a feature of Apabhramsa than of Prakrit, e.g. mukka 'released' in Apabhramsa forms a new verb mukkai 'to release' and a past participle mukkia. The suffix -kka thus came to be regarded as a usual element in the formation of verbs and this was a very active principle in the modern vernaculars and has given rise to a number of verbs, as can be seen at at glance from Professor TURNER's Dictionary. The similarity with the extension of the -lle suffix is striking and sometimes the same roots were affected: e.g.V/muc- to release', Apabhramsa mellai and mukkai 'to release'; vac'to speak', vollai and vokkai 'to speak', 'to call. From the roots in -e, and -, -3, the suffix was extended among verbs of entirely different nature by the time of the modern vernaculars. This accounts for instance for the modern Hindi rokna 'to hinder from Sanskrit Vrudh, Prakrit rundhai, rumbhai. The analogical extension among verbs with roots in -e, - and is however not the only source of -kk- verbs in Prakrit. In Sanskrit there is -72 PRAKRIT THAKKA, 'TIRED' 315 a group of words, mainly onomatopoeic in origin and formed with the verb kr- 'to make': e.g. putkara 'puffing, thutkara 'spitting', camatkara 'astonishment. From these were derived Prakrit verbs of the types pukkarei 'to call out' with the past participle pukkia Page #47
316 L. A. SCHWARZSCHILD paikka's 'servant'. These words come from *rajikya, *nayikya, and perhaps *padatikya, based on raja, nayaka, and padati respectively. The majority of the instances where -kka is derived from Sanskrit -kya may be explained as adjectives with the suffix -ya formed from nouns in -ka in Sanskrit. These adjectives then often became nouns, as for instance trailokya > Prakrit telokka, which originally meant 'associated with the three worlds', hence 'the three worlds' already in Sanskrit. Similar in development must be "gonikya 'associated with cows' > Prakrit gonikka 'a herd of cattle' (quoted in the Desinamamala) and probably also Prakrit mahisikka 'a herd of buffalo' (ibid.). A doubtful example of Prakrit -kka from -kya is *satakya 'used for chariots'>"satakka modern Hindi sarak 'road' from *sataka, by metathesis from sakata 'a chariot'. From the numerous occurrences of -kka, whether from consonant +-ka or from -kya there was abstracted the new Prakrit suffix -kka, which had the advantage over the 'svarthe' -ka of being more distinctive: -ka had lost its initial -k- when added to a word ending in a vowel. The substitution of this new suffix for the older and less expressive 'svarthe' -ka accounts for a number of words that were listed by the Prakrit grammarians and the modern grammars following on them as showing irregular doubling of the consonant --. Examples are maukka from Sanskrit myduka 'soft' (used also in the sense of Sanskrit mydutva 'softness'), tuphikka 'silent' from Sanskrit tasnika, mandukka 'a frog' from Sanskrit manduka; ekka 'one' from Sanskrit eka, a word where the doubling was particularly important as avoiding homonymic clashes. Some words with -Icka have been explained as Sanskrit loanwords in Prakrit, and the double consonant is therefore regarded as showing the learned preservation of the Sanskrit intervocalic consonant -k-. These words probably belong rather to this category of substitution of the more distinctive prakritic suffix -kka for ka sakka from Sanskrit svaka 'one's own', kanikkd from Sanskrit kavika 'a grain', salakkia which comes from Sanskrit sarika 'a kind of bird'. Sometimes there was no 'svarthe' -ka in the corresponding Sanskrit word and the suffix -kka was newly introduced in Prakrit. This has occurred for instance in Prakrit ledhukka, from Sanskrit lestu 'a lump'. In some words of very uncertain origin, often listed as Desi words by the Indian grammarians this final is also found: e.g. phirakka 'a waggon', tikka 'a spot' (though this might have some connection with Sanskrit tilaka 'a mark'), pikka 'pure', khadakki 'a side door' and a number of others. There 15. J. BLOCH, loe, eit. regards this as a borrowing from Persian. 16. P. THIME, op. cit. p. 184-185. 17. R. L. TURNER, op. cit. s.v. tilo. -74 PRAKRIT THAKKA, "TIRED' 317 are many words in the modern vernaculars that go back to originals in -kka not attested's in Prakrit or Apabhramsa, but of considerable geographic extension, such as "tukkal a piece' > modern Hindi fuk, *tokka 'a basket' > modern Hindi tolera. In some of these cases the suffix -kka has been substituted for the original final of the word: thus modern Hindi mulcka 'a fist' from "mukka for mutthi < multi, Hindi math, 'a handle'. There is then no doubt of the existence of a very lively nominal and verbal suffix -kka in Prakrit, though it has no one particular point of origin in Sanskrit. The maximum extension of this suffix was probably at the time of the formation of the modern vernaculars. It is impossible to attribute any uniform meaning to this suffix -kka: among modern verbs one finds such pairs as Hindi ghalakna (formed with the -kk- suffix) 'to lean over' and dhalna 'to fall'; Nepali jhulkanu (with suffix) 'to rise (of sun)' and jhulnu 'to swing'. In some verbs the suffix might have an emphatic meaning and this may be due to its association with onomatopoeic verbs. In some nouns one could see in it a diminutive meaning, but on the whole the function of this suffix is one of formal enlargement, expressiveness and avoiding of confusion with other words. None of the functions of the suffix -kka mentioned hitherto would explain directly the development of thakka 'tired', 'stationary' from the root stha- 'to stand'. Stha- does not share similarity in consonants with the group of verbs pac-, sus- etc. Thakka cannot therefore be simply classed as one of the past participles of the type mukka. It is certainly a later formation than mukka, as it does not figure in the early canonical Prakrit texts, and only is a feature of Jaina Maharastri and particularly of Apabhramsa. By the time of the formation of the adjective thakka there must have been already a definite suffix -kka, which had little enough semantic significance, but was phonetically strong. Thus it is most likely that thakka belongs to the third category of words discussed, namely that it represents an enlargement of the adjective tha this was similarly enlarged and gave *thiakka whence the modern 18. Some of these have been explained by J. BLOCE (loc. eit) as borrowings from the Dravidian. 19. This might be connected with the Sanskrit root, truf- 'to break". 20. This is a particular characteristic of 'slang' endings (cf. Jespersen op. cit. p. 300) as for instance the ending -y which can replace all sorts of finals in Australian: football footy; Tasmania > Tasy; swagman > swaggy etc. -75

Page #48
313 L. A. SCHWARZSCHILD Hindi thike 'correct'; and theya durable', Prakrit the (y) gave heaklee, whence the modern Hindi thele 'support. These two words are of parallel formation to thakka tired', 'stationary' but probably later in date, and they are not attested in the Prakrit or Apabhramia texts so far available. The enclitie adjective tha seems to occur only with a dental initial in Prakrit, whereas the other derivatives of sthd-, this and the (y)vary in their initial and generally prefer the cerebral which has survived in the modern forms. The development of the meaning of the word thalka does not present any major difficulties. The concept of being tired' is usually expressed by past participles of verbs meaning to work hard': thus Sanskrityram to exert oneself, past participle ininta tired'; but it can also be expressed by words meaning 'having just stopped work', 'being stationary': this seems to be the case particularly in the lower strata of the language (eg. English 'finished"). Thallow maintains the old meaning of being stationary side by side with that of tired' and the two meanings are also found in the cloai, the verb formed from it. The concept of being tired, just like that of sundry other physical conditions, being fat, hungry, drunk etc. is apt to give rise to a variety of slang expressions, which ometimes replace the standard word. The word thakka was probably accepted into the literary dialects of Prakrit more readily than words of similar formation on account of the hopeless confusion that overtook the standard word for tired': Sanskrit kranta "tired' > Prakrit anta; Sanskrit idnta peaceful > Prakrit santa. It remains certain that the adjective thalle, formed as it is with the help of a popular prakritie suflix can only be explained by conditions within Middle Indo-Aryan. GLEANINGS FROM THE VASUDEVAHINDI By L. A. SCHWARESCHILD I. The prefix 1- for Ou Sansierit avd-, apa-. In his article "The Vasudevahindi, a specimen of archaic Jain Maharastri' BSOS 1935, Professor Alsdorf has discussed some of the important morphological peculiarities of this most interesting text.* There remain some minor phonetic problems connected with words occurring in this work: one of these is the replacement of the Oderived from Sanskrit and and apa- byt. The Sanskrit prefix ava- 'down' normally appears as o- in the various Middle Indo-Aryan dialects according to the general rules of phonetic development. There are however some words in which this o- has become According to the standard works of Geiger and Pischel? o and t, likee and I were easily interchangeable in closed syllables in Middle Indo-Aryan. But it has been shown by Berger that this was not the case, and that the opposition between o and 14, and e and i still had some functional yield. For open syllables a free interchange between o and has never been claimed, and it is therefore all the more surprising to find in the Vasudevahindi uyariya p. 46 line 11, from Sanskrit avatarita, ulmo from avatirna etc. ulambio p. 147 line 6, from aualambita, tuiddhani p. 135 line 26 from aaviddhani, uloyana p. 17 line, 18, from analokana, and further useraha p. 102 line 7, from apasaratha, thanand p. 102 line 4 from apabhd pana The meaning of these forms leaves no doubt that they correspond to Sanskrit words beginning with ava- down' and apa- 'away'. Thus uyar-avatar. always means to get down': e.g. p. 46 lines 1011 sa vi Samadatta ayriye rahdo 'and Samadatta also was helped down from the chariot'; on p. 147 line 3 in speaking of a well the story-teller says cham syarami md tubbhe 'I am going down, not you'; p. 16 line 4 winno sibigdo 'he got down from the palanquin'. In some instances Prakrit sya. "to get down' alternates with the References throughout are to the Bhavnagar edition of 1931. Geleer, Pal Literatur und Sprache. Strasburg 1916. 15. Pischel, Grammatik der Praliterchen, Strasburg 1900. 84 and 119 3. Berger, Zel Probleme der mittelindisches Loutlehre Munich 1955. P 51 3a. For conditioned changes of this kind of. H C. Bhayani's article "Gujerati mam nasiya vyanjan parveni mul, num Lurupe parivartan. Jon. San. Re. So. 1 .77 -

Page #49
Nos. 1 & 2] GLEANINGS FROM THE VASUDEVAHINDI BHARATIYA VIDYA [VOL XVIII normal oyar-avatar- and sometimes even with the semi-learned avayar-avatar p. 134 lines 14ff. uyaraha,kisa vilambaha come down into the water)', says one of two friends 'why are you delaying?' Addhanam parikkamiya pa sahasa jalan avayariyauvan 'when one has been walking some distance one should not suddenly get down into the water, answers the other. The words in which u-has replaced Sanskrit ave- or apa-before a single consonant do not strictly speaking constitute a linguistic peculiarity of the Vasudevahindi, but their frequency in this text is striking. There are parallels in other Jain Maharastri works; e.g. in Jacobi's Ausgewahite Erzahlungen in Maharastri we find tulugga avarugna 'sick'. This word recurs in a passage of very similar content in the Nirayavaliyao, though there are a number of variant readings, olugga and even ulagga, which may be derived from avalagna, uithaapalersta appears in the Nayadhammakahao I. 8. but here too there are some variant readings. Further sporadic examples occur in other Middle Indo-Aryan dialects, thus Buddhist Hybrid Sanskrit ulcirati for avakiratt is quoted by Edgerton. There are some indirect survivals (vide infra) of this development in Apabhrarsa, but it does not seem to have been extended, with a few possible exceptions e.g. Nepali wira for oirocavagirate. It is at once noticeable that apart from the Vasudevahindi this development is very restricted, whereas before double consonants the change of arx- and apa- to - is not uncommon in middle IndoAryan. Some of the many examples are: nuggaha cavagraha 'obstacle' which occurs in the Jain Sauraseni of the Pravacanasara, in Ardhamagadhi and in Jain Maharastri; ukkhand-from avaskand'to ambush', 'to attack which appears in Ardhamagadhi, and utthambh from avastambhto support'. Less numerous and in a different category are those words in which the prefix ava-(and apa-) was originally followed by a word beginning with a single consonant which was subsequently doubled, while the o- of the prefix became te.g. uklas- from analys- to draw away (Ayarahgasutta II, 3.1.15), uttamsa from avatamsa 'head ornament' (Gaudavaho). Words in which the o- from ava- and apa- was replaced by 1may therefore be conveniently divided into three types: (1) qua-, apa-appear as t- and the following single consonant develops regularly. (2) ava-, apa-appear as 14- before a double consonant. (3) ava- and apa-appear as t- and there is reduplication of the original single consonant. The question arises: are these three developments part of the same linguistic process? Requirements of metre might explain some, but by no means all the examples of 1- for ava-, apa-, and naturally not those in the prose of the Vasudevahindi. The explanation is probably partially phonetie: the great similarity in pronunciation between o and u, particularly when they were both short, facilitated a confusion with the prefix ud- 'up'. This has already been suggested a long time ago by Wackernagel and later by Bloch.8 The prefixes concerned differ markedly in meaning; ava- down and ud- 'up' remained strongly contrasted in some compounds, as for instance in the two Jain technical terms of opposite meaning: osappini aaarpini 'descending cycle and suppini utsarpint 'ascending cycle', which recur throughout Ardhamagadhi (e.g. Th&nanga I.1.) and Jaina Miharastri. On the other hand the basic meaning of some verbs obscured the differences in sense between the prefixes, and a glance at a dictionary 10 will show how this applies already in Sanskrit to some of the words under discussion: tro- to cross traverse', 'accomplish' etc. ava + tre- descend to (from)', 'come down to earth', 'become in carnate', 'get over, 'attain ete. ud + tra- 'come out of", emerge from', 'get over', 'overcome'. stabh-to fix', 'establish' etc. ava + stabh-' to fix', 'support', 'grasp', 'take prisoner. ud + stabh-to fix aloft', 'set up'. kre- 'to draw'. ava + kys- 'to draw away', 'attract'. ud + krs-'to draw out', 'raise'. apa + lers-'to draw away', 'remove', 'put aside'. The overlapping in meaning between the compound verbs concerned together with the phonetic causes mentioned above brought about a confusion between these prefixes in Middle Indo-Aryan: 0-Page #50
Nos. 1 & 2] GLEANINGS FROM THE VASUDEVAHINDI 25 BHARATIYA VIDYA [Vol. xvm This confusion between prefixes explains words in category 2 where av-0- was in a closed syllable, but it cannot account for the difference between category 1 (avatarami uyarami) and 3 (avaler- ulclas). One would only expect the latter, as the final consonant of the prefix ud-caused reduplication of the initial consonant of the following word. The changes undergone by the Sanskrit system of composition by prefix in Middle Indo-Aryan help to explain the varied developments. Sometimes phonetic change has caused a weakening of the system and compounds were no longer recognised as such, particularly if they had become dissociated in meaning from the simple verb: thus Apabhramsa baisai 'sits' is no longer felt as a compound verb derived from spa + visati, the simple visai tviddha, apasarati osaras + -> sarai ete. As in a number of other instances the archaic Jain Maharastri of the Vasudevahindi has thus given us an indication of a transitory feature of the spoken language which is little attested elsewhere. For a while there existed side by side the two factors which account for the differences in development between the two categories 1 and 3: consciousness of a prefix u plus double consonant on the one hand, and on the other consciousness that a simple was also a prefix. Gradually the first tendency gained ascendaney, seeing that it was in agreement with the general trend to preserve the initial consonant of the word that followed the prefix and to remodel compounds. In the Vasudevahindi there is some overlapping in meaning, as already pointed out for Sanskrit, but on the whole syarat cautarati 'to descend is still kept separate from uttarai utarats 'to emerge from'. In Apabhramia uttarai has ousted syarai and we find for instance in the Paumasiricarius uttariya in the meaning "put down'. It is from this form with the double consonant that Gujarati utarou, Hindi utarnd etc. are derived. Similarly Hindi tard to recede' is not derived directly from Sanskrit tutarati 'to leap up', but from apasarati 'to recede' via Middle Indo-Aryan sraite fuarai (Vasudevahindi). With recomposition and doubling of the consonant this word gave ussarai in later Jain Maharastri and in Apabhramba. The intermediate stage shown by the Vasudevahindi gives an indication of the complexities of developments in Middle Indo-Aryan: it shows just one detail of the many changes and remodellings that helped to weaken the old system of composition. The complete decay of the system and the gradual rise of new methods of expression is characteristic of Apabhrana. 13 or a different interpretation of these two examples ef. Pischel, op. cit. 141. 14. The etymology volkantavalduta glven in the Pala-saddha-mahannavos satisfactory as it involves a complete reversal of the meaning of the word. 15 CE HC Bhayanland MC Modl, Pow e r Singh Jain series, Bombay 1968 Glossary Ev. 16 Gujarati - may represent a recombosition of this form or it may more probably show the influence of oral. CI. P. B. Pandit loe. cit - 80 -

Page #51
SOME ASPECTS OF THE HISTORY OF MODERN HINDI NAHIN "NO", "NOT" By L. A. SCHWARZ CHILD IN AN EXAMINATION of word-phrases A. Meillet stated long ago : "Des mots comme oui, non representent le plus haut degre d'abstraction que puisse atteindre ainsi une reponse consistant en un seul mot." As has been repeatedly pointed out, the classical IndoEuropean languages did not have any exact equivalent of this abstract method of expression. The majority of the modern languages on the other hand have arrived at these convenient "special expletive interjections". Hindi nahin, like the cognste Manithi ndhi, Gujerati nahi(s), etc, typifies this development of most modern Indo-European languages both syntactically and formally it is used as an equivalent of"no" (though it may serve also as negative adverb), and it represents an enlargement of the old Indo-European negative particle, Sanskrit wo. This formal and syntactie transformation of Sanskrit ne into modern nahin has been variously explained. The standard theories involve the addition to the negative particle of some part of a substantive verb, a development by no means isolated in the Indo-European languages. They may be summarized as follows: 1. Theories in which parts of the verb "to be" are added. (a) Kellogg stated: "The common negative natin, Braj whi has arisen from the combination of the negative with the 3rd singular Ali of the substantive verb. (6) S. K. Chatterji thinks that a i based on Sanskrit asti may have been added to m. (e) Dwijendranath Basno believes that only a derivation from THE HISTORY OF MODERN DE VANTNO", "NOR 45 no+dait can account for the Bengali forms. Similar explanations were given also by Sen and others. II. Theories in which parts of the verb bh"to be" are added. (a) L. P. Teritorisuggests that the origin of the Old Western Rajasthani forms was from news and humPage #52
46 TAR HISTORY OF MODERN HINDI ANTYKO OT changes in the form of the negative particle. There appears in the canon a series of enlargements of the negative particle for emphasis and distinctiveness. Some of these enlargements date back to Sanskrit and differ from their Sanskrit prototypes mainly by their frequency. In the very oldest parts of the canon such as the Ayamngasutta pa is still the most frequent type of negation, sometimes even ousting md with prohibitions, but elsewhere in the canon 90 Prakrit naki, Sanskrit wo + api > Prakrit maur, and makhalte > Bathu. In later dialects, such as dramatic Sauraseni and the Maharastri of the lyrics (eg. Vajjalaggam) akhte has been weakened further to make and it survives into Apabhruins and even in the early vernacular texts of W. India, e.g. the Gurjararasavali. Navi natthi is frequent in Ardha-Magadhi as in Pali and it has clearly lost its association with the 3rd person singular and has become stereotyped, as is shown by its use with plurals, e.g. Vivagasuya, story of Mrgaputra: watthisse daragaasa hatthd vd payd kannd o acchi ni yasa d... "the boy had no hands or feet or ears or eyes or nose". Natthi in such cases is scarcely more than just a reinforced form of the negative particle. Its survival into Apabhrama and as the modern Gujerati nathi is well known. TIL STORY OF MODERN HINDI NANIN XOXOT 47 Apart from these combinations of the negative particle inherited directly from Sanskrit there are also in the Jain canon new negative particles where purely Prakritie elements make up the reinforcement. There is, for instance, the form dim. This occurs in a fixed locution where an offender asks for forgiveness, ending with the words adim bhujjo kernayde and I will not do it again" (e.g. Uvas gadasio i113 ed. Vaidya). Naim has been explained by Dr. Vaidya in his note on the passage in question. He compares it with punin which is found in Ardha-Magadhi for pu p unar "agnin". This extension seems to have started among the pronominal adverbs of time such as Sanskrit Kudacit "sometimes" > Prakrit kayfi. The close association between m and punar in particular can be seen from such Prakrit forms as w e Page #53
48 THE HISTORY OF MODERN DE SANT "0","0" frequent in the works of Bhisa, eg. Svapnavisavadatta Act II, where & maid asks: Bhattidarie, jadi so vidi o bliane... "Princess, if the king were to be ugly ..." Vali mali answers Vasavadatta, dantzario erra. "No, he is handsome." In some of the slightly later dramas one occasionally comes across cases where make without repetition conveys the meaning of "no"e.g. Malavikagnimitra Act III, Malavika: mapaso chandena mawless ? "Do you say this of your own accord?" Maid: Nahi Shattino edim... allhardin."No, these are the words of my master." Even here there is a variant reading with repetition of a But as regards for the post-canonical Prakrit dialects do not appear to have either of the unusual reinforced negatives found in the canon, dim and whe, though aim is permitted for Prakerit by Hemacandra's Grammar (11.190). In 11.191 Hemacand even gives mfim, which must be derived from ma not used with injunctions, while the final syllable is due to the influence of naim the word maim is not to be found in any texts. The Ardha-Migadhi negatives dim and hi are absent even from a popular JainaMaharastri text like the Vasudevahindi, where apart from all the ordinary enlarged forms of the negative we find only nai Sanskrit na cid, Naim and mihi recur in Apabhrama: they are found in the texts of the Digambaras and Svetambarns alike. The resem blance between the Jain canon and Apabhrami s regards negation is more than a coincidence. Ndim und ndhinised in the particularly emphatic passages quoted from the canon, presumably belonged to the popular language and survived as emphatic negatives in those parts of India where the literary Apabhramia were formed. They were then spread far afield by the literary Apabhiramies, as is shown by the wide distribution of the derivatives of Apabhramsa niki Unlike whi, Apabhraps me seems to have left few direct descendants (possibly Bengali ay may be counted as one of them). This was due to the fact that Apabhramadim was less distinct as a negative, being identical with Apabhrama din (cf. also Apabhrama nam, sim, sirai), which had the meaning of "like", "as if", and was clearly the ancestor of modern Hindi nim. The confusion between him notanda "like" in Apabbra TRE HISTORY OF MODERN HINDE VANTO", "SOL" 49 affected the other negative particles, so that one finds now "not" Prakrit na ya "and not","nor", a very frequent combination already in Sanskrit and especially so in Middle Indo-Aryan. It would thus form a parallel to Apabhrama now "not" Page #54
50 THE HISTORY OF MODERN HINDI NAHIN "NO", "NOT" naya to nd by the Apabhramsa period, although details of this phonetic change are still uncertain. The appearance of this form na na ca gave new vitality to the popular emphatic negative vali. Na itself left a number of derivatives in the modern vernaculars, it is found for instance in Kashmiri and Lahnda and in modern Eastern Hindi as well as in earlier texts from that region (e.g. the works of Jayasi and Tulsidas). The later phonetic development of the Apabhramsa negative ndhi(m) does not present many problems. The lengthening of the final syllable that characterizes the Hindi derivative is probably based on the influence of the frequent final -hin of adverbs such as kahin. The correspondence of the final of these locative adverbs with the final of the negation in some of the other modern IndoAryan languages lends support to this view: Gujerati has tahi and nahi(m), while Old Gujerati has variants such as kahia(m) for the locative adverb and nahia(m) for the negative; Nepali has kahi and nahi, etc. The syntactic advance shown by the Prakrit of the dramas in the use of nahi Page #55
BHARATIYA VIDYA I VolXIX (c) Dave suggested a derivation from the Sanskrit tanaya "son', and this view was considered the least objectionable by W. N. Brown. (d) Tessitoris suggested a derivation from amanaka 'one's own with loss of the initial vowel by aphaeresis. (e) Modi mentioned that there is a possibility of deriving it from a pronominal form of tad". The difficulties in the case of the postposition resi, Tesamusi are perhaps even greater and no suggestions as to its origin appear to have been made. Both taxaya and resi, resmi are so well attested that there is little doubt as to their role and meaning. Tanaya was a postpositional adjective and was usually added to a noun or pronoun in the genitive case. Being an adjective ramaya agreed in gender and number with the noun that it described and which generally followed it, e.g. Jixindaha tandu dhammu the law of the Lord Jina' Paumasiri cariu IV, 64). It was also occasionally used to form a compound with the noun before it which was then left undeclined, and in such a ease tanaya fulfilled a more truly postpositional function, e.g. divammi tammi dittho kim sartho Rayagihatanao 'has the caravan from Rajagrha been seen on this island?! (Nanpanicamikaha IV. 84). Sometimes the use of tayaya has been further extended, particularly in later Apabhramsa, and it has reached the transitional stages bet ween a postpositional adjective and an indeclinable postposition. It could be used as Bhayanil has shown in the phrase... arena loranera 'by reason of ... and with the elision of cranega it tended to become a causal postposition in its own right, being listed as such by Hemacandra IV.425 idarthye Icehim-chim-resi. resim-tanenih, 'Kehi, tehim.resi, resim and tonna are used in a causal sense'. In most of its earlier occurrences tanaya was used with pronouns rather than nouns. Thus it figures in the Paramatmaprakala: mahutanci 'mine', in the Sapayadhamma-doha: tasu tanaim 'his, and in the Paumacariu of Svayambhu: kaho tanan dhana 'whose wealth? In this use with a pronoun to form a possessive, tanaya resembles an even earlier attested postposition, kert, which is found in Mahl. Tastre, e.g. in maha keramine (Hila). There is even evidence that like tera, tanaya was actually used to form a possessive adjective 6. T. N. Dave, A Study of the Gujarati language in the 16th century, London, 1905 58 7. W. Norman Brown, "Some Postpositions behaving as Prepositions in the Old Gujarati Vatavilla di Lin e s 1958 p 31 3.LAP Tester Indian Antiary p208 3. M. C. Modi, Gorje Berada, 1956 Clay S. Tai 10. HC. Bhayani, Pemer of Kevinja Sambidere. Bombay, 1958, p. 67 Nos. 3 & 4) NOTES ON TWO POSTPOSITIONS OP INDO-ARYAN that was felt as a unity, tujjhanau (attested in the Kuvalayamala) from tujjha+tandu.2 The declensional system of Apabhramia maintained some clarity. Thus mahu in mehutanai stands for the genitive or dative-ablative singular of the pronoun and would therefore in its context indicate possession without the addition of tanai. Similarly in the case of nouns, phalai saggo-dumalo could mean the fruit of the heavenly tree' without the addition of tanaim as in the Paumacariu 3.7.1. phalaim saggo-dumaho tanim. As in the case of the history of languages outside Indo-Aryan, for example in Romance, pronouns tended to keep some case distinctions longer than nouns, and in fact they even retain a diversity of declensional forms in Modern IndoAryan. One would therefore expect fewer, not more postpositions with pronouns than with nouns. Yet the opposite state of affairs prevails and postpositions are more frequent with pronouns than with nouns, particularly in Apabhrama and to a lesser extent even in the earlier stages of the modern languages, for instance in Awadhi, as pointed out by B. Saksena. The frequency of tawaya with pronouns and its use with nouns clearly showing the genitive ending indicate that this postpositional adjective was not used simply as a substitute for the ending, and the decline of the case-system and the rise of the postpositions in the case of anaya at least cannot be dismissed just as a matter of cause and effect. If one explains tandya as an emphatic form of the possessive, then its frequent use with pronouns is quite understandable: emphasis is felt particularly necessary with pronouns and mahu-tagaya etc. would simply correspond to a more emphatic method of expression than mahu 'of me', perhaps more emphatic even than the Prakrit mala-lera "mine", Apabhramsa mera, and it would be equivalent to my very own' This aspect of Apabhrama usage also throws some light on the etymology of tanaya. It renders improbable the suggestion by Beams that tanaya was derived from the Sanskrit suffix -tang. This suffix was used exclusively to form derivatives associated with time, e.g. cirantana, puritana 'ancient', sadatana 'everlasting' etc. As pointed out by J. Wackernagel 15 it occurs occasionally in rare and late for mations associated with words of a local rather than a temporal meaning, but it never had any strong possessive sense and there is no sign of it ever having become emphatic. Moreover Sanskrit -tana, 11. This form is quoted by A. Master, "Gleanings from the Kumalaemata-sha", BSOS 21950. 27. However in a very recent article in the Jubilee Volume of Braced (Adyar 1951). "The Late MTA Possive Suff Hc Bhayant has shown that forms like t hese are probably unconnected with the possessive postposition fame 12. B. Saksena, The Bello of Awad, Allahabad, 1907, p. 214 13. J. Wackerna indische Grammatik edited by A Debrunner, Vol. 1 2 Gottingen, 1954, p. 55. - 90 - - 91 -

Page #56
80 BHARATIYA VIDYA (Vol. XIX a mere suffix, appears unlikely as a prototype of tanaya which occurs in Apabhramsa mostly as an independent word. Emphatic expressions in the popular language generally have an element of novelty and tempt one to discard any explanation that is based on a well-worn word or phrase which is likely to be weakened in meaning. Bloch's suggestion that tanaya is a direct derivative from the verb tan- 'to stretch' appears therefore to have a certain measure of probability, particularly as such a formation seems closely parallel to the other usual possessive postposition, kera karya from kr- 'to make'. But karya is by origin an obligatory participle and the presence of the vowel -a- before the -ya suffix is unheard of in such participles. A direct derivation of tanaya by means of the suffix -ya is therefore highly unlikely in Sanskrit14 as well as in Middle Indo-Aryan. There is even general agreement that the Sanskrit noun tanaya 'son' has no direct connection with the verb tan-. In Middle Indo-Aryan there do exist derivatives in -aya from verbs, but these are agent nouns, as this -aya comes from Sanskrit -aka: e.g. Sanskrit himsaka > Prakrit himsaya 'one who injures', Sanskrit vinasaka > Prakrit vinasaya 'one who destroys', from the verbs hims- and nas. In such derivatives the vowel of the first syllable is generally lengthened 1e and one would therefore expect *tanaya. It is however mainly the meaning that renders the derivation of tapaya from the verb tan- impossible, as it seems hard to imagine how an agent noun meaning 'that which stretches or extends' could possibly have become an emphatic possessive adjective meaning 'one's very own'. The explanation suggested by Bloch therefore remains extremely doubtful both on phonetic and semantic grounds. The emphatic use of the postposition tanaya and its close link with pronouns render unlikely Dave's suggestion of a derivation from the Sanskrit tanaya. Tanaya 'son', 'descendant' is sometimes used in Sanskrit as an adjective 'belonging to one's own family'. There is however no evidence in Prakrit for this adjectival use of tanaya: it seems to have been simplified and restricted in meaning, and is found in Prakrit only with the meaning of 'son' which is irreconcilable with the later Prakrit and Apabhramsa emphatic postposition tanaya. Modi's interesting suggestion that the postposition tanaya is derived from the pronoun tad was not developed by him any further. It is in agreement with the geographical distribution of the 14. J. Wackernagel, op. cit., p. 213. 15. M. Mayrhofer, Kurzgefasstes etymologisches Worterbuch des Altindischen, Heidelberg, 1953, ff., s. v. tanaya. 16. J. Wackernagel, op. cit., p. 146. -92 Nos. 3 & 41 NOTES ON TWO POSTPOSITIONS OF INDO-ARYAN word. Tanaya is found only in Southern and Western Apabhramsa and in Prakrit texts influenced by Apabhramaa. It is very usual in Old Western Rajasthani and has survived in poetic usage in modern Gujarati. Probable further derivatives of tanaya are confined to the western and central areas, for example the postposition ne in Gujarati, Panjabi and Hindi, and nem, nim in Marathi. 17 It is more doubtful whether the modern Awadhi ten 'from' is a derivative of tapaya. One would therefore be inclined to associate tanaya with a prototype that was characteristic of the west of India. The oblique of the pronoun tad appears in Hindi and the associated group of languages as tin, and only in the extreme west and scattered other areas does one find an oblique of the demonstrative pronoun that resembles tanaya: Sindhi has an oblique plural base tan-, Mewari an oblique plural tanam and an oblique singular tani, while Kumaoni has an oblique plural tanana or tanu. These forms are probably derived from the Prakrit genitive plural of the pronoun tad, which was tanam, and the short vowel in the first syllable of the modern forms is due to weakening in an unaccented word. The association between the genitive and the possessive is close and the suggestion made by Modi would therefore carry some conviction were it not for certain objections of a phonetic nature. The genitive plural tanam, which occurred along with tesim as the most usual form in literary Maharastri, was generally replaced in Apabhramsa by taham. Tanam, according to the evidence of the survivals in the modern languages, must have continued to exist in the spoken language over some parts of western India. Had the postposition tanaya been directly derived from tanam, one might at least in its earlier occurrences expect to find a long vowel in the first syllable: tanaya. 81 From the point of view of meaning the most probable source of the postposition tapaya would be a word meaning 'self'. Hence Tessitori's theory of derivation from atman 'self', atmanaka 'one's own' has much to commend it. But some phonetic difficulties lie in the way of this theory too. The loss of the initial vowel is to be expected in a word used as a postposition, and so there is no need even to consider the alternative Vedic form of atman, namely tman 'self' as a possible prototype. The real difficulty lies in the fact that the change of -tm- to -pp- in atman> appana just like the closely associated change of -to- to -pp- and of -du- to -bb- was characteristic mainly of the western areas of India, particularly in the Prakrit and early Apabhramsa period. This is shown by the frequency of appana Page #57
82 BHARATIYA VIDYA [Vol. XIX of intermingling of these forms: thus the originally western form apaatman appears in the Old Bengali Caryas and figures also in the modern languages of the east of India, while the eastern type of form is actually used as an emphatic possessive several times in the very late Apabhramsa of the Prakrtapaingala, attanaya -ya suffix and to join the more usual type of feminine declension in -a.24 Tanu 'person', 'body' was not one of the words so treated in Prakrit and Apabhramsa. On account of its meaning it was generally found at the end of possessive compounds in the nominative or vocative or at the beginning of other compounds and was therefore not often de 23. It is probable for instance that tanuhei, a word of doubtful interpretation, which occurs in the Paumacariu, ed. Bhayani, vol. I, 8.5.3. represents tanu-bhedin "body-splitting". 24. G. V. Tagare, Historical Grammar of Apabhramsa, Poona, 1948, p. 178. -95

Page #58
BHARATIYA VIDYA [Vol. XIX clined e.g.: tapu-kanti-bhariya-gayan' angandi 'the beauty of whose body filled all heaven' (Kumarapalapratibodha J. 93). It thus came about that tanu was ultimately no longer considered as a -u stem. In the late Apabhramsa of the Sandesarasaka it forms a locative singular tani (vs. 187 and 189) and shows every sign of being treated exactly like the ordinary masculine-a declension whose nominative singular ended in -u, and with which it was therefore liable to be confused. There is no doubt, however, that this confusion with the masculine stems in a belongs to a late stage of Apabhramsa and it would not account entirely for the earliest occurrences of the form tanaya. It seems that several words in -u, both feminine and masculine, formed derivatives even in the Prakrit period in which this final -u did not figure: kangu 'bracelet' has a derivative kangani of similar meaning: pasu, from Sanskrit pasu 'animal' gave rise to pasaya 'a kind of animal', which is found already in the Jain canonical texts but is sometimes wrongly considered as a Desi word; teu from tejas forms a derivative teaya; etc. There is thus no morphological objection to a derivation of tanaya from the word tanu 'person', -body'. Part of the reason why the type of formation shown in the North-Western Prakrit, tanuvaka 'one's own', did not become popular over a wider area may be the existence of a homonymous word, the adjective tapu 'little', 'slight'. Tanuvaya was too liable to be associated with tapu 'little', generally lengthened to tapua, which was a widely used word and itself the basis of several other words taquei 'to weaken', tapudi 'to be weak'. The form tanaya 'one's own' on the other hand did not cause any confusion with any other adjective, it only resembled the noun tanaya 'son', too different in use to lead to any major confusions. 84 Tanaya 'own' is therefore probably a derivative of tanu 'body', 'person' characteristic of the western and north-western areas of India, and it is reflected as such in Apabhramsa and Old Western Raajsthani as well as modern Gujarati, while some less certain derivatives are found over a wider area. For a period tanaya was felt as an emphatic possessive adjective 'one's own', 'personal' and even in late Apabhramsa and early Gujarati it had not yet undergone all the loss of independence that distinguishes a true postposition. In fact it appears in the Old Gujarati of the Vasantavilasa-phagu almost as a preposition, being placed before its noun as for instance in ritu taniya-vasanta 'the season of spring'.7 Tanaya is thus a clear example of a postposition in the making, as yet little affected by the lack of independence and accentuation as well as by the rapid phonetic changes that account for modern postpositions such as ne, which may be derived from it. -96 Nos. 3 & 4] NOTES ON TWO POSTPOSITIONS OF INDO-ARYAN The postposition resi, resammi is in quite a different category from tanaya. It is known only from a much smaller range of texts and over a shorter period. It already has all the typical features of a postposition which makes the search for its origins so difficult. Resi(m), resammi appears only in comparatively late Apabhramsa and in very late Prakrit texts influenced by Apabhramsa. 25 It is usual in the older texts in Gujarati, e.g. in the Vasantavilasa-phagu and in some of the poetry from Rajasthan, such as the Dhola-Maru ra Duha and the Beli Krisan Rukmini-ri, but it does not appear to have survived right into the modern spoken languages. The meaning of the postposition resi(m) is clear from its earliest occurrences: it is a dative postposition indicating purpose e.g. in the Kumarapalapratibodha S. 49.9, ta valivi pucchiu samanu, sauna-parikkhana -resi 'they caused the monk to turn back and questioned him in order to test the bird'. Sometimes, as here the word preceding resi(m) is compounded with it and left uninflected, but sometimes the instrumental-locative ending is used before resi(m), as in Hemacandra IV. 425 annahi resi 'for someone else'. Resi appears in various forms: the earliest and most frequent form is resi(m), but resammi and even resimmi are occasionally found. It seems clear that resammi and resimmi simply represent resi(m) with the addition of the locative singular ending (a)immi. This locative singular ending strictly speaking belongs to Prakrit (AMg., JM., JS. and M.) but it is found also fairly frequently in Apabhramsa. Long ago Grierson26 discovered similar developments in the modern languages. Giving some tables of modern postpositions he stated: "It is an interesting fact that every dative suffix in the above two tables is identical with the (modern) locative of a genitive suffix." Resammi, resimmi probably represent the first instance of this tendency to mark dative postpositions by the addition of a locative ending. Grierson's statement further helps to elucidate the ultimate origin of resi(m), it strengthens one's suspicion that resi(n) was derived from a very usual and early attested genitive postposition, namely kera< Sanskrit karya, Resi (m) has the general appearance of a genitive plural of a pronominal form: tesim, kesim, annesim, savvesim etc. were the usual genitive plural forms of pronouns in AMg. JM. and JS, and according to Hemacandra (III. 61) they could be used for the feminine as well as the masculine of the genitive plural. Now kera < karya was not in itself a pronoun, but was used so much with pronouns, particularly in the formation of the possessive pronominal adjectives maha-kera > mera etc., that it was liable to be influenced by the pronominal type of declension. A genitive plural keresim, 25. H. C. Bhayani, "Some Interesting Features of the Prakrit of the Napapaficamlkaha", Bharatiya Vidya Vol. XII, 1951, p. 157. 26. G. A. Grierson, op. cit., p. 476. -97 85

Page #59
BHARATIYA VIDYA [Vol. XIX though apparently not attested at present, does not therefore contravene the normal rules of declension. The elision of the first syllable, keresim resim is frequent in derivatives of kera and is found for instance in Rajasthani rau kerau. Resi (m) thus represents the survival of a definite case form and is not an adjectival postposition: in many respects it represents a more advanced state of affairs than tapaya. It is therefore also less accented and can never fulfil the function of a preposition as tapaya did in the Vasantavilasa-phagu. Resi(m), resammi is intimately linked with the noun that it governs: it has become a true postposition. 86 Both tapaya and resi(n) served at first to give geater emphasis to the meaning of a phrase than was possible by the simple use of endings. This emphatic use is particularly noticeable in tapaya, attested as it is at an earlier date. Gradually with frequent use this emphasis was weakened, and what had been a very expressive construction became ordinary; tanaya and resi(m) became mere postpositions, as usual as the simple endings. Encouraged by the weakening of the declension system, the use of these 'emphatic expressions' thus helped to hasten the further decay of the declension system. -98 SOME INDO-ARYAN WORDS MEANING ALL' L. A. SCHWARZSCHILD MOORE, VICTORIA, AUSTRALIA MUCH HAS BEEN WRITTEN on words meaning 'all' in the Indo-European languages, from Brug mann's important monograph to recent discussions of the later history of the Latin fetus and omnis. The main interest of these words lies in some peculiarities of declension and in the great variety of methods of expression. According to Bryndal's classification the idea of totality in volves four subsidiary notions: all. 1. Completeness: (Latin totus), whole, 2. Universality: (Latin omnis), 'all." 3. A distributive or iterative meaning: (Latin quisque), 'every, each." 4. Generality: (Latin quisquam), any." Languages do not necessarily distinguish between these notions, and even when they do there are frequent transitions from one subsidiary meaning to the other among words used to express totality. The liveliest, most expressive and on the whole most easily replenished group of these words is that which expresses completeness. Often adjectives of completeness tend to be used gradually to express universality, and later they may become iterative or be reduced to a vague general meaning. Such developments have taken place in IndoAryan as much as in Romance and elsewhere. The unaccented word sama was already in process of disappearing in Vedic. It conveyed the meaning of every' and 'any' and was sometimes a very weak indefinite pronoun, e. g., in R. V. 9. 29. 5b samasya kasya cit of any quoted by Wackernagel. There are indications from other K. Brugmann, Die Ausdrucke fur den Begriff der Totalitat in den indogermanischen Sprachen. Progr. (Leipzig, 1893-4). 8.J Andersson, Etudes sur la syntage et le amantique do mot francais tout. Etudes Romanes de Lund (Lund, 1964) V. Bryndal, Omnis et Tots: analyse et etymologie, Melanges linguistiques offerts a M. I. Pedersen. Acta Jutlandica, Aarskrift for Aarhus Universitet (1937), PP 200 288. J. Wackernagel, Altindische Grammatik, III (Got tingen 1830), 582. Indo-European languages and particularly from Avestan (Avestan hama all") that this word had originally designated completeness and univer sality. The Vedic adjective ladeat ever-recurring sometimes had the meaning of 'all' and 'every, 4. g., in R. V. 7. 18. 18 lavastas latravas all the foes.' saivat is even found in the phrase faivallu riky in all the habitations which is clearly su viku. In the classical language the adequivalent to the more usual Vedic expression jective faseat is obsolete but it survives in adver bial form an aleat 'always' Viva like sama was probably at first an adjec tive meaning "whole.'s In Vedic it covered all the subsidiary notions of totality (except to some extent that of generality) and was much used. In later Vedic and the Brahmana period riva lost ground gradually to sarea and was restricted to fixed locutions and archaising phrases in classical Sanskrit. Sarva had originally been yet another adjective indicating completeness as is shown by the cognate Avestan haurea, English whole, etc. It was the most widely used adjective of totality in Sanskrit and covered the subsidiary notions. Despite the virtual disappearance of the Vedic sama, viera and the adjective faivat, there was no shortage in Classical Sanskrit of words meaning all though they expressed mainly the subsidiary notion of completeness. The most important of these words was sakala, which has sometimes also assumed a universal and distributive meaning. More definitely restricted to the sense of completeness were smarta, samagra, akhila and sampurna. Middle Indo-Aryan in this respect, as in other features of vocabulary resembles Sanskrit rather than Vedic. Descendants of visa therefore occur but rarely and then only in specialised usages: 13 J. Wackernagel, ibid, refers to Zuhats I. P. XXV, 201 See also J. Gonda, Reflections on Barra in Vedic Texte, Chatterji Jubiles Volume, Indian Linguistics 16 (1955). 99

Page #60
14 SCHWARZSCHILD: Some Indo-Aryan Words Meaning All rissa< viva is found in some archaising formulas in Pali In Jaina Prakrit it is usual only in proper names, eg Vissasena and Vissabi, which are mentioned a number of times in the Svetambara canon. The form visa < vive is due to a sporadic sound change in Ardha-Magadht and Maharastri and occurs also in proper names in the canon, e. g., Visasena in the Sayagadanga, 1.6.22. As in Sanskrit there is no sign in the literary Prakrits of descendants of Vedic sama and of the adjective salvat. In Middle Indo-Aryan it was the descendants of sarea that were most frequently used to express totality. Of the Sanskrit words that denoted the aubsidiary notion of completeness akhila does not seem to have been popular in Middle Indo-Aryan. It was used mainly in stylised works such as the Gandavaho where it occurs in the form akila and it was also occasionally used as a tatsams in other texts. Sayala, sagala < sabala was more common, but in quite a few cases it also appears as a tatsama or semi-tatsama. Sanskrit samasta became samattha in Prakrit by the ordinary phonetic development. This word was too much alike in form and meaning with samatta < samapta com plete,finished not to be confused with it. How complete this confusion was can be seen from the fact that Hemscandra derived samatta from samasta and endeavored to explain this by an unusual change of st to -- instead of tth(Hemacandra 2. 45). As a result of this confusion the Prakrit derivatives of samasta did not become important as adjectives of totality. Sampussa < sampurna was still felt as a compound of purna fall and was therefore kept more in the specialised sense of complete, 'full' rather than as an ordinary adjective of totality, while samagga< samagra also does not seem to have become usual or popular. There was thus in Middle Indo-Aryan only one word that covered all the subsidiary notions of totality: Prakrit sees, Pali sabba, while sagala and some less usual words ahila, samagga and samalta indicated completeness. Late Apabhramia texts show the emergence of a new adjective of The Pall formula vissen dhamman quoted in this respect by J. Wackernagel, loc. cit., was even misunderstood by the commentator of the Dhammapada, unknown was viss rierom. Cl. W. Geiger, Pali Literatur und Sprache (Strasburg 1916), $118. I. Pischel, Grammatik der Prakritsprachen (Strass burg 1900), $64. totality, sahu. This word is given by Hemacandra in the rale sarvasga saho (4. 366) shu is optionally substituted for sarea. It also occurs in a verse quoted under this rule. This verse is almost identical with verse 88 of the earlier Pahudadoba of Ramasimha, though there sayalu is used instead of alku. Sahu is found a few times in the plural but on the whole it is an adjective of completeness occurring mainly in the singular and it is not much declined. Savve, sarvi sarve is used for the plural and also indicates the subsidiary notions of totality. In some later texts, such as the Paumasiricariu of Dhahila and the Carcari of Jinadatta sahu appears with a short a and the Old Gujarati form corresponding to it is sahu. There is still some uncertainty as to the origin of this word. The most widely accepted theory is that of Pischel who derives athu from Sanskrit saival. To strengthen this theory Berger 10 quotes an intermediate form ass. Tessitori," Modi and others follow a similar explanation to that of Pischel. Hemacandra equated the word aliu with sarsa, and this view was taken up again by Dave" who derived the Old Gujarati sahu from Sanskrit sarsa, Pkt. sarva, Apabhramin save, with the addition of hu, a derivative of the Sanskrit particle khalu used for emphasis. Although it has gained the support of so many scholars the theory that anal became adhu in Apabhrampla has a number of weaknesses. Some of these are phonetic. Berger (loc. cit.) stated that the lengthening of the a was not clear, but there are plenty of parallel cases (eg, those quoted by Pischel, loc. cit.), and it seems that the simplification of a sibilant plus -- or -- with compensatory lengthening of the preceding vowel was not unusual in Ardha-Magadhi, Jain Maharastri and Maharastri (cf. visa for viss quoted previ a. V. Tagare in his Historical Grammar of Apo Mana (Poona 19481, p. 447 quotes salu as cecurring verse so of this text, bas . L. Jain's edition bas R. Pischel, op. cit. 1864 and 202. H. Berger, Zuel Probleme der mittelindisches Laut lekre (Munich 1955), p. 80. "L. P. Tessitori, Grammar of the Old Westers Rajasthani Indians Antiquary, 1914-16, $98. Gurjarardalselt. Edited by Thakore, Desal and M. C. Modi (Barols 1956). Glossary, .. sahu. T. N. Dave, A study of the Gujarati Language is the 16th century (London 1935), p. 35 and p. 190. -100 SCHWARZSCHILD: Some Indo-Aryan Words Meaning All ously). More serious is the objection that athu, sahu appears always as such. There is no indiestion anywhere that the -- had originally been a sibilant although the word is particularly well established in areas where the change of a to a made least impact (Gujarat). One might argue that this was a case where the change of a to a was facilitated by the presence of another sibilant which caused dissimilation. The most obvious example of such a development that could be cited is the verb ida- to tell which nearly always appears as a rather than ade- in Middle IndoAryan. Such a dissimilation is not impossible for baleat>alsa. Nevertheless the complete absence of forms of this word with a simple - or of any other intermediate form makes the derivation of aths from faivat unlikely. Pischel and those who followed his explanation were concerned mainly with phonetic development and did not state clearly whether they thought of the Vedic adjective laivat-ever-recurring, all or of the Classical Sanskrit adverb saivat 'always' as the antecedent of athu. In either case there are difficulties. Apabhrapia shu is rarely declined, and this might lead one to regard the Classical Sanskrit adverb faivat as its most probable source, but the change of meaning from always to complete, entire, all' cannot claim widespread support from parallel developments. The Vedic adjective lavat-all' offers no semantic difficulties,, but is even less probable as a direct ancestor of the late Apabhrain aths; so-called Vedic survivals that never penetrated into the literary works of Sanskrit, Prakrit or even early Apabhramia are open to grave doubts. 15 The difficulties that lie in the way of the more popular explanation do not however vindicate the alternative theory put forward by Dave. The phonetic development of sarea + khalu to athu is not clear, particularly as there is no indication that such a combination was frequent either at the Prakrit or the Apabhramla stage of Indo-Aryan. Sahu, sahu makes its appearance so late in Apabhramula that it may well be a formation within Apabhramia itself rather than an inherited form. There is evidence in the Apabhramia texts to sup port this view. are signs of some phonetic irregularities in the Already in earlier Apabhrames literature there development of sarma 'all, probably owing to the frequent unaccented use of the word. There was sonant, semetimes, as one might expect, with reduction of the double consonant to a single concompensatory lengthening of the preceding vowel, of the Pahudadoha and the Bharisatta-kaha we sometimes without. In the Western Apabhrampia therefore find save as well as the phonetically regular sera, while the Paumacaria and the later Sandelarisaks have adva,advu. The nominative and accusative singular masculine and neuter is sometimes distinguished further by the loss of the single -- before the final -u, and we find stu" as well as adru in the Paumacariu. The introduction of a into alu Page #61
SCHWARISCHILD Ideas Word Meaning AB SCHWACHILD: Some Indoren Wonde Meniny having abandoned the wrice of the king which the car that in spite of the inter (apparently) s www boos, but which in stand development of the the dris M eand which is rally allo w plate from all in Ape (Alsdorf:in Wahrheit Leide ). One lodo Aryan follows pro c heme of can how do this type of phrases to Indie n ts which wonde mening all are partially is where Hindi s t antly supplemented and stimately sup followed by it no s , i r the whole planted by wondering whole or complete night (Padumst edited by Lam Dhar 442) Bhayani has shown to be the needent of the Western Rajasthani and specially in Gujarati moder Hindi na LP . T or t ed with the addition of the emphate particle that the Old West Rajasthani in t e hanc Although it ultimately served no Raba, well and the were based only in the limited (et modern Gujarati on > Old Western Rajathani aka The ithalat tree of informace where change of the nominative-se mainen kit wala whole' all'appears as and all tow el to all m ein Apam but in Om Wester these development and due to the info of the more regently whale. The The metion of word h lration and Dare to explain this word from the rest of the declension of derived from Sant Age w orthy Bein singular wu usturally more to which involve highly improbable change of es completeness, while the plural meaning. It w as mach more posible that the are expressed w i ty. Ultimately this pirated and we brought about by the co have given the speaker the illusion, shared by lance of the word gala whole with a Pischel and other that show quite separate of identical meaning. It is also possible that adjective of completeness, conted with similar infance out for other irregular and that its introduction into tenery pirations in words mainall. Modern Inde. Apabhram had been favoured by the compan Aryan A p pene over large part of tive Benity of surviving adjectives of totality. Northern India as Buch forme including the We have this the curious ce of an old adjective r a mbha ou already in the text of the of universality and totality in general, D A VIL K. Chatterji gested that this s splitting into two to form av et t be due to the ince of the ta b le All and suwhole bly, but the possibility remains that the One of the peculiarities of both in and Anabrumin d h uwu the of this is the lack of declension. This has nothing to do supirate consonant too. with any posibility of a derivation from a hole arti ve only in it is merel feature of the adjectives of totality limited ar e would expect to find in There are correspondences to this in other lan Modern Inde. Ara the introduction of some se gun , French. The age originated from adjective indicating the notion of completeness the tendency of the word for all to imply alto happened with the use of derivatives of ether or nmpletely and to approximate in the Sanskri t of Hindi Bell Panjabi, this way to an adverb. This come about especially and ete whole, all. The beginning when the word for all wed with another of the dual change of m ing of Sanskrit adre adjective or past participle. The beginning of art to whole all' is already such a development can be for instance in the bottle in A bram in the Kumlis Paumscriu VII. 7. 4: b inuhos lantbatha of Nomaprabha (Sthalabhadraka the whole army was turned back of the army the 1 1 in Aldorf's edition) was completely turned back Lank of declension contion in the modern Indo-Aryan ang chanuariy-blagga-sar und noted by Trastori it was particularly iddum sajalahi thir; Boticeable in Old Western Rajasthani; in Modern the best s Gujarati t r the coral resort to the bitter a in fact was not frequenti a of its beauty (or all its beauty has been 1-2: to have din importance in oldered by her lips and ma-ma ca nister HylVar upp pirasidomos dula More B ett (1955) There are plastic for the for by MC Moll y . ChalO Delof the AB , where that th e Lone Caltha 10). 18 State T his formas de sorte Sort u ri 1990). Gary, the one or for , The Be a te with it , but they last Allah . wyd - 102 -103.

Page #62
THE INDECLINABLE JE IN MIDDLE INDO-ARYAN L. A. SCHWARZSCHILD Australia "The emotive aspect of word-meaning is even more variable and unstable than its cognitive side"-in this and similar statements it has often been pointed out that those parts of speech which are emotive, emphatic particles and exclamations in particular, are very variable in their use. They do not form part of any regular system, as do declensional and conjugational forms and they are therefore more liable to sporadic alterations: They are sometimes used too much and then fall out of favour again according to the demands of fashion and even according to individual taste. This is shown by the loss in English of expressions such as "verily", "in sooth", and exclamations such as the angry "sirrah". Similar and constant changes are noticeable even in a country of profound respect for the past such as India. The study of this ephemeral type of usage is beset with difficulties as can be shown by an examination of the Middle Indo-Aryan particle je, used with the infinitive. The type of infinitive that occurs most frequently in the Svetambara Jain canon is that in -ttae. The infinitive in -um, -ium derived from the Sanskrit (i)tum is less usual, but gradually becomes more frequent in the later portions of the canon. It is sometimes associated with the particle je as in bhareum je, kareum je found in the Uttaradhyayanasutra (19,40). There are also isolated examples of the infinitive in -um followed by the particle je in the Panhavagaranaim, and there is one instance in the Nayadhammakahao (I. 9.). This usage persists in some of the later Jain Maharustri texts, such as the Paumacariya, the Samaraiccakaha, the Dhurtakhyana and the Kuvalayamala, The usage is particularly frequent in Jain Sauraseni, for instance in the Bhagavati Aradhana and the Malacara. On the other hand the infinitive with je is missing from Some of the other narrative texts, notably the Vasudevahindi and the Jambucariya; it is practically unknown in standard classical Maharastri and occurs only once in a text of the extent of the Ravanavaho (kaum je, 4.36), and also once in Hala and then only as a variant reading, while it is completely absent from the Vajjalaggam. The fact that this usage was not generally current in Maharastri, the prakerstam prakrtam, led to its omission from the grammar of Vararuci, and it therefore also tended to be ignored by the modern grammars. Hemacandra mentions the emphatic particle je (II. 217), but not its 211 -104 LA SCHWARZSCHILD use with the infinitive. He does however give an example of the use of je with the infinitive, the example being the variant reading from Hala. In modern grammars the special use of the particle je with the infinitive is barely mentioned anywhere except in the works of Schubring, S. Sen and A. N. Upadhye and H. C. Bhayani. One of the most striking features of the particle je when used with the infinitive is that it does not at first sight appear to add to the meaning of the sentence. Thus in the Jain Sauraseni of the Bhagavati Aradhana (p. 362 of the edition in the Munisri Anantakirti Series), one finds a succession of phrases containing infinitives, some with the particle je and some without it: paso va baddhidum je, chittum mahila asi va purisassa; sellam va bindhidum je, pasko nimajjidum mahila, 'woman is like a fetter for tying a man, like a sword for hewing him down, like a dart for piercing him, like a quagmire for drowning him'. This alternation is noticeable even in the earliest texts where the infinitive with je is found, e.g. in the Uttarddhyayanasutra 19, 40: jaha aggisiha ditta paum hoi sudukkara, taha dukkaram kareum je, taruppe samanattanam, 'just as it is difficult to control a fiery flame that has been lit, it is difficult to be an ascetic in one's youth'. It becomes clear even from a few examples that the infinitive always precedes the particle je, which has therefore been considered enclitic. In fact the position of je is nearly always final. It occurs at the end of the sentence in prose, e.g. in the Kuvalayamala (p. 111 of the edition in the Singhi Jain Series): ta na juttam mama jivium je 'therefore it is not right that I should live', and p. 108, jalanihi na tirae langheum je 'it is impossible to cross the ocean'. In poetry the characteristic position of je is at the end of a verse or half-verse: in the examples quoted from the Bhagavati Aradhana and the Uttaradhyayanasutra three of the half-verses end in baddhidum je, bindhidum je, and kareum je, but there is no je after chittum, nimajjidum and paum, where it would not be in a final position. Seeing that it concluded a phrase, je was never used with an infinitive that depended on a following adjective or past participle. This means that it is completely absent in the frequent type of phrase kareum payatto (or pautto), kareum samadhatto, 'having set out to do something". and kareum vavasiya 'having decided to do something'. It is similarly excluded from any other type of expression where the infinitive is closely linked with the following word. It therefore becomes apparent in both prose and verse that the particle je when used after the infinitive conveyed a final emphasis: the particle singled out the infinitive from the remainder of the sentence. As is so often the case with particles, je soon lost its empha 212 -105

Page #63
THE INDECLINABLE JE IN MIDDLE INDO-ARYAN tic significance, and began to be treated as a mere adjunct to the infinitive form. Although in its earlier appearances it had been exclusively used with the infinitive as such, and only in a final position, it occurs in the later Jain Maharastri of the Nanapamcamikaha with an infinitive form that is actualy used in the meaning of an absolutive, kaum je (L. 79). It was also extended to other positions in the sentence. In Apabhramsa and in texts influenced by Apabhramsa je appears in the weakened form ji and figures as a mere expletive anywhere in the sentence, mainly after pronouns, particularly in the expression so je, so ji, equivalent to 'he indeed' or 'just he'. The word is therefore listed by Hemacandra as a mere expletive (je in 2.217, and ji in 4.420), and it has survived as such in the modern Gujarati -1.7 The original emphatic meaning is in keeping with the etymology of the particle je. There is no doubt whatever that je developed from the Prakrit jeva, jjeva 'thus', 'indeed', and that this form jeva came from the Sanskrit eva 'thus', 'actually', 'indeed', which was generally used in Sanskrit to lend emphasis to the word that preceded it. It is the details of the irregular phonetic development of eva to jeva and je that are most complex. In order to retain their emotive and emphatic value, particles tend to be reinforced. There are two major ways in which such reinforcement takes place, and this can be illustrated from IndoAryan as well as elsewhere: 1. New sounds are added to the particles to strengthen them, and the source of the new sounds is not always clear (e.g. Latin tantum gives Old French i-tant 'so much"). 2. Two or more particles are heaped together so that their cumulative force is felt, as for instance with the postpositive particle vai which forms ha sma vai, and ha vai in Vedic and later eva vai, api vai, tu val etc. Similarly the Latin particle sic 'thus' is strengthened by in to give in sic, whence Old French ensi, modern ainsi 'thus'. Both these types of reinforcement have been suggested by scholars in attempts to explain the development of eva to jeva, je. According to those who favour the first theory, jeva represents the normal phonetic development from an earlier word yeva 'thus which is found in Pali and in the Asokan inscriptions. Geiger suggested that the initial y of this word was a sandhi consonant, which was inserted before eva if the preceding word ended with a vowel or with anusvara. He quoted other words which he thought showed the introduction of such a sandhi consonant in similar circumstances. In the case of yeva jeva, je this would agree quite well 213 -106 L.A. SCHWARZSCHILD with the usage found in the Jain Maharastri and Jain Sauraseni texts where je generally follows the anusvara of the infinitive. There are cases where other consonants appear to have been inserted before eva: there was a tendency in Asokan Prakrit to add h before this word, to and hevam, hemeva therefore appear occasionally, mainly in inscriptions of eastern provenance. The main weakness of the theory of Geiger is that there are no true parallel examples: of the insertion of a sandhi y in Middle Indo-Aryan at the beginning of any word other than eva. As pointed out already by Edgerton1 the parallel cases quoted by Geiger are in fact to be explained by quite different causes, and the Pali form yeva Prakrit jeva, je therefore stands quite isolated. There have been attempts to explain the Prakrit word jeva, je according to the second process of reinforcement which is so frequent in the case of particles, namely the accumulation of particles. According to this theory jeva is not directly connected with the Pali yeva, but is derived from two particles, ca and eva which normally gave ceva, ccia.12 The Asokan forms hevam etc. could come from a similar accumulation of particles, ha+eva heva. There is no doubt that the compound particle ceva, ccia was in fact often used in Prakrit for emphasis; there is even one instance in the text of Hala (v. 524) where there is hesitation between je and ccia after an infinitive. The parallel between the two forms is continued later, when ceia > Old Marathi ci survives as the modern Marathi -e, while je, Apabhramsa ji survives as the modern Gujarati -j. But it is probable that the similarity arose mainly from the fact that ccia, ceva ca eva and jeva, je were both derived from eva. The close resemblance between ceva, ccia and jeva, je does not necessarily imply that there was a direct link between the two words and that jeva was in fact a derivative of ceva ca+ eva. The derivation of jeva from ceva is quite possible phonetically, as weakening of the initial e to i would not be out of place in an enclitic particle, 13 The main difficulty of this explanation however is that the Pali word yeva, which appears to be inseparable from the Prakrit jeva, je, is left unaccounted for: Pali yeva cannot be derived from ceva Page #64
THE INDECLINABLE JE IN MIDDLE INDO-ARYAN LA SCHWARZSCHILD tions for instance there are on the one hand some relative pronouns and derivatives of relative pronouns which have lost their initial y : ata < yatra, ath Page #65
THE INDECLINABLE JE IN MIDDLE INDO-ARYAN irregular development from the Sanskrit word eva 'indeed', via the intermediate forms yeva and jeva. It was most unstable in use, varying from an emphatic form to an expletive; on the one hand it could be combined with ham as hanje, a slightly disdainful form of address, and have a highly specialised use, and on the other hand it could be used on its own, as in later Prakrit and Apabhramsa where it occurred very generally, and was practically meaningless. NOTES 1. S. Ullmann, The Principles of Semantics (Glasgow, 1951), p. 104. W. Schubring. Das Mehanisihasutta (Berlin, 1918), p. 92. 2. 3. S. Sen, A Critical Introduction to the Pashanigarandi P. 13. 4. Dhurtakhyana, edited by A. N. Upadhye (Bombay, 1944), p. 53. 5. H. C. Bhayani, "Some interesting features of the Prakrit of the Nanapamoamikaha". Bharatiya Vidya XII (1951), p. 161. 6. There is one example in Buddhist Hybrid Sanskrit of the form yeva being used in this position, gantum yevadhyavasito "he is determined to go', which is quoted by Edgerton, Buddhist Hybrid Sanskrit Dictionary (New Haven, 1953), s.v. gera 7. T. N. Dave, A Study of the Gujarati Language in the 16th Century (London, 1935), glossary s.v. ja. 8. R. Pischel, Grammatik der Prakritsprachen (Strassburg, 1900), p. 229. 9. W. Geiger, Pali Literatur und Sprache (Strassburg, 1916), 66. 10. J. Bloch, Les Inspriptions d' Asoka (Paris, 1950), p. 52. 11. F. Edgerton, Buddhist Hybrid Sanskrit Grammar (New Haven, 1953), p. 37. 12. H. C. Bhayani, "Prince of Wales Museum Stone Inscription from Dhar", Bharatiya Vidya XIX (1959), p. 118. 13. ef. the occasional weakening of the particle on to yn in Prakrit, Pischel, op. cit., p. 135. 14. M. A. Mehendale, Historical Grammar of Inscriptional Prakrits (Poona, 1948), P. 189. 15. J. Bloch, L'Indo-Aryan (Paris, 1934), p. 200-201. 16. Further examples are given by R. Pischel, op. cit., p. 229. 17. R. Pischel, op. cit., p. 190. 217 (Wurzburg, 1939), -110 NOTES ON SOME WORDS MEANING "IMMEDIATELY" IN MIDDLE INDO-ARYAN By L. A. SCHWARZSCHILD ONE OF THE major advantages of the study of synonyms is that it provides a link between descriptive and historical linguistics. In Professor Ullmann's words: "Synonymic structure implies the formation of associative groups between senses on the langue level. The firmness and constancy of such associations can best be gauged from the diachronistic developments which they may initiate." Among the first words to be studied from this point of view were the Middle English adverbs which meant "swiftly" and later "immediately" Words of this meaning are of interest also in Middle Indo-Aryan where they form a closely-knit group. This is proved by their formal as well as their semantic development. Probably the most important words of this group are the derivatives of Sanskrit sahasa. Being the instrumental case of sahas "force", sahas was used adverbially and meant in Vedic "forcibly" and hence "suddenly", "all at once"." on the spot "a In late Classical Sanskrit it developed the further meaning of "rashly", but this late meaning is not reflected in Middle IndoAryan. There is at first sight nothing curious about the development of sahasa in Middle Indo-Aryan where it becomes quite normally sahasd, with optional shortening of the final - as usual in adverbs. But on reading a well-known Maharastri text such as the Vajjalaggam one is surprised to find that sahasa is followed by the particle tti from Sanskrit iti "thus". The same characteristic is found in other Maharastri texts, e.g. the Lilavaikah, where sahasa is followed by -tti whenever it appears with a short final, and this sahasa-ti is more usual than the simple sahasa. Sahasa-tti is very frequent in Jain Maharastri texts and is used almost to the exclusion of sahasa in the Apabhramea texts of both Southern and Western 18. Ullmann, The Principles of Semantics (Glasgow, 1951), p. 111. G. Stern, Swift, swiftly" and their Synonyms. A Contribution to Semantic Analysis and Theory. Goteborgs Hogskolas Arsskrift, xvii (Goteborg, 1921). For a similar semantle change ef. Buddhist Sanskrit pratiyaly era quoted by L. Renon, JA, cexlvii (1959), p. 261. R. Pischel, Grammatik der Prakritsprachen (Strassburg, 1900), SS 113. Vajjalagga, ed. J. Laber, Bibliotheca Indica (Calcutta, 1944), v. 388. -111

Page #66
SOMS WORDS MEANING THE DEATRLYIN MIDDLE INDO-ARYAN 1 40 ONS WORDS BEASISODELATI I DDLE INDO-ARYAN provenance, eg, in the Harivamparina of Puspadata, and the Paumastricarit of Dhabils. Asthi is an enclitic particle one might assume that it has simply been added to lift the preceding word out of its context for emphasis, so that what might be translated as very suddenly It is tempting to compare solas-co v ery suddenly which is found occasionally in Jain Maritti, e in the gate Erkungen edited by H. Jacobi. But such an explanation is impossible. Even in those dialects where was is so sul eg. in the Mihrist of the Lilavalkah there has been no extension of the use of thi. Apart from its occurrence with it appears only where Sanskrit would have it, that is to mark the end of any quotation of names, words, thoughts, or sounds, Nowhere is there any laxity in the use of the which might explain the formation of sahati. The origin of nasi becomes clearer only when one examines some of the other words meaning " suddenly" *immediately" in Middle Indo-Aryan. The word which most obviously springs to mind is Sanskrit jhaili, Middle Indo-Aryan jhadatti, laditti, and latti, suddenly," "immediately," but the history of this word in Indo-Aryan is also not simple. It is generally agreed that lait is derived from har "knock": sudden movement", with the addition of the particle ini. Such is the view of Walernagel, Turner, and Mayrhofer."What is presumed to be of onomatopoeie origin. Some difficulties are involved in the assumption that jhaiti was formed within Sanskrit itself. In Sanskrit the combination with ili was peculiar to jhaiti and the rare variant jhagiti and was otherwise contined to definitely imitative words where the quotation with it was justified, as for instance in indad "bang". Such words had no meaning apart from the idea of noise that they conveyed and did not form part of the normal vocabulary. Jhait, on the other hand, did not convey any idea of noise in Sanskrit: it was an ordinary adverb used in much the same way as head, anandar, etc. The isolated position of jail in Sanskrit and its late appearance make it probable that this word was remodelled borrowing from the populer languages and that it was formed in Prakrit as jhadatti, the form in which it is first attested. Jhalatti was by no means a unique formation in Prakrit. Already in some of the comparatively early parts of the Svetambara Jain canon, such as the Naydhammakahao, there appears a word dhati which is used in connection with a heavy fall. This word is linked with the Prakrit verb dhani "to fall", modern Hindi dhanna" to collapse". Dhasi may have been an independent onematopoeie formation, or it may be derived from Sanskrit dh a ti > Prakrit dha )" to fall", under the influence of parts of the verb that had no nasal, such as the past participle passive dhasta Whatever its origin, hasi was certainly felt to be onomatopoeie in Middle Indo-Aryan and meant to fall with a heavy thud". This fact is illustrated further by various other highly expressive derivatives : d ekke a thumping noise", dhasakkei "to make a thumping noise". Hindi dhasa "to sink (of land)", and Priskrit dhasatte" to topple over", which is found in the anonymous Maoipaticarita of the eighth or ninth century Dhasa-ti thus literally meant making the sound dhaa", ie," with a crash and it shows every sign of colloquial and possibly slang origin. Dhati is frequent in the language of the Jain canon and there is u fixed formula dhawuhi dhanawiyalasi cewchim wivedic "he full flat on the ground with a crash". This cours, for instance, in the Niraya valiyao, where it refers to Kopika who is smitten by remorse on realizing that he has caused the death of his father. Sometimes one may be justified in assuming that hasati conveys the ides of speed as well as noise, as, for instance, in another example from the Nirayavaliysio v. 12: si Roll Devl... pardusiyatlari compaglayi dati dharaniyasi x adiya, "Queen Kali fell flat on the ground suddenly, like a campala-creeper felled by an axe. In this case it is clearly the idea of suddenness rather than that of a heavy thu that is being emphasized, and dhaaott might well be translated by suddenly" or "straight away". Jhadat is very similar in formation to dhasti. It is connected Wackernagel, Alice Gri d Gottingen, 1957), voli, & *R L Tur 4 Compare and cal Dictionary of the Nepali Los Joe (London, 1981)... Jej M. Mayrhofe, pe d e desen (Heidel berg, 1953), sati The pouble Mandarin of some of these compos tele red by Prof . J. Kupe, e in Zeitar Iw . 241. 1 See jaar , ed. K. Thelor, M. D. Desi and MC Modi (Baroda. 1956), slovary v. dhe The Prakrit Version of the Mesipi-Carie, ed. . Williams (London, 1978). SEL

Page #67
42 SOME WORDS MEANING SEDETILSTS MIDDLX DARYAX with the Prakrit verb jadai "to collapse", which is of onomatopoeic origin from "jhar "knock", "sudden movement" (see above). As in the case of dhani the onomatopoeie quality of the verb is continued in some of the very expressive derivatives, eg. Apabhirama Malappal "to snatch", Guj. jada pet; sladakka! " sudden onslaught" and its descendants in the modern languages. Any remaining doubts about the onomatopoeie value of this group of words are removed by the expression jalajulajladajala translated by Dr. Bhayani as striking so as to produce la ajali sounds". Jumattithus at first meant "making the sound jad" i.e. " with a crush", and it was exactly parallel in formation to dhasti. As in the case of dhastithe idea of noise or of a blow seems to have gradually disappeared in this word and it came to denote the sense of suddenly","immediately which is the only surviving one. Parallel developments for such a semantic change are not hard to find, e.g. French tout-d-coup. The Prakrit word jhadatti was given a Sunscrit garb and introduced into Banskrit at a comparatively late stage shariti, which does not figure in the Kidila. The Sanskrit jhafiti was subsequently borrowed back in some of the more artificial late Jain Mahliristri and Maharastri texts where it appears as jhaditi, e., in the Surasundarfcaria. There is yet one more form of this word, namely Shoti which occurs in Prakrit and Apabhrams and is shortened form of shalatti Under the influence of the word jhadatti and possibly on account of the use of the word dhanit suddenly", "immediately in the Jain canon the synonymous sahad suddenly","immediately took on a finali in Middle Indo-Aryan, although sahast had no onomatopoeic value whatever. The use of i would not have been justified in this instance by the ordinary rules of the Middle Indo-Aryan languages. In other words, i was treated almost like a suffix characteristic of adverbs meaning "suddenly" "immediately". Tadatti suddenly","immediately" is yet another adverb that formed part of this closely knit group in Middle Indo-Aryan but it seems to be a comparatively recent formation and does not appear until rather late in Middle Indo-Aryan. This word has sometimes been associated with it "lightning" by scholars, .. Permacare by Swayambhidea, ed H. Bhayani, vol. 1 (Bombay, 1953). SOME WORDS MEANING IMMEDIATELY IN MIDDLEIDARYAS 43 by Dr. Williams who quotes todittadiyalike lightning from the Aupapatikastra in support of this theory! Todotti suddenly immediately is probably connected only indirectly with Sanskrit fadir "lightning", which was not a very sal word in the popular language and had been largely replaced by derivatives of midyut, > Middle Indo-Aryan ju l , jul, etc. whence modem Hindi bijt, and by derivatives of mani. Tadatti appears to be exactly parallel in formation with the earlier dhati and with jhadat suddenly","immediately. It was derived from the verb tad to strike" which was felt to be onomatopoeie, as is shown by numerous onomatopoeie derivatives that are formed from it, .. Apabhramistada ka "crashing and fadatadatadants "breaking with the sound a fatalento of similar meaning is found in the Jain canon. Tadatli therefore meant originally " with a crushing sound and it actually occurs with that meaning in quite a few instances in Apabhramla literature, e.. in the verse quoted by Hemacandes IV 302 wys w antiae piu diffhow what addhiwalay withihiyo add ghuffalotti which is translated in the Bombay Sanskrit and Prakrit Series is "As the lover was suddenly sighted by a lady who was driving away the crows, half of her bracelets dropped down on the earth and the newt) half cncked with noise". Here tada ti with noise" is associated with mosallisuddenly " with which it rhymes, In practically all its occurrences todotti is used with either Bahasa or jhadatti, and the semantic and formal link between all three words is evident. Although fadatti conveys the idea of sound occasionally in the above example, in the majority of cases the idea of speed which is prevalent in nahast and jhadatta is paramount also in tadati. This can be seen from a verse which is quoted by Hemand very near to the one quoted above, i.e. in illustration of IV 357: hadapi adatti hari Milo-them kaimo heart, break at once for with a crush"), why do you delay " Often adatti has to be translated simply by suddenly" "immediately" as is done by Professor Alsdorf in the Harvamispania 86.3.2, where the word occurs in association with ladatti: plasir-plande tidai adatti padithalaitalappae Hari jhalatti 1. W opil. p. Har u n , Aldo Hamburg. 1990. 312 HR LT, it, where 115

Page #68
44 SOME WORDS MEANING IMMEDIATELY IN MIDDLE INDO-ARYAN "The snake struck suddenly, with its hood extended, Hari parried immediately with the palm of his hand ". In late Jain Maharastri where jhadatti occurs in the semi-Sanskrit form of jhaditti, tadatti is made to conform and appears as taditti, as, for instance, in the Manipaticarita of Haribhadra. The parallelism between jhadatti and tadatti is therefore complete. Historical evidence thus shows the existence in Middle Indo-Aryan of a closely associated group of synonyms meaning "suddenly", "immediately": dhasatti which is found in the Jain canon is succeeded by jhadatti, talatti, and sahasatti which continually influenced each other formally and semantically. In Apabhramas there were already forces at work which caused the gradual break-up of this system of synonyms. The group of words meaning "suddenly", "immediately" formed part of a much larger system, the adverbs of time. As they were gradually made to conform with this larger system their existence as a separate group was obliterated. This tendency gave rise to forms such as jhattihim (Sandelarisaka), where the adverbial ending him (based on the pronominal locative singular ending) was added to jhatti, probably under the influence of words such as tahim "there". "then", and enhim "now". Similarly in the Jasaharacariu one finds jhadavi, which shows a remodelling of jhadatti on the general pattern of adverbs of time composed with ri>. Il n'y a pas de doute sur le sens de ces mots bien attestes, mais leur interpretation etymologique souleve des difficultes et des problemes qui sont d'une portee plus generale et que la presente etude ne saurait qu'amorcer. Il s'agit des problemes qui entourent l'origine du theme verbal ghumm- et l'emploi du suffixe ira. LE THEME VERBAL GHUMM Le verbe sanskrit ghury-<< s'agiter, branler s, dont on se sert si souvent pour traduire ghumm- est inconnu du vedique et des Brahmapas et n'est atteste que dans l'epopee et dans les textes sanskrits ulterieurs. T. Burrrow (1) a suggere pour ghurg- une etymologie dravidienne tres convainA cante. Mais cette etymologie ne suffit pas pour expliquer le mot prakrit CCL-1 5 - 117

Page #69
66 ANNEE 1962 ghumm- qui ne saurait etre un derive regulier de ghurn. Les grammai riens prakrits, de Vararuci jusqu'a Ramasarman Tarkavagisa citent toujours (a gholai et parfois ghummai comme dhatvadesa ou remplacant du verbe sanskrit ghurg: ils reconnaissaient done tres bien que ni gholai ni ghummai n'etaient des derives reguliers de ghumSelon les regles phonetiques ghurg- aurait du devenir ghupp, ou ghunnen prakrit. Ce mot se trouve en effet dans nos textes, mais tres rarement; il y en a un seul exemple dans le canon jaina (ghunnanta, Panhava. garapam I), et des exemples tres peu nombreux dans des textes du moyen indo-aryen la tardif (Desinamamala, Prakrtapaingala). Le mot ghunnderive regulier de ghuro, a ete remplace completement par ghumm en maharastri et en apabhramsa. Ghumm- s'agiter, branler >> existe deja dans les textes canoniques des Jaina: on se sert du participe passe ghummiya dans la description d'un naufrage, Nayadhammakahdo IX: gurujanadipphavaraha sujanakulakannaya ghummamani viva, (le navire) agite comme une jeune fille de bonne famille qui s'est rendue coupable devant les yeux de ses parents. Si le mot ghumm- qui parait de si bonne heure en prakrit n'est pas un derive du sanskrit epique et classique ghurn-, il semble tout de meme appartenir a la meme famille etymologique, mais son origine immediate reste obscure et il faut l'examiner a la lumiere des autres verbes en -mm- du moyen indo-aryen. -mm- comme finale d'un theme verbal est inconnu du sanskrit (4) mais assez commun en prakrit. Souvent cet -mm- du prakrit n'est que le resultat des changements phonetiques reguliers, et surtout de l'assimilation -my--mm -my- figurait dans le theme du present de quelques verbes sanskrits qui appartenaient a la quatrieme classe. On pourrait citer a titre d'exemple: klamyati s'epuiser, etre fatigue qui donne en maharastri kilammai (Lilavatkaha). * tamyati s'epuiser, faiblir qui devient tammai en maharastri (par exemple chez Hala). damyati dompter (ses passions), d'ou dammai en maharastri (Uvaesamala). sramyati) se fatiguer, pratiquer l'ascetisme qui donne sammai en ardha-magadhi (Uttaradhyayanasutra 1. 37). Ce dernier verbe ne se laissait plus distinguer en moyen indo-aryen de sammai tire de samyati >, -118 GHUMMIRA, CHOLIRA AGITE, BRANLANT. 67 et kyamate a cote de ksamyate pardonner, souffrir. Ce double usage a laisse des traces en prakrit ou l'on se sert de bhamai et de bhammai, de khamai et de khammai. Le verbe intensif cankramate, cankramyate se developpe d'une facon toute semblable, d'ou cankamai, cankammai en prakrit. Meme dans les derives de klam- etc., on trouve parfois des formes a -m-simple surtout dans les parties du verbe ou-my- ne figurait pas en sanskrit et ou la consonne simple s'accordait avec les regles phonetiques : par exemple damiana absolutif de damai. Mais en maharastri c'est la forme du present ou-mm- etait de regle qui s'est etablie fermement dans ces verbes par analogie avec le present, temps le plus usite. Il y avait done en moyen indo-aryen un groupe de quatre verbes herites du sanskrit, a signification tres semblable, et dont le theme finissait par un -m-gemine: ce groupe comprenait kilammai s'epuiser, etre fatigue, tammai s'epuiser, faiblir, dammai dompter (ses passions), sammaise fatiguer, cesser >>. Trois autres verbes herites du sanskrit, bhammai et cankammai errer et khammai souffrir etaient souvent associes a ce groupe. Une consonne geminee -mm- tiree de -my- parait aussi au present du passif de quelques verbes en -m- final en sanskrit. On trouve donc en maharastri gammai < gamyate, rammai ramyate a cote de gamijjai, gamiai gamyate et de ramijjai ramyate, et dhammai < dhamyate, nisammai nilamyate entendre (ef. Vararuci VIII, 58). Mais il y a d'autres cas ou un-m- gemine figure au passif en moyen indo-aryen sans se laisser justifier par les regles phonetiques. Il s'agit de : summai (en maharastri jaina, maharastri et apabhrama)-sruyate() sentendre ; cimmai (ne se trouve que chez les grammairiens)= ciyate recueillir >>; hammai (ardha-magadhi, maharastri jaina) hanyate tuer>>; khammai (maharastri) khanyate ou khayate creuser >>; jammai (maharasti)=jayate, (atmanepada) naitre >>. a On a beaucoup discute sur l'origine de ces formes du passif. H. Jacobi (7) suivi par K. F. Johansson (s, a cru a l'influence analogique du verbe gam- aller, tandis que Pischel etait convaincu que les passifs en -mmetaient dus plutot a des formations denominatives (basees sur janman naissance, hanman meurtre >>), ou bien a toute une serie de changements phonetiques dont plusieurs (-->-m-;-m->-mm-) etaient peu communs a cette epoque: ruyate> "suvai sumai summai, 5. - 119

Page #70
68 ANNEE 1962 et de meme clyate> cimmai. L'influence des noms abstraits janman et hanman qui devaient donner en prakrit jamma et "hamma a ete pour quelque chose dans la formation de hammai et de jammai. Mais l'importance de l'analogie dans le developpement des themes verbaux et du systeme verbal tout entier en moyen indo-aryen est plus considerable que ne le supposait Pischel), Les theories de Jacobi ne cessent de convaincre apres plus de soixante-dix ans : le verbe gam- aller avait de la ressemblance avec han- tuer et khan- creuser >> (par exemple a l'absolutif gantina: hantuna, khantina, et au gerondif gantara: hantavra, khantarva) et l'on a done cree un nouveau passif hammai, khammai a l'imitation de gammai. Ce sont les themes nouveaux du passif hammai et khammai qui ont a leur tour entraine cimmai (ci-) et summai (Viru-), a cause de la ressemblance des formes du present actif: hanai, khanai correspondaient a cinai et sunai (10). Le developpement regulier des formes sanskrites du passif de ci- et de sru- aurait donne lieu a des confusions (11), c'etait donc la un terrain tout particulierement favorable au jeu de l'analogie. Dans tous ces verbes -mm- semble avoir joue le role d'un morpheme du passif, comme d'autres consonnes geminees dont l'usage s'est etendu en prakrit grace a l'analogie, ef. -pp-, -- Toutefois jammai, present du verbe jan- naltre en prakrit n'appartenait pas au passif proprement dit, c'etait plutot une forme de l'atmanepada comme jayate il naft en sanskrit, et il ne se laissait done pas distinguer des verbes de la quatrieme classe en -my--mm- qui etaient presque tous des verbes intransitifs. Hammai de han- tuer figure parfois comme present de l'actif aussi bien que du passif (cf. chez Vararuci VIII, 45 et Hemacandra, Kumarapalacarita VII, 79). Cet emploi est du probablement a l'influence du nom "hamma meurtre auquel il semblait se rattacher comme verbe denominatif. Il ne nous reste done que gammai, rammai, dhammai, nisammai (de nisamay-entendre) et khammai (13), cimmai et summai ou-mm- fit fonction de mor pheme du passif. A cote des themes verbaux en -mm- du passif et de l'actif qu'on a cites jusqu'ici, et qui sont tous herites du sanskrit il y en a d'autres dont l'origine n'est pas certaine.. Il s'agit de nissammai a etre fatigue >> (maharasti): nisammai se reposer (maharastri); pasammai etre tranquille (maharastri): timmai etre mouille (maharastri); -120 GHUMMIRA, GHOLIRA AGITE, BRANLANT. 69 thimmai (doublet de timmai en apabhramsa selon la grammaire de Markandeya); jimmai manger (chez Hemacandra, IV, 230); dummai se desesperer>> (maharastri et apabhrama); * ghummai s'agiter, branler (ardha-magadhi, maharastri et apabhrasa); K khummai plier (?), (ardha-magadhi, et chez les grammairiens Ramasarman et Markandeya); cummai embrasser (Ramasarman); pummai regarder >> (Markandeya). Les trois premiers verbes de cette liste ne devraient pas y figurer, mais c'est Johansson qui a groupe nisammai et pasammai avec dummai et ghummai, et il a tout explique par l'analogie. Il suppose que les participes passes du sanskrit nisanna de ni+sad-a s'asseoir, s'abaisser >>, et prasanna de pra + sad- auraient abouti en prakrit a un nouveau present nisannai, pasannai (qui d'ailleurs ne se trouve nulle part, ni dans les textes litteraires ni chez les grammairiens) et qui a son tour aurait abouti a de nouvelles formes du present nisammai et pasammai sous l'influence de "hannai, doublet (introuvable) de hammai, passif analogique de han- tuer >>. Il n'y a aucune indication dans l'usage qu'on fait en prakrit de nisammai et de pasammai qu'il s'agisse d'un passif, et la verite est beaucoup plus simple que ne le croyait Johansson. Nisammai et pasammai sont plutot des derives reguliers des verbes de la quatrieme classe ni+lam-a se reposer >>, et pra+Sam-etre tranquille >> comme l'ont toujours cru les commentateurs. En maharastri on trouve aussi un present nissammai, parfois nisammas, tire de ni + Sram-etre fatigue qu'on ne saurait rattacher a ni+sad- non plus. Il y en a des exemples assez clairs en maharastri, ef. Lilavaikaha, 1082: pamhantara-parigholira-nayana-nisammanta-vayanaimles visages se fatiguaient a cause des yeux qu'on voyait rouler entre les cils >>. Ici l'interpretation par ni sanna de ni+sad- ne saurait donner une signification sensee et le commentateur jaina a donne le mot tamyade se fatiguant comme traduction sanskrite. Les grammairiens evidemment etaient du meme avis: ni(s) sammai, nisammai et pasammai ne figurent pas dans leur Dhatvadesa et ils les consideraient donc comme des derives reguliers des verbes sanskrits ni+ram-, ni+sam et pra+ sam-, qui n'avaient rien a faire avec nisanna et pasanna (1) Les theories de K. F. Johansson, qui a tout explique par l'analogie, ont fait tomber en discredit les resultats fermement acquis par H. Jacobi, -121

Page #71
70 ANNEE 1962 et c'est cela qui explique probablement pourquoi R. Pischel a rejete d'un seul coup la theorie de l'analogie toute entiere. Il n'y a pas eu de changement systematique d'un -n final en -m dans les themes verbaux, ni par des regles phonetiques extraordinaires comme le supposaient S. Goldschmidt et un peu Pischel, ni par l'analogie, comme le supposait Johans son. On doit donc chercher ailleurs l'explication de dummai, ghummai et des autres mots de notre liste. Johansson a tente d'expliquer le verbe prakrit dummai e se desesperer >>, par les changements analogiques qu'aurait subis le verbe sanskrit dubruler >>. Mais il n'y a aucune ressemblance entre dummai et le verbe passif summai, tire de sru- entendre >>. Les formes intermediaires manquent totalement. La consonne geminee de dummai ne semble nulle part jouer le role d'un morpheme du passif comme dans summai; c'est dummijjai qui sert de passif (chez Hala). Il faut donc rejeter l'etymologie dummai < du- briller>>. Le participe passe dummia desespere qu'on trouve chez Hala est frequent en prakrit et en apabhramsa et c'est probablement la forme la plus ancienne du verbe dummai. Pour l'expliquer il faut partir du mot sanskrit durmanas prakrit dummana qui a l'esprit malheureux, solution qu'avait deja entrevue Weber (14), mais sans s'y fier. En prakrit on avait cesse de considerer dummapa comme un compose de manas esprit : dummana ressemblait plutot a de nombreux adjectifs verbaux en -ana tires du sanskrit. Les adjectifs en -apa, suffixe actif, s'opposaient nettement au participe passif : karana qui fait jalana qui brule >>, harana qui emportes, karia fait >>; jalia brule; haria emporte On a interprete dummana comme un adjectif verbal et de la on a tire un participe passe dummia attriste, et un verbe dummai se desesperer>>. Une serie complete de doublets phonetiques porte preuve a cette interpretation pouvait tomber devant une consonne en allongeant la voyelle precedente (Pischel, SS 62). On trouve donc dumana (ardhamagadhi), d'ou dumia attriste et dumai se desesperer (maharastri). Churn-n'etait pas un mot herite en sanskrit, et son origine dravidienne l'exposait a des changements extraordinaires, puisqu'il n'avait pas d'appui dans la langue. En outre -nn- n'existait quasiment pas en prakrit comme finale de theme verbal. Le participe passe ghummia qui devait remplacer ghunnia>. -122 GHUMMIRA, CHOLIRA AGITE, BRANLANT. 71 Ghummia (ardha-migadhi) est donc la forme la plus ancienne du verbe ghumm- et c'est de ce participe passe qu'on a tire le present ghummai branler >>, si commun en maharastri. Les deux verbes ghummai et dummai restaient lies etroitement en moyen indo-aryen et ils devaient se rapprocher peu a peu dans leur sens : en apabhramea dummia signifiait trouble>>, agite>> (au sens figure), et plus tard, en hindi moderne damna * s'agiter, branler est devenu a peu pres synonyme de ghumma branler, tourbillonner >>. Khummily)a plie paralt en ardha-magadhi. Ce mot ne se laisse guere separer du participe passe khunna opprime, ecrase (du sanskrit kunna > qui se trouve chez les grammairiens tardifs Ramafarman Tarkavagisa et Markandeya. Il n'y a aucun doute que ce mot ait existe dans la langue parlee, car il a survecu dans les langues modernes, parfois au sens figure, ef. nepali khumcinu se retrecir, etre deprime (Turner, Nepali Dic tionary, p. 24). Ces trois mots en -umm- si etroitement lies par leur origine, ont ete rejoints par pummai voir (apabhramsa) d'origine obscure, et par cumbati embrasser devenu cummai par assimilation de la consonne b. Cummai n'est connu que dans le moyen indo-aryen tres tardif (Prakrtapaingala, Ramasarman Tarkavagisa) et dans les langues modernes (hindi cumnd, panjabi cummad, etc.). Probablement vers la meme epoque lambate lambai se pencher, sombrer>> est devenu lammai par l'assimilation du b, (cf. les derives, hindi lamaknd, etc.). Ce mot ressem blait trop au groupe dummai s'affliger, etc., pour ne pas en subir l'influence, et au lieu de lammai l'on devait done dire "lummai, qui a survecu surtout au sens de << faiblir>>. C'est la l'explication du verbe hindi lamna faiblir >>. Le verbe prakrit ghummai fait done partie d'un groupe de mots qui ne s'est etendu que peu, mais qui est reste uni par des liens etroits jusqu'a l'epoque moderne. L'origine du theme verbal ghol- s'agiter, branler reste obscure, mais on s'accorde en general a le considerer comme apparente a ghumms'agiter, branler >>. LE SUFFIXE -IRA. En sanskrit le suffixe-ira s'employait avec certains verbes pour former des adjectifs au sens de qui a l'habitude de... >>. Dans plusieurs cas * -123

Page #72
ANNEE 1962 c'etaient des formations tres anciennes et a peine reconnaissables par exemple isira vigoureux d'is- animer et peut-etre sithira mou de 7th-(?). Mais le suffixe restait bien vivant et servait a former des mots nouveaux en sanskrit (cf. mucira genereux >>, qu'on trouve dans les lexiques, et rucira << lumineux >>). Le suffixe -ira est apparente a -ra qu'il sert parfois a remplacer (cf. vedique ajira a cote d'ajra et dans les lexiques bhidira et bhidra, chidira et chidra) (13), Mais le suffixe -ira figure aussi en sanskrit comme suffixe secondaire au sens possessif, par exemple dans medhira qui possede de l'intelligence (medha) >>. Ce double emploi du suffixe -ira a survecu en prakrit; Hemacandra s'en rendait deja compte, et il a note -ira comme deux suffixes differents : 72 II. 145 ladyarthasyeral), ira indique le sens de l'habitude, etc., et II. 159 alvillollalavantamantetteramana matob, alu, illa, ulla, ala, vanta, manta, itta, ira et mana remplacent le suffixe mat >>. Le suffixe secondaire -ira est devenu un peu plus frequent en prakrit; il ne garde pas toujours le sens possessif et il sert parfois a remplacer des syllabes finales insolites dans des mots rares ou empruntes: divira scribe du moyen persan dipir se trouve dans les inscriptions kharosthies (avec i long ou bref ?) et a passe en sanskrit; nacira du moyen persan naxir parait aussi dans les inscriptions de Niya et dans une inscription de l'Inde propre (18); gahira en maharastri tend a remplacer gahira >, namra courbe >>, prakrit namira. Mais ces mots ne sont attestes qu'a une epoque assez tardive et l'explication anaptyctique ne vaut pas pour la plupart des adjectifs en -ira du moyen indo-aryen. Le sens des mots prakrits en -ira est aussi varie que la forme et l'on ne peut pas en tirer grand-chose pour expliquer les voies de l'expansion de ce suffixe, Presque toutes les consonnes et meme des voyelles figurent a la fin des racines verbales auxquelles l'on a ajoute le suffixe -124 CHUMMIRA, GHOLIRA AGITE, BRANLANT 73 -ira. Il semble done que l'expansion de ce suffixe se soit produite par des voies multiples. Les debuts de l'extension se laissent deja entrevoir en sanskrit: rucira luisant semble avoir ete cree sous l'influence de rudhira rouge (16) (extension due a une ressemblance de sens et de forme); rucira: rucluire a son tour a donne lieu a mucira (dans les lexiques) < genereux >> de muc-liberer >>; l'extension ici est due a une ressemblance purement formelle. A l'imitation de timira sombre presque tous les verbes en -mont forme des adjectifs en -ira en maharastri : c'est la en partie du moins l'explication du mot prakrit ghummira agite, branlant >>. Mais l'extension du suffixe -ira a rapport aussi au sens : il y a peu de mots qui se pretent aussi bien a l'idee de l'habitude que les verbes qui designent vaciller, aller et venir, etc., et ce sont donc ceux-la qui ont pris le suffixe -ira de bonne heure en prakrit. Parisakkira qui marche se trouve deja dans le canon jaina, et dans les textes maharastris on peut rencontrer hindira et hallira qui va et vient (Samaraiccakaha); vevira qui tremble (surtout en apabhramia); pakampira tremble >>; bhamira, gamira qui marches, et ghummira qui branle (ef. l'influence probable de timira); vallira qui marche >>; rankholira, dolira et andolira <>; rasira qui crie >>; sahira << qui raconte >>, etc. qui * Ayampira et parisakkira representent les seuls exemples clairs de l'extension du suffixe primaire -ira dans le canon jaina. L'usage de ce suffixe semble plutot avoir ete un des traits caracteristiques de la maha. ristri litteraire : les mots en -ina abondent dans le Galdavaho et le Setubandha et surtout chez Hala. Vu l'importance de Hala pour la grammaire prakrite (17), il n'est pas surprenant que le suffixe-ira se retrouve deja dans la grammaire de Vararuci: IV, 24 trna ira sile, -ira s'emploie au lieu du suffixe de l'agent-tr pour designer une habitude >>. -ira est beaucoup moins frequent dans le Vajjalaggam, et si l'on exclut les mots formes a l'aide du suffixe secondaire, par exemple ullira et ujjagira, il ne reste que ring mots en -ira dans ce texte, dont un dans un vers qui se -125

Page #73
74 ANNEE 1962 retrouve chez Hala (Hala, v. 202). -ira n'est pas frequent non plus dans les textes maharastris independants de la tradition de Hala, il n'y en a qu'un ou deux exemples dans la Vasudevahindi, la Samardiccakaha, et la Lilavaikaha. -ira appartient donc a la maharastri litteraire de Hala et s'est repandu de la en apabhramsa et parfois meme dans les autres prakrits litteraires: Pischel cite quelques exemples tires de la sauraseni des drames tardifs (1), C'etait un suffixe tres utile, surtout quand les derives des noms de l'agent en -r s'etaient separes du verbe a tel point que l'on ne s'en servait quasiment plus. Le suffixe -ira s'est fermement etabli en apabhramea grace a l'influence de la maharastri litteraire, et s'est de plus en plus generalise. Il n'etait plus restreint a l'idee de l'habitude, il est plutot devenu une espece de participe present (ou passe) de l'actif, et c'est ceci qui explique la traduction sanskrite ghurnayamana (participe present) dont se servent les commentaires pour rendre gholira et ghummira. Cet elargissement du sens devait amener le declin du suffixe -ira. Ayant perdu son sens specialise -ira ne faisait que doubler le participe. Dans le Sanatkumara-caritam l'usage de ce suffixe est devenu si general, surtout avec des verbes a preverbe, qu'on est tente de l'interpreter comme un procede litteraire. Pour ne citer que les verbes precedes par pari: parikampira, parikilira, parigamira, paricintira, parituffira, paridhavira, parimillira, parivilasira figurent tous dans ce texte et sont presque inconnus ailleurs. Ce procede litteraire a l'air artificiel, et l'on voit bien que le sufixe -ira etait devenu trop vague pour rester expressif et vivant. Les mots ghummira et gholira laissent donc entrevoir deux tendances linguistiques: un groupe de mots qui ne s'est pas beaucoup etendu peut rester bien vivant et garder son unite (ef. hindi moderne: ghumna, dumna, lumna), mais un suffixe devenu trop commun est tombe en desuetude (-ira). NOTES T. BURROW, BSOAS, XII, p. 378. Cr. M. MAYRHOFER, Kurzgefasstes etymologisches Worterbuch des Altindischen, s. v. gharn, qui cite aussi les opinions contraires de Scheftelowitz, KZ, 53, p. 260 et de F. R. J. Kuiper. Vararuci, VIII, 6; Hemacandra, IV, 117; Ramasarman Tarkavagida, VIII, 2: Markandeya, VII, 9 et IX, 119. Le seul Kramadievara ne cite que ghopai. Chulai et ghammai figurent aussi dans la Definamamala. Chun- se trouve aussi cher Vararuci, VIII, 6, a cote de ghol -126 CHUMMIRA, CHOLIRA AGITE, BRANLANT. 75 Hammati marcher ne se trouve que chez les grammairiens et est attribue par Patanjali au pays de Suristra Le doublet dramati (e classe) est peu usite. Il est evident, d'apres le temoignage des textes, que summai ne vient pas de sap (malgre Pisci, Grammatik der Prakritsprachen, Strasbourg, 1900, 261). H. Jacont, Uber unregelmassige Passiva im Prikrit, in KZ, 28, 1887. F. JOHANSSON, Uber analoge Neubilding der Verbalflexion im sind, und balt slaviechen, in KZ, 32, p. 449 et saiv R. PiscuEL, op. cit. Cf. au contraire l'importance de l'analogie dans l'extension enorme des verbes en -l- et -kl- en moyen indo-aryen tardif (e) L'influence du passif regulier nisammai etre entendu peut avoir ete pour quelque chose dans le developpement du passif summai etre entendu >> ru entendre tendait a se confondre au passif avec snap dormir: ef. suni, aurai sapiti. khammai est actif chez les grammairiens Ramadarman Tarkavagisa, VIII, 16 et Markandeya, VII, 145. (1)jimai manger est d'origine kolarienne (ef. T. Bennow, TPS, 1946). Le doublet jimmal se trouve parfois dans les textes tardifs. Ce manque de certitude quant a la consonne n'a rien de surprenant dans un mot emprunte. C'est la forme a consonne simple qui est presque de regle et qui a survecu dans les langues modernes (hindi jernd, etc.). Le verbe sanskrit tim etre immobile, etre mouille est apparente a tam faiblir aussi bien qu'a stim->prakrit thimmai etre immobile, etre mouille, Ces deux verbes tim- et stim ne sont connus en sanskrit qu'au participe passe, limita et stimita, bien que les grammairiens citent parfois le present timyati, stimyati (4 classe, sous l'in fluence de tam-?). Il faut probablement expliquer les mots prakrits timmai et thimmai par l'influence du mot apparente tammai faiblir tam et des autres verbes de la 4 classe, semblables par le sens, kilammai, sammai. A. WEBER, ZDMG, 28, 352. T. BURROW, The Language of the Kharoshiki Documents from Chinese Turkestan, Cambridge, 1937, p. 98 et 99. J. WACKERNAGKL, Altindische Grammatik, 11, 2, p. 361. 1) Cf. L. Nerr-DoLc, Les Grammairiens prakrits, Paris, 1988, R. PISCHEL, op. cit., 596. - 127

Page #74
518 SCHWAR SCHILD: First, Second and Third in Middle Indo-Aryan FIRST, SECOND AND THIRD IN MIDDLE INDO-ARYAN LA SCHWARISCH V , ATLA M ACCORDING TO Y DE SAURER "motivated fint was obviously bad on a tem quite differ worde, which are linked with others aus derivativesent from The words for sint in the or infected forms, are much more frequent in modern languages, ... Hindi pala, Gujarati very grammatical languages soch Sanskrit pakeld, Marnthi pala semain disociated from than in modern lexicological languages, where cha oneThe derivation of these modern forms each word tends to be a separate unit that has no from the Apabiramis pubilla or pola preobvious etymological a ntion with anything tha is well known, but the US of the reele. This A S ure himself pointed out, is placement of prathama padhama by prathis by no means an sholute rule, and there are numer Warrant some further stady. ons exceptions to it. The ordinal number other Theater which us in prathama 'first than Art in some Indo-European W ed in Sanskrit for forming many ordinals form one map of exception is very gam-above out and for contrastive adjective with matilling. like Latin, unidos cond u perlative force ... Adima initial, first was quite separate from the cardion tumeral duo c a st. parcima last western. In epic two, wheness in more "lexicological language Sanskrit the cro w ns extended and it rederived from Latis much modern French, placed the superlative and in ordinals of maldeurme d' is obviously bad ondertiples of ten, this nanatime appen for ma n a There is lit doubt that similar te- ninetieth Bone of its frequent use with banes modellin nder the influence of the cardinal ending in the wut the considered to be ben took place in the words meanincondimai WA deveni ces where the vowel and third, but not first in the course of the Dot justised, incarime for carama development of Sanskrit into Middle and Modern last. At the same time in early Middle IndoIndo-Ary. Ary B show by Pali und Prakrit texts like, To Vedi and Sanskrit, as in practically all the relative superlative was weakened and graduIndo-European languages, the numeral pinam ally replaend by the comparative. This meat *. de Buure Cour le platigue erale, RL Two, pr e dtywaloploat Di od. Paris, 1949). 16 of the language Lendon, 1931) Wackel, Alrike Great Broa, dr World Bakit nama 180), p. 405, (New Haven, 1960). pp. 131er that the need was felt for a more emphate mafix pointed out, is foreign to the word 'first this where the superlative in a had been tased, per is generally emphntie and has a stronger alinity ticularly in the contrastive adjectives. Sachan with the contestire adjectives than with the emphatic and popular suffix was illa, which indi. numerale. An Iolated exception to this is aange eated possession as well as intensity, in the land in modern Marwa to mean the first day Ardha-Magadhl words fanals Oy, dila of the month, which is clearly derived from 'true' la appears the sun emphatie replace akaone. There were however in Middle Indoment for the suffis , both in contestire and Arya a mber of other derivatives of eles De other adjectives. This riddhima and riddhi which were not strictly speaking ordinal numer wealthy retsed in Jain MAr ri, anda , but which were to have an important effect hima and gentilla 'knotty and other similar on the development of other or numerals doublets. Sometimes al was simply added to in the early stages of Modern Indo-Aryan. the adjectives in we therefore find in The Sanskrit word dit indiri, second Andha-M udh punchinilleessternmost from and third, have arrived with modifications in pursuicima (purp m ) and paddimills Gujerati and in the thin languages, first from pratham. But in the majority of Marwil door wo o nd and jo "third." cases the suffix illa shows one of the main char- similar forms are found in the earlier phases of acteristics of being a ling" suffix: like the other Modern Indo-Aryan language, in Old it is substituted for the entire final of the word Bengali: doja and tits, tod, and in 16th century in which it appears. Hence one finds for instance A t dan anda (Tulad . But in the in the Ardh- M adhi of the Blagalera: majority of the present-day language these etywuarilla uppermot, haftilla "lowest, uttarilla mological form of the ordinal numerals have been Fnorthernmost, which are based on a restricted to specialised T e h Hindi httime and uttarima (Bhag. XVI 8). Identiel do and the second and third day of the lunar forms are found in the Jain Sarnet of the th. They have been replaced in their original Digambar texta e a and sarilla on mening of second and third by an entirely in the Bhagave dhand. Everything the new type of form, widely distributed throughout point to a very early substitution of illa for the the centre and east of India Hindi ddard and u x . One may therefore postulate an early ani, Bengali doard and tent. There is no direct Middle Indo-Aryan form a tulla for protams, evidence of the exit of such forms in Apa where cerebralisation of the did not tale place brumath em to have spread mainly in Prathilla e padilla in Apabhrami -illa from the modern innganges. The earlier Middle Indobeing an emphite wolfix became merely pleonastie Aryan tests Pall, the Jain on and classical (cf. He II 164), and by the Apabrams period Maristel all have derivatives of the classical it had become work that beloonlly a need Sanskrit ordinale to express the motion of second w telt to strengthen the word padla by the end third well some older formations addition of the comparative suitx - Page #75
SCHWARZRCHILD Pint' 'econd and Third in Middle Indo-Aryan 519 520 SCHWARZSCHEILD: First second' and 'Third in Middle Indo-Aryan skrit past participle apfa moved. This theory of Sanskrit renderings and possible etymological Wis not generally scepted, mainly on phonetic connections for the word kaaral. The Pand grounds. Much more convincing is the view of add- on mention to Sanskrit e nivo. 8. K. Chatter who derived the final -ars of lontstekeserake and ca arge. Eks-sarga having these numer from the verbal adjective and the mind intent on one object is very unlikely from Vatto move, which w e in Sanskrit A prototype for eklarata single "undi. to form the adjective pura preceding vided on mantle ground, and almoat impos This explanation has been accepted a probable wible on phonetie grounds. The obvious explansby Turner, Sakson and other scholars. A study tion is that alkaera a) was formed from the verb of Middle Indo-Arya indicates that the developart mowind wa parallel in composition to ment of the modern forms WI suggested by puro repurare preceding Basaraka) Professor Chatterji, although in some way more this formed would naturally have the meaning complet of 'going on one's own, aingle, and hence - In Middle Indo-Aryan test there actually divided in the Jain Bripture. The variant exists a formabant (with variante altre forma in Andhs-Magadh with lengthening of the or e san) which has the appearance of the type final of the areenily explained by the slogy of intermediate form that is required by the theory of elddade eleven and in solitary the of S. K. Chister. This word con number former word in particular brought about an ex of times in the Avetimbare canon, but always tension of the long to other compounds of ele in the same context. Typical of such an instance in Middle Indo-Aryan, e., del trenty-one! is a phrase in the last sentence of the Past in the Jain Mahat of the Pedro garan Panganie go pusaklando (20.79) und einen ninety-one in the Ardhndone whayunt banaal da Cr die Midhr of the Bandage. The word ekko wild the Pan garanti forms one rada), although highly specialed and rebook of Holy Writ. It hasten dided chapters. stricted in me in the Jain onnon survived to which are read in many days. Ella r eale later date and still exists in the Modern Hindi this used of chapters in the canon which are not word awar, ilar single, Nepali chane (cf. also divided into lessons. This is proved further by Middle Bengali kadra). statement in the Samanlage where details of Its probably the analogical influence of the contents of the other sacred texts are given b raka) 'single' that brought about the for Se namangalade docce e de althanshamation of new adjectives from the cardinal tesisa haand tell uddes , the numerals: do-sort and fint. These adjectives od Ang which forms part of the canon being parallel with a grabakara) single tains two book of Holy Writ, consisting of twenty. undivided probably meant originally two tothree chapters to be read at thirty-three different the making a complete two, three togetber, rending times. Abhayadera's commentary to this making a complete three. This sense of making e mention in explanation that the fint complete soran number is typical of ordinals chapter has four lessons, the second the the in Indo-Aryan, has been shown by Wacker third four, the fourth and sifth have two opengel who quotes se one of his examples the each, while the remaining en chapters of the famous penge from the Athardeda de first book are condivided" He continuent de mannede, n ad sedan at mamayant par la volta, "what two man sitting together speak Ishande, and in the second took there are seven boat, King Varan know that third (making great divided chaptersThis explanation complete the number thre). It was probably in counts for the thirty-three reading time necessary this manner that doar and fant developed into for the twenty-three chapters and corresponds to the actual division of the 80sidalgo L. DT . Sheth, Pets ADO (Caletta, it is 1993-6). handed down to n. The Adverbs , twee, three The commentator and lower give a variety times in Inter Avestan e Wacker III 427). For phantle they do not appear to repre.S. K. Chat , Origin and Development of the Bes t parallel development in Irania goN Langue alia, 1996), p.700. Wacker III. p. 400 ordinal namerlain Modern Indo-Aryan, replacing Sitting there among his companions and among the derivatives of dust and try Over most of cities of the verses of excellent poeta, the lord the centre and the east of India. The word for of men at once grew contented. A variant reading fourth and higher numeral are on the whole of v. 634 of the Litoral contains ek baarilessed in the completive type of enumeration sind. The ebding bere is no doubt due to the that has given time to the task of dosent and fearlesson of genitive plural indicating time, wuch ordisala. Cand found in the Inte Jainis Balanmor maydn. Derasionally the word Mainstrt of the Sudanthacaria) bw therefore curam has been confused with alkar ) by not become in ordinal, but has survivedu Hindi inter comentator and has been closed as ekada causar fourfold (particularly of a necklace), a once" (cf. Ludraba v. 178), but in the texts dice game, while the higher ers do not som it always signifies at ODON? simultaneously." to have followed the analogy of abbana) at the time of the endir indicates, chariot all. Although the formation of didant and and is like aku ). W adjectival in origin and we therefore probably due to the Prakrit word to some extent still felt to be an adjective in 'nglerundivided, it is necessary. Prakrit. The derivation of haaram is dearly in order to trace the history of these modern not from a n opportunity tor from ornal to take into account several other deriva cekamaralla) 'single' a suggested by the Pie tives of oba "one" in Middle Indo-Aryan. saddens . The correct explanation of A fairly common word in Prakrit is alkarim) the word has been indirectly hinted at by Modi Fone which occurs in the Jain can and in the glory to the currant . It seems tinues to be used in Apabhramis (g., Pag- certain that haram was derived from a Carin 2.14.1). Nebu ) is generally taken by one t ri ko b asarito which was the gramm s to be derivative of Sanskrit wided the adverbial ending in the case of chiar sing. Bpecial rules are formulated by a chleasant. B asarian thus mount Hemanden to explain it and its variant elle originally like one and when used with a word in Prakrit (He, 2.102, where hower it is taken for time simultangalya n ' The evidence Seguivalent of Sanskrit elada 'one'), and of the Middle Indo-Aryan texts apport this views in Apabhram (He 4.498 chako d). The ir- ra r ambling our both in Poland regalar change of the final - of kas to fis Prakrit ( s. 1.148) in maria) Ad appears almost certainly due to the analogical influence ogulsionally in Prakrit from the Ardha-Maradhi of other adverbs, particularly wom) Mulyono, of the Jain ponowards (of, Nyldhammado rather than to the survival of an Indo-European 1.1.2). It is found for instance in the Jain uffix is suggested by 8. Sen. Boch analogi- Mihrist of the Manipuliaria of Hariblandt cal changes in the ending of adverbs and con- 4751 Wana m akalda EMjunctions are not in Prakrit, as for instance Dare-dergryde, translated by R. Williams in us and joid) from yadd when in the charming story of Sapipati, which is A very similar adverb, derived from okolne like a divinity spreading the same of tranquility." figures in Prakrit. This is cleasarian at once. Sori e d by Hemania in the und Fsimultaneously which is mentioned by Hema- placarita I. 90, and in his , I. 143. candra (2.313: kirim hagikaprali). In Sarita) like although never a common in Middle some of its occurrence this word appears with the ending of the locative singular feminine as this arvation m ined by A. N. Uparge sary or c are preamably with the hixetition of the L and (Tombay, 1949). 136, olision of the boun w w w time o re 131. in the Mahir of the Lila Burjana (v.131): , K, Thakure, MD Did M. C odi, Gankwad's Oriental Series No. CXVIII toliko sahihim bas-chandanwattiahig si Paradt, 1956). 15. ha-auto naruto h inde WILL Turner, op. cit., 590. For details of the development of editorle. Berger, El Prod * Bukums Sen, Comparative rammer of meer Mindesben Laatler (Muleb, 1956) Inders Clutt, 1950). 162 42 Quelques Adverbes Peon du Moyen In W i se Port Versions of the Magipat dien . 4. CCXLV (1987). 24 carita (London, 1969), p. 337 p - 130 -

Page #76
SCHWARZSCHILD: "First," "Second' and 'Third' in Middle Indo-Aryan 521 an early form of dosard "second" with loss of the final ra, which was felt to be a suffixal element (cf. its use in doka-ri below). Tagare" lists the suffix-ra as being a rarely used pleonastic suffix in Apabhramia. ra was frequent in Sanskrit as a possessive adjectival suffix and there is no doubt that it was very much further extended as a suffix Ekkasarian simultaneously, at once has sur- in the Modern Indo-Aryan languages. It is therevived in Modern Hindi and its dialects as the fore possible that dosa was a back-formation from adverb aksar all at one time,' as distinct from dosard 'second' and belonged to such low strata the adjective ikear, aksar single. There are indi- of the language that it was thought to be a Dest cations that the adverb aksar was more wide-spread word. A number of Desi words can be explained in the earlier phases of Modern Indo-Aryan, and in this way; some have been recognised as Indothat like its Prakrit prototype ekhasariam it could Aryan long ago by Pischel," while others have occasionally be used adjectivally: in the Old Gu- been examined more recently, especially by P. jerati of the Gurjaranialeak there appears a Tedesco. Dosa 'half may therefore be considered nominative plural ekkasura 'all at one time.' The as an indication-however uncertain-of the exist change in the final syllable of the word eksar dosart 'second before the akkasarian shows that there may have been some time of Hemacandra, and it tends to show that confusion with the very similar word ekkasara(ka) the analogy of ekkasara(ka), which brought about > ekar() > modern aksar, iksar single, "un- the formation of the new ordinal numerals, bedivided." The adverb aksar 'all at once,' 'all to- longed to the Apabhramea rather than the Modern gether' has caused analogical developments among Indo-Aryan period, numerals in the dialects of Hindi: for instance the adverbial forms do-sar doubly, and fi-sar "triply in Bhojpuri are almost certainly based on ekser all together." Indo-Aryan as its synonym sarisa dohns "twofold. The Middle Indo-Aryan adverbs doha "G. V. Tagare, Historical Grammer of Apobrando (Poona, 1945), p. 341. Delindmandli of Hemocendre, ed. I. Pichel, d od, Bombay Sanskrit XVII (1938), p. 8. "In his article "Sanskrit uch-to gleam," JA08 71.8 (1957), for instance, P. Tedesco derives the Dea word punch to wipe from Indo-Aryan. -132 522 SCHWARZSCHILD: First, Second" and "Third in Middle Indo-Aryan doubly and tiha, Pali tidha 'triply from Sanskrit deidha and tridha, probably also played some part in the formation of the new adjectives. The long vowel in the Bengali words dohard, etc., may indicate that these words date back to the period when the adverbs dohd and tiha were still pronounced with a long final vowel, which influenced the new adjectives. The development of tha+ ra> fehra, Bengali fehand, was exactly parallel to that of dokra, Bengali dohari. The analogy of these forms has spread to the remainder of the first four numerals, whence Hindi ekakra 'single' and caward 'fourfold' and the corresponding Bengali forma, but the analogy was not extended to the higher numerals. The new adjectives ekahr, dohra, tehra and cashrd just like iksar, dusrl, tard and causar show that the first four numbers were felt to be a group in Indo-Aryan, a series par excellence. They are also a further indication of the tendency in late Middle Indo-Aryan to replace etymological forma of numerical adjectives by "motivated" words based on the cardinal numerals. -133

Page #77
SOME SPORADIC CHANGES OF VOWELS IN MIDDLE INDO-ARYAN LA SCHWARZSCHILD Melbourne Complaints have sometimes been made that the so-called "sporadic changes which include assimilation, dissimilation, metathesis and epen thesis have not been studied as thoroughly as the regular sound-changes. This applies as much to Middle Indo-Aryan as to other languages. The Prakrit grammarians have simply listed such sporadic changes as exceptions to the general rules, and the modern grammars have taken over their lists with some additional examples. This is true for instance of the cases of sporadic change of the vowel w which are listed by Pischel. This article represents an attempt at a further analysis of some of the problems involved. While Pischel simply lists the cases where the Sanskrit vowel has been replaced by other vowels in Middle Indo-Aryan, some more recent grammars make confusing statements to the effect that becomes a "by dissimilation or assimilation". Some of this confusion comes from a failure to distinguish between two different processes, dissimilation and dirferentiation. Differentiation occurs when two similar sounds are in immediate contact and the speakers distinguish between the two sounds, thereby changing the phonetic characteristics of one of them. Meiller and Grammonta consider differentiation as the reverse of assimilation: differentiation is a means by which assimilation is avoided and it helps to preserve the identity of the word which might be lost if assimilation took place. On the other hand dissimilation occurs when two identical sounds are close together, yet not in contact, and there is a tendency to avoid the repetition of the one sound. This might be considered as a more truly L. A. SCHWARZSCHILD sporadic change than differentiation, it is often merely due to a slip of the tongue and is linked with the subconscious dislike of repetition. On the basis of this definition it is clear that cases of dissimilation of vowels are quite common in some languages as in French: Fr. selon soblongu Page #78
SPORADIC CHANGES OF VOWELS IN MIDDLE INDO-ARYAN 27 example of vocalie dissimilation and certainly points to the influence of the longer derivatives of the word which had the vowel a in the initial syllable, that is the forms gariyas and garisha, already quoted, as well as other derivatives, especially gariman 'weight' > Pkt garima, and garina 'weight': the longer form garuka was associated with this group rather than with the simple word guru which survived in Middle Indo-Aryan as guru. The Prakrit word purisa, Pali porisa, from puruso'manis at least partially due to the substitution of the more usual suffix -isa (e.g. as found in words like erisa, harisa): -usa was also replaced by-asa in pharasa parusa 'rough'. The unusual development in the first syllable of the Pali form porisa has been explained by H. Berger on the basis of the generalisation of the metrically long Vedic form pirusa." One of the possible cases of dissimilation of the voweli has been interpreted as an archaic form: sadhila "loose, which occurs in Prakrit along with the more usual sidhila sithila, has been explained from spthila "loose'. V Srath 'to become loose' 10 Some of the other words which could be considered as examples of vocalic dissimilation are rare words, or unusual or borrowed, in any case they are words that have no strong links with others; they do not belong to a big etymological group and therefore have little support in the language and are more liable to "sporadic" changes. Such a word is the Dravidian borrowing aguru, Prakrit agaru, agalu, agurk, which was obviously influenced by Skt guru and Pkt garua with which it was associated by popular etymology.". Sometimes the origin of the word is so unclear that it is difficult to say which is the original vowel. This is the case for instance with the probable Munda borrowing which is represented by Vedic chubuka chin', Sanskrit cubuka, cibuka, Prakrit cibua. Equally uncertain are Sanskrit fumburu, Prakrit tumuru, tumharu and imbaru. Pali timbaru 'a tree': Sanskrit kilinja, Prakrit and Pali kilanja: Sanskrit mucilinda, Prakrit mucilanda 'a tree as well as Sanskrit Phillika, Prakrit jhallika 'an insect'. These words cannot be considered as clear examples of dissimilation of the vowels i and w, but it is probable that dissimilation did play a part in the unusual sound-changes undergone by these rare words. Sometimes the change of vowels is due to analogical influences rather than to dissimilation: thus the Prakrit neura, alura Sanskrit napura * H. Berger, "Pali pria "Mensch, Wiwr Zeitschrift fur die K e Simi Ostasien, I. 1957. 14 R. I. Turner Dictionary of the Near Lawape London, 1931), s.v. ila 11 For a Dravidian loan-word Nee T. Burrow, The Sanskrit Language (London, 1955).. 30, and for some of the further history of the word see Jean Filliozat "L'Agalloche et les Manuscrits sur Bois dans l'Inde", J.A, 1958, pp. 85 IT, L. ASCHWARZSCHILD anklet' has been shown already by Pischel to be a remodelling under the influence of keura Prakrit bhumuha, bhamuha, but the suffix-maya was substituted for the original final of the word. This is indicated already by Hemacandra (1 167). The change of blumaya to the alternative form bhamaya is probably due to the influence of the original word bhrumukha > bharuha 'brow. The development of brumukha to hhamtha represents what is perhaps the only clear example of vocalic dissimilation in Middle Indo-Aryan. It is typical of examples that abound in other languages, the change being regressive rather than progressive: the earlier of two sounds is altered rather than the later, because of the tendency of the speakers to think ahead. The word bhamuha had be. come completely dissociated from the original bhrabrow and was therefore liable to further sporadic changes in Apabhramsa, where it appears as bhoha. The scarcity of examples shows that dissimilation in Middle Indo-Aryan was very rare indeed, probably in the spoken as well as the written language. The infrequency of vocalic dissimilation in Middle Indo-Aryan, as opposed to the Romance languages for instance, may be explained by the more even accentuation of Middle Indo-Aryan: whereas particularly in French the stress accent was so heavy that vowels in unaccented syllables, preceding the stressed vowels, tended to be dissimilated. It would appear that differentiation on the other hand was not so infrequent in Middle Indo-Aryan, though it has not usually been recognised as such. The most common kind of differentiation is of the type mukula > m a > mala 'bud', where the first of the two vowels w is altered as much as it possibly can be and loses its major phonetic characteristics: it is unrounded as well as lowered, becoming a which for some time at least was pronounced in a separate syllable from the following u. There are a number of very similar cases: muka diadem' appears as maida in Ardhamagadht as well as in later Jain texts, mukunda 'a kind of drum' becomes maunda, mukura becomes maira. Some of these words appear in lexical Sanskrit with the vowel a in the first syllable, and sometimes also in Buddhist Hybrid Sanskrit, e.g. makufa 'crest', 'diadem (Divyavadana), makura 'mirror', 'bud' (lexical Sanskrit), and it is probable that the vowel a in the initial syllable of these forms is due to the # The tendency for dissimilation to be regressive rather than progressive is discussed by R. G. Kent in Language XII, pp. 245-258. - 136 - -137

Page #79
SPORADIC CHANGES OF VOWELS IN MIDDLE INDO-ARYAN 29 influence of Middle Indo-Aryan. It is possible that kutuhala < kohala "pleasure", "curiosity shows the same type of differentiation, as suggested by Pischel, and that kutuhala was changed to *kauhala, whence kohala. The occurrence of the variant form kouhalla in Hemacandra's grammar and of kouhala in the Jain canon might suggest that in some areas perhaps the differentiation did not go so far in this word, and the u of the initial syllable was only lowered to o, and not lowered further and unrounded to a. It is however more likely that the word kutuhala > kohala does not represent a straight-forward example of differentiation, but shows the influence of the derivative abstract noun kautuhala > kouhala 'curiosity": this is almost certainly the case in the Sauraseni and Jain Sauraseni form koduhala. The usual Prakrit and Apabhramsa word somala "tender' has been explained by Pischel as being parallel in development to kohala: sukumara > suumara, suumala saumara > somala, except that in this case the intermediate form samara is actually attested in Maharastri (Setubandha). Professor Alsdorf has suggested a different etymology for the word somala, namely saumya+ala. Even if one is persuaded by the presence of the intermediate form satimdra that the old derivation from sukumara 'tender' is probably correct, it is very likely that the word sukumara > somala was at least associated by the speakers with saumya > soma 'good', 'gentle and this may account for the prevalence of the form somala rather than saumara. Although the usual process of differentiation of u-u is thus to a there are some other cases where it seems that the second of the two vowels has been differentiated to a. The noun dukula 'a kind of cloth' appears in Middle Indo-Aryan as duula (Hala), dugulla (Ardhamagadhi and Jain. Maharastri) and dualla (Hemacandra). This last form is almost certainly due to the substitution of the suffix -alla for the final ula, -ulla; the influence of the two other words of similar meaning ambara > ambala (Pali) 'cloth', 'garment and kambala 'blanket', 'woollen garment was probably not instrumental in this substitution, as these two words do not show a double / in their final. Jugupsa> dugumcha, du(g)amcha "disdain' is also an example of the complete remodelling of a word rather than simple differentiation; the initial syllable was obviously influenced by the pejorative prefix dur->du-. There is thus little doubt that the usual differentiation of u-u was to all. There is a further and barely recognisable example of such a differentiation in sarisa 'a good man' which occurs in Jain Maharastri and Maharastri. This is probably not always to be de Harivansapurana von Puspadanta, ed. Ludwig Alsdorf (Hamburg, 1936), Glossary s.v. somala. -138 30 L. A. SCHWARZSCHILD rived from satpurusa, but sometimes corresponds to supurusa, which appears also as suurisa and surisa (Hc. 1.8). The two vowels all occurring in hiatus formed a most distinctive group: they were as far apart in pronunciation as possible, and because of this distinctiveness the grouping seems to have been favoured in Middle Indo-Aryan. The same applies to the vowels at in hiatus, though perhaps to a slightly lesser degree owing to the probable frontal articulation of the short a in Middle Indo-Aryan. These two groups, af and all thus became very prominent in Prakrit and Apabhramsa. They were pronounced in many tatsama words from Sanskrit which had originally contained the diphthongs ai and aw that had disappeared from the language by Middle Indo-Aryan. all was quite particularly frequent, and only in some cases the variant semi-tatsama form with o was allowed by the grammarians; af was slightly less common and the variants containing e were used more freely. The groups all and at were also introduced in some cases by differentiation, not only in the examples of u-u> af listed above, but also in a few words where there has been further differentiation from eu and ei, by the lowering of e to a in contrast with the following higher vowels. This might account for the unusual form auna, from euna, eguna < ekona lacking one' which is found in numerical compounds in the Jain canon, as for instance in aunattisa 'twenty-nine' (Nayadhammakahao). The meaning of the word aina renders the derivation from aguna, suggested by Pischel, quite impossible. By a change very similar to that of ekona > eina to auna, Prakrit chailla (Gujarati chel) has replaced chei(D)la from Sanskrit chekila "clever'. A most striking instance of the way in which the distinctive group of vowels all was favoured in Middle Indo-Aryan is shown by the cases in which the u was due to the vocalisation of labial consonants. After vowels other than a such vocalisations do not seem to have occurred until very late in Apabhramsa and in the modern Indo-Aryan languages; thus sapatni 'rival wife' became savatti and sautti in Jain Maharastri (Supasanahacaria), pravahana 'ship' became pauhana, navati 'ninety' became naui. When following on vowels other than a, the labial consonants do not undergo such a development until considerably later, if at all, and we for instance always find devara 'husband's younger brother', and sevaya For the pronunciation of short a in Sanskrit see W. S. Allen, Sandhi (The Hague, 1962), p. 30, and P. Thieme, Panini and the Veda (Allahabad, 1935), pp. 89 ff. and 118. 15 R. Pischel, op. cit., p. 58. Cf. R. L. Turner, A Comparative Dictionary of the Indo-Aryan Languages, Fasc. II (London, 1963), s.v. ekona. The Old Hindi forms agunis etc. might be explicable by a contamination between alina and eguna. -139

Page #80
SCHWACHILD: The Middle Indoren Preto THE MIDDLE INDO-ARYAN PREFIX - OFF AND SOME PHONOLOGICAL PROBLEMS ASSOCIATED WITH IT LA SCHWAR CHILD SPORADIC CHANGES OF VOWELS IN MIDDLE INDO-ARYAN 31 Servant' (apart from an isolated example of suya seveka in the Kalpasitra)" and waya wepaka'shivering' in Middle Indo-Aryan, and forms such as deurdin devral husband's younger brother's wife. and diso (Nepali)Page #81
SCHWARSCHILD: The Middle Indo-Aryan Profia 10 351 352 SCHWARISCHILD: The Middle Indo-Arya Prefu 10 ol pesn the use of an initial is restricted to 'The most frequent use of initial in Prakritis some forms of the verbs a lo transport and however is verbal prefix. In a number of care to speak and even there it is of secondary the origins of this prefix are perfectly clear, and origin, arising from the change in the cause there are obvious antecedents in the earlier lanof waland from reduplication in the ens of no g e. The corresponding verb in Banskrit are "Lo speak. There are some proper names, but the used with the prefixes per thus are generally offer to root date to be of any chandai < a + child to cut off, and the signifieanon regarding the phonology of Sanskrit, past participle occhina chinna are very e. Voldaaka, plscene in the Rajotarari frequent in Andhamgadhi, Jain Maharasti and s, and Vapaa penone of the twelfth Mabini , in the Rain ); toka < or thirteenth century, A.b. Only very few common yapabarsals to remove pet rid of' is also nouns begin with in Sanskrit. These are either common in Middle Indo Aryinwide on not attested at all in ondinnry Sanskrit texta, decided' occurs in the Gandhar Dharmapada appearing only in the work of the lexicographer (v. 5); Bodas d e destrition of se or they are found so Inte that they cannot reflect tions and MEG < imite finished off any early stage of the Ishuset ex., were extinguished' occur in Andhamini und Join kind of large snake moro, s kind of rice, Mahir It is probably on the analogy of these vaja and sold are numes of plante, torking and similar forms that para destruction and soldha refer to home of le red and of a living creature always appears to ebemut colour respectively, rodala a kind of in Jain texte, fish, and ala 'copyint or writer. All these The case of the verb oral, or to leave'is words are of late or unknow origin, and are rather more complex The Prikrit numarans bown by their meaning alone to be fringe-ele- and the older glossaries generally derive these ments in the langung, belonging mainly to those forme from wor lodimi. Peschel showed spheres (tion and funn) where borrowing from a long time ago that the Prakrit word probably non-Indo-Aryan langsinges would be mo r e the blending of meer veral ots: he pected. Others such wrontalk' (erler - suggested all to leave, and to run (e. skritta) represent reborrowing from Prakrit. Prakritsarai). The presence of the forms you , As mentioned above, the underlying phonetieso 'to be handed over of payment in C e for the one of initial and had the Niyu texte har rendered the probleme much Vanished by the Classical Banskrit period, but more complex, and T. Burrow we led to the nevertheless words beginning with the syllable onclusion that yet soother verb was involved are marked out immedintely a foreign words in and that the verb from which hoth and Sanskrit the Prakrit verb putra were derived was s ind In Prakrit on the other hand initial did . "The past participle torta regularly beDeer and initial wou by no me uncommon, come w eils in Niya Prakrit. This derivation In some cases represents a regular boetic would provide of Sounting for the development of earlier forms . Bankrit presence of the vowel in the stem of the verb oma > Prairit romsky' in other it moet togiral in Middle Indo-Aryan, and it is supported certainly representa borrowing from non-Indo by the words are a ll olinadi quated by Aryan languag, .. Prakrit sodrada (or bo Barrow from the Buddhist Hybrid Sanskrit of draad dosha both menting young man har da, ed. J. Fough London, See Mar, JOS LXT PO 45 W 12 Allen, Ban The 1), p. 72. 3. Wockel R. Piebel, Grammar Prop heta Asche Gram p 29 quotes the contineture, 1800), p. 10 view of Benfey, Bloomfield and others "T. Harrow, The Language of the Khas F. 3). Kuiper. Proto Mund Wande in d it frow C r eaton (Cambridge, 1887). pp. (Amster , 1948), p. 30 the Maharastu. Burrow interpret the - in the prefix te-and the populars-has natural disapstema & svarabhakti-vowel between and r: peared in the course of the development of the yanarayati > EORG. Later irai and sarai modern languages with the continued tendency berme dully confused in meaning and form for to become a bence Hindi boni - In some part of the verboard in the area coduadrala trider origin from spante Sanskrit verbs is still clearly Sometimes it is not possible to determine with visable: the Prikrit matha 'le' is almost certainty whether Middle Indo-Aryan word certainly derived from sanarsa, and the ab- has been inherited from the older Inngange, or solutive post from an y , while there is whether it has been newly created: thus welanta In alternative absolutive art from your one away might be derivative of Sanskrit But the evidence of the forme pisarcito go way yopalnica, but it could also have been formed (from ) and i n to fall from vi + within Middle Indo-Aryan with the help of the ri) shows that in Middle Indo-Aryan the verbs pretixo. But there is a series of words in which to run and to rest on, to direct toward h eberly been substituted in Middle Indo were hopelessly confused, and this strengthens Aryan for the older and less expressive prefix the case for the complex origin of nai, viral thus wil disappeared corresponds to from the verbs - and Sri- with some influence o poet, gone, i to open for blog soma corresponds to do to open out of From the use of verbs such as the ones listed, blossoms): iga abandoning occur in the which began with the profix o derived from some textes s 'abandoning you, yapay it became apparent that there The prefix is often used with verba that have Was in Middle Indo-Aryan a prefix to which to antecedent in Vedic or Sanskrit. Thusokho implied off why and was rather more em- riya 'bejewelled, decorated is probably connected phatie than This new prefix was then added with the Del word Khareya 'bad, blossom. to other verbs in Prakrit and Apabrand and The empatie meaning of is not much in became almost the rule with certain verbs with evidence here, and it more or less corresponds to which its meaning blended particularly well the meaning of al in Mita adorned. In This new prefix 10. Wis felt to be distinct from compounds with such Dest words and sometimes the original prefix and from an op- elsewhere - m y alternate with another prefix which have survived in a contracted form, of Middle Indo-Aryan origin, which is ul peruspe B e mi leamed formustion. In this timately based on udt - For instance theontracted for the meaning of all to pour out can hardly be separated away is not sunlly very clear. Thes e from tillatte spilt, emptied', i 'broken, aya decision is always found in that formspileseems to belong to the same root assoravia in Jain M inistre, and not a suhana l alled, which could perhaps be regizded saakara'oonduet, setion is an important tech- Dei word, since the standard derivation from nical term in Jain religious literature. When a taparopita taken away, deprived of is not alloword sodre does occur in Apabhrami, il l e ther satisfactory. crisiyd's collection of the much more concrete and foreful meaning of flower' which occur in the LaLlah, has its sweeping away (rubbish from a courtyard) Dearest equivalent in a collected and (Pa r is 78.3.5.), and it is probably derived weigu to collect from Sanskrit + V ). from the Middle Indo-Aryan profix combined in very few cases the original emphatie meaning with the verb Maret var to take away, mother of v olwy appears to have been boet that directly from the Sanskrit yapabayati. The Aabhrams word sodral almost certainly b roma slabe baad on the Middle Indo-Aryan e goH. Chayani, Baralia Vidys represents the origin of Modern Hindi bwborn to XIX. 115 sweep. The distinction between the semi-learned e Bharatiya Vidyo XVIII, p. 25-23 s 143 -

Page #82
SCHWAR CHILD: The Middle Indo-Aryo Profis 353 361 BCHWAR CHILD: The Middle Indo-Aryan Prufa - C m is so more after initial stabilisation does seem showed that the change of tow was later to beur in close proximity with other labial Pand, but not s r . While this evidence may a nts in Middle Indo-Aryan anda be regarded conclusive only for the early Occasionally changed to in p om a nd formative year of the Middle Indo-Aryan prathama "first. languages, it would make us inclined to attribute As shown above, the phoneties which rather late Apsbhramla date to the further lavoured the use of contrusting vowel after change in pronunciation by which initial may initial had disappeared early. It is therefore have developed into unrounded bilabial not surprising that Middle Indo-Aryan should be frestive with complete closure in articulation tolerant of initial and But spart from the listely bemeentes with over large very noticeable extension of the prefix parts of Easter Indis way. disc above, the sound combinations The Middle Indo-Aryan development show and were not very common and there was that there has no been a radical change in the O general tendency towards Founding of a Tihonemie system. The Old Indo-Aryan system after From these indicatione it would and that contained two semivowels which were closely initial pronounced without lirondine parallel and were mainly used with contrasting in Middle Indo-Aryan and that it was probably vowels in the initial position. y probably end still a labio-dental frientive. This view con to have any independent phonemie vale in most firmed by the conclusions of H. Berger, whe Middle Indo-Aryan languages; but the history of the prefix help to show that initial Pelapot. 1. Berger, Probleme der Minden mained distincta labio dental for a consider Loutre, Munich, 1905), pp. 51 sble period. completely, in rokkhandaya arriving coming totum, the future wocchi, ete. The verb olla close' which is connected with khanda attack to walk which may have been based on a and ultimately with and leap. In the major followed this analog. Very similar in develop ity of compounds the original meaning of the more word waar o m what is very clear. Typical of such wonde is the adje s to be woven, which occurs in the Gandhar tive are put gone, which is listed in the Dharmapade (v. 149). wow has been explained Delinmemeld, but the derivation of this word by J. Brough is derived from the rare verb and its connection with a frightened top to weave the vocalism being influenced are quite unknown. by Prekrit du, which represents either the past The Jain Maharisrt word w atthalso participle omwoven', or the noun of welt. e past and derived from the modern Marathi word of Gujarati el yot e pt. This etymology is supported the form id appearing in the Delima by the existence in Prakrit of words and the older Marthi 's may all be yote change. The prefisso, y connected with the verb to weave and the introdud into the word on it seemed same past participle omwoven with me to fit in partlarly well with the meaning and hesitation over the presence or absence of the there were stamber of synonyms beginning initial consonant. The Prakrit word je load with two roli, s hare one. One might (Hindi bujh) is very similar information: it is feel tempted to believe that the use of the form derived from the obligatory participle who och instead of recenthe, which would have that which is to be carried, (w), which was been thermal derivative of l i fe.ced by the article and the not due to the influence of the prisot to finitive On this verbal no ho possible pounding of the vowel a to, brought on there w as a new Middle Indo-Aryan sbout by the initial labial consonant. But sich verb hal toesty. tendency to rounding seems to have been almost some of the other words in which there might unknown in Middle Indo Aryan. In most cases appear to be evidence of labilisation of a to where one might suspect rounding there is by preceding are either of restricted occur more probable alterative explanation. Bakrit ne or of very uncertain origin. The und habeat l y represented in Middle wine, which been just one in the Indo-Aryan by the best. The change of A r t of the Paco of Sambha to in this wondhas now been ined by H C de pe to be connected with the Sanskrit Bhayanis due to the influence of adha and wond e bon. A few words of unknown origin other parts of the verb to trupor, convey might show this type of rounding, but none of the with which it was still soorinted. The history viable examples ne convin t hu Sanerit of the Inte Middle Indo-Aryan verbs olla 'to ba n d is almost certainly related to Prikrit walk d ollar to speak' is very complex, but raha, cha, but the change in the vowel the presence of the vowel in the verbs is might well be connected with the posible units unlikely to be simply due to i t . If origin of this word. A few Prakrit verbs are most ome were to seept the view that made to speak' difficult to explain in this respects they are is ultimately based on the would alwa tofan Sanskrit ), and seem to be due to the analogy of those parts of Sankt-) to fear, well as the post the verb that contained Bankritt, participe n , und anskrit wignal and whenee slothe Middle Todo-Aryan infinitive ein rightened. All these forme do seem to Pr o f show some kind of labilisation after the initial Madee, ed. He Bhayan, but other puer Vol. IIT Bombay. 10. 224 to excluded, and it Not e rys Wade is difficult for instance to discite this group JAO 11, 2016 of words from here frightened. Although it

Page #83
A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN By L. A. SCHWARZSCHILD THE IMPERATIVE OCCUPIES A unique position in the conjugation system. In its real function it implies the presence of a speaker issuing a command to one or more listeners, or making an order referring to a third person or persons. This means that the conjugation tends to be defective. Because of its immediate association with the speaker the imperative is generally linked with the present tense, but through its meaning it is also associated with the future tense. The uncertainty of whether a command will be carried out links the imperative with the subjunctive, while the idea of volition brings it close to the optative. These special features and varying associations make the history of the imperative extremely complex in Middle Indo-Aryan as in other languages. An attempt is made here to study two aspects of this history: the use of the subject pronoun with the imperative, and the distribution of the forms of the second person singular of the imperative. Other features of the imperative, in particular its relation to the passive, will be discussed later. One of the interesting features of the use of the imperative in Middle Indo-Aryan is the frequent occurrence of the pronominal subject. As has often been pointed out in works on syntax,' this usage was not unheard of in the classical languages including Sanskrit, but in all cases it seems to have been confined to popular speech and does not generally figure in the more stylized texts in these languages: e.g. Latin narra tu mihi (Cic.) "tell thou me", Sanskrit ma tvam vairagyam gaccha (Kathasaritsagara) "do not thou grow weary of life". This type of construction is very frequent indeed in Prakrit, particularly in Ardhamagadhi and in Jain Maharastri, as can be illustrated by numerous examples from the Svetambara Jain canon, both from the older and the more recent sections: Se nam paro navagao navagayam vadejja: ausamto samana, eyam ta tumam chattagam va cammacheyanagam va ginhahi, eyani tumam viruvaruvani satthajayani dharehi, eyam va tumam daragam va pajjehi, tusinio uvehejja. "Another passenger may say to him as he is sitting in the boat: "Venerable monk, take thou this parasol or this rug, hold thou these various weapons, give thou this child a drink,' but he should look on silently." (Ayarangasutta, 11.3.2.) Especially frequent in the canon is the phrase: gacchaha nam tubbhe devanuppiya, "go ye, beloved of the gods," which generally introduces instructions to the household servants (e.g. Nayadhammakahao 1.16). Similar examples are found in Jain Sauraseni, though perhaps not quite as abundantly, e.g. jinasu tumam, "conquer thou" (Bhagavati Aradhana, v. 1447). In the other literary Prakrit dialects this usage is also found, though less generally, as in Sauraseni: tam tuman labhasu, "take thou this" (Karpuramanjari, Act I), and there are also instances in Buddhist Hybrid Sanskrit, e.g. na tuvam manasikara, "do not concern yourself with it."* H. Hirt, Indogermanische Grammatik, Part VI: Syntax (Heidelberg, 1934), p. 154. Quoted as an example of the use of the imperative with the prohibitive particle ma by J. S. Speyer, Vedische und Sanskrit-Syntax (Strassburg, 1896), p. 58. * Quoted by F. Edgerton, Buddhist Hybrid Sanskrit Grammar (New Haven, 1953), p. 203. This example is based on an emendation and a more definite case is given on p. 108: sa va manjasiri precha sadhana, "do thou, O Sadhana, ask Manjusri." -146 93 A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN The pronoun may precede the verb, as in the long example from the Aydrangasutta quoted above, or it may follow the verb as in some of the other examples and in fairly usual types of Ardhamagadhi phrases like: paccapinahi nam tuman Dovaim devim, "bring thou back Queen Draupadi" (Ndyadhammakahao XVI). There is nevertheless some uniformity in the position of the pronoun subject of the imperative in early Prakrit. It usually follows the first word of a clause, or is separated from the first word only by a particle, especially the emphatic particle pam. This rule may be seen to apply in all the examples quoted above. The regularity with which the pronoun subject of the imperative occupies this position is closely related to the unstylized usage of Sanskrit and of the classical languages. This second position in the sentence is generally unaccented, and it would seem in fact that the pronoun subject of the imperative was used to give personal emphasis to the command, but was not itself heavily accented. Particularly in the singular the imperative of the second person was distinctive in form, and in the plural it was identical only with the second person plural of the indicative in Prakrit, so there was no formal need for the pronoun to indicate the person of the subject: it was purely a matter of emphasis characteristic of popular style, and this emphasis became very frequent in Ardhamagadhi and Jain Maharastri. In the meantime the popular language developed further and some of the changes are reflected in later and unstylized Prakrit texts. Here the use of the pronoun subject of the imperative is much rarer, and is on the whole confined to cases where there is a contrast between persons. This is particularly marked in popular Maharastri prose, e.g. in the Kuvalayamala1: deva, aham ceva vaccami, eittha tumam, "sir, I alone am going, stay thou" (157.5), or again in contrast to the speaker, senavalno, vaccaha, niyattaha tubbhe, "generals, go, return ye" (146.7). The construction with the pronoun subject of the imperative is practically missing from a popular text like the Vasudevahindi, but it does occur in a later semi-stylized Maharastri work, the Lilavaikahd: ma tamma tumam, ma jharasu, ma vimunica attanam, "do not falter, do not faint, do not abandon thyself" (v. 573). It is noticeable that here the pronoun subject tumam no longer occupies the second position in the clause. It is in fact characteristic of these later popular texts that the pronoun subject, when used, appears almost invariably at the beginning or the end of the sentence: e.g. in the Cauppannamahapurisacariyam, where it is used to mark contrast, citthasu tumam, java aham chiviinam uyagam agacchami, "you stay, while I wash and come back again" (p. 160), gavesatssami ham ti, tuman puna... citthasu, "I will go and search, you stay" (p. 161), and also tumam tam eva pucchahi, "you indeed ask her" (p. 157), where the pronoun subject is clearly in the initial position. The construction in which the pronoun subject of the imperative occupied the second position had earlier been a mark of colloquial speech, but became so general in Jaina Prakrit that it was ultimately felt to be both stylized and meaningless and ceased to be a popular feature. The use of the pronoun subject later gradually came back into the language in cases where a contrast between persons was to be emphasized, but it tended to remain outside the sentence, either at the beginning or at the end. Quotations from Prakrit and Apabhramda texts are taken from the standard critical editions, notably those in the Singh Jain Series. The Kuvalayamala, edited by A. N. Upadhye (Bombay, 1959), forms vol. no. 45 of this series. Cauppannamahapurisacarlyam, ed. A. M. Bhojak, Prakrit Text Society Series no. 3 (Benares, 1961). - 147

Page #84
A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN 95 94 STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN Apabhrama continues the tendencies shown by the later popular Prakrit texts, and the use of the pronoun subject of the imperative is rare. When the pronoun occurs it occupies an initial or final position in the sentence. At least in some cases it may have had a vocative function. Examples are: fuar appuu jaevi levi eki, "you yourself go and bring him in" (Karakandacarlu of Kanakamara 4.3); kari jayakaru tuhu,"shout hail" (Paimacariu of Svayambhudeva 12.10.9); ruh hohi deva khayarahirdu, "o lord, be thou king of the Vidyadharas" (Harapapurana of Puspadanta 91.11.6), and bharu bharu Saccahame, saccan tuha, "oh Satyabhama, speak thou the truth" (ibid. 88.19.8). In the last two sentences the use of the pronoun subject is associated with the vocative. There are only very few instances in Apabhramsa where the pronoun subject of the imperative does occur in the middle of the sentence, but then it is always linked with the vocative, e.g. jam janahitam pahu tuli mi kare, o lord, do thou as thou knowest" (Palmacariu 6.16.8). and malu taniya pihimi tuu bhuile bhdya, "o brother, enjoy thou (sovereignty over) my country" (ibid. 4.12.6). In Apabhrama even some of the singular endings used for the imperative were no longer distinguishable from the indicative, e.g. -ahl was used for both moods, and the distinction between the two moods must have been mainly prosodic. It is also possible that occasionally in Apabhrama texts we may glimpse an incipient tendency to differentiate between the two moods by the use of the pronoun subject, which was more usual in the indicative than in the imperative: for instance in one passage of the Palimacariu of Svayambhodeva (19.15) there are two indicatives with the pronoun subject of the second person and two imperatives without subject, hd purta, putta, kahim gayali ruhu, luhalu mul, nikkdrane povahi kain rulum? "o son, show your face, why have you gone? O son, dry your face, why do you weep without cause?" This tendency, which is barely noticeable, does not appear to have been further developed, and formal distinctions between the imperative and the indicative moods seem to have been gradually re-established, at least in the second person, as for instance in Hindi. The general Apabhramsa conditions survive in modern Hindi, where the pronoun subject, when used with the imperative, tends to stay on the outside of the sentence, generally at the beginning, e.g. tu cori na kar, "thou shalt not steal," um to Yadavon ko maro, "do you then destroy the Yadavas," in both familiar and polite forms: (a) ja, "go (thou)," (tunt) do, "go (you)," and (dp) jaie, please go." The pronoun subject of the imperative, which had occupied an unaccented position in the early forms of Middle Indo-Aryan, in conformity with the popular usage of other Indo-European languages, later moved into a more prominent position at the beginning and sometimes at the end of the sentence. In late popular Prakrit and in Apabhrama the pronoun subject was no longer just added for a little extra emphasis after the beginning of the sentence, but it fulfilled a special function, particularly when there was a need for contrast between persons, and it was also frequently joined with the vocative to bring home to the person addressed that the command was meant for him, and the command was The vocative is often to be regarded as the psychological subject of the clause. See A. Sechehaye, Essai sur la strhne logique de la phrase (Paris, 1950), pp. 27-28. IS. H. Kellog, Grammar of the Hindi Language, 3rd ed. (London, 1938), p. 460; and A. Sharma, Basie Grammar of Modern Hindi (Government of India, Ministry of Education and Scientific Research, 1958), p. 61 thereby made much more personal. Any incipient tendency for the use of the pronoun to fulfil a morphological function, and to distinguish between moods, did not develop further in Apabhramsa and in modern Hindi. Apabhramsa resembles Medieval French by the manner in which the pronoun subject of the imperative was used, but in modern French developments have been quite different and the presence or absence of the pronoun subject serves as a means to differentiate between the indicative and the imperative mood. Modern Hindi on the other hand has retained the flexibility of late Middle Indo-Aryan, where the use of the pronoun subject can provide a special nuance of meaning in rendering a command more personal. Few problems in Middle Indo-Aryan have caused quite as much discussion as the derivation of the second person singular imperative ending -s. According to the first of the two major theories, Prakrit -54 and Pali - come from the atmanepada ending - This theory was originated long ago by Lassen, and upheld by Weber and Jacobi, and more recently and convincingly by Edgerton, who argues that the Prakrit and Pali forms cannot be separated from one another and that the use of an atmanepada ending in the parasmaipada was due to the general decay of the armianepada in Middle Indo-Aryan An opposing theory was held by Pischel," and followed by more recent scholars.' Pischel believed that the ending -S4 was due to analogy, which brought the imperative endings -51, ()u exactly in line with the indicative endings -si, -OL. A very tempting compromise between these two theories has been suggested by J. Bloch, and it seems highly probable that Pali - and Prakrit -su were derived from the dtmanepada ending -a, and that the final vowel was adapted to that of the third person singular ending -()u. The single consonant in the Prakrit ending was the result of further remodelling. In the heat of the discussion over the origins of the ending su, problems surrounding the other forms taken by the second person singular of the imperative have often been overlooked. The distribution of these forms is of interest, and may help to throw light on the development of the imperative in general. In the Ardhamagadhi of the Jain canon we find only rare survivals of the Vedic and Sanskrit use of the bare stem for the second person singular imperative in the thematic conjugation. The ending -su is hardly ever used, and then only in metrical portions of the canon: for instance in the Savagadanga there is a passage (1.4) which describes the joys" of married life and contains a long series of commands addressed by the lady to her husband, and yet the termination -su only appears once, in genhau "take". The use of the Second person singular termination -li on the other hand, which was confined to the athematic conjugation in Sanskrit has been greatly extended in Ardhama gadhi. It is found in verbs ending in the vowel -o and especially in Verbs ending in -e. As is often the case, the analogical extension of the termination seems to have started from a few common Verbs: eli "go", dehl "give", dhehi "place". From there-hi seems to have become associated 1 Ch Lassen. Tr o nes Line Prakriege (Bonn, 1837), pp. 179, 338. A Weber. Uber das Sapatarakam des Hala (Leipzig, 1870), p. 61. *H. Jacobi isewhle Erwagen in Maharashir (Leipzig, 1886), p. 54. * F. Edgerton, op. cit., p. 147. R. Pischel, Grammatik der Prakril-Sprachen (Strassburg, 1900), p. 331. * G. V. Tasare, Historical Grammar of Apubhra (Poona, 1948), p. 298. "J. Bloch, L'Indo-Aryen (Paris, 1934), p. 249. - 148 - - 149 -

Page #85
96 A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN with the vowel -- and its use was extended to verbs ending in -e from -ya of the causative. In the Jain canon-hi was furthermore the most common termination of the second person singular imperative of all other verbs, and before it thematica was lengthened: c.g. khandhl "dig". payacchi "hand over" (Siyagadanga 14). No real reason for the lengthening of the thematic vowel before I has as yet been advanced. Final-dhi was etymologically justified only in verbs of the second class ending in long-d, such as ydhi "blow", pdhi "protect" and the much more usual verb ydi "go". Just as the ending-chi was extended from chi "go" and some similar forms, dhl "go" may have exerted a parallel influence. This influence spread very early to the ninth conjugational class. The old second person singular imperative ending dna, taken by some verbs of the ninth class in Vedio and Sanskrit, had no support within the conjugation system and disappeared without a trace. As the third person imperative of a ninth class verb was jand(tu "he should know", which resembled law, etc., Middle Indo-Aryan speakers began to use a second person singular andhi, corresponding to dhi. From the ninth conjugational class the termination-dhi became associated particularly with a preceding. and the analogy was not always felt in verbs of the ninth class where this was obscured by phonetic developments; whence genes instead of gendhi "take" in the passage of the Siyagadangauffa already quoted. This association with was very close and was carried to verbs of other conjugational classes in the Sauraseni and Magadhi of the dramas, where the ending-dhi is otherwise unknown and only w is used, e.g. dhl "listen", bhandhi "speak" (Malavikdgnimitra). Despite this extension of -dhi in Saurasent, the link with the ninth class is still the strongest: bhanasu may occur Occasionally in Sauraseni alongside bhanall, but jdmasu has not been found despite much searching. In Ardhamagadhi the use of the long before-hi was so general that it seems likely that yet another influence was at work, quite apart from the influence of jl, which affected the ninth class, and hence verbs ending in-. It is probable that the imperative was linked by the speakers with the optative, with which it was so closely associated in meaning: e.g. in the same passage as the list of imperatives quoted from the Swagadargutta, there occur the optatives carejasi, vaijeljd, miluje, parivaljdsl. It is certain that the old subjunctive mood, prior to its disappearance had some influence on the imperative, and this may account for some Asokan forms: S (Kalsi), "may he obey," palakkandu and others quoted by Bloch as well as the curious plural suru (Girnar). "may they hear." But the subjunctive disappeared so early that it seems much more probable that the imperative and the optative, both used side by side in a similar context, were associated by early Ardhamagadhi speakers. The fact that the lengthening of the vowel immediately preceding the ending-himight to some extent have emanated from the optative is confirmed by some unusual forms in Ardhamagadhi where there is actually a mixture between imperative and optative forms: Yanded, "may you praise," paljudeji animanidhi (Udsagadando). A different mixed form is found especially in Maharastri: kualdo, Mhandisu, sad , sar di (Lildwalkahd), kuid (Valjdaggam, where the chidyd has kurvyda), khameds (Palma cariya). Lengthening of this kind is quite unusual before the endings, it is found only in A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN 97 association with the optative. But in the case of the ending-hi in Ardhamagadhi the further extension of such a long form was facilitated by the existence of the analogical forms based on jdhi and on the ninth class as well as the very frequent causative ending-chi, which brought about a rhythmical association between the ending-hi and a long vowel. The distribution of the forms of the second person singular of the imperative in other dialects is quite complex. In Jain Sauraseni, as illustrated by the usage of the Bhagavati Andhand, the termination e-l is found with verbs ending in , including the causatives, as in Ardhamagadhi and Sauraseni, e.g. cintei "think", nisevel "attend to": but forms like cites are equally frequent. Apart from this the ending -hi is confined to occasional use with verbs belonging to or modelled on the ninth class: jddl (Bhag. Arddh. v. 803). vijanah (Moldedra). Sometimes the vowel is short before this ending as in millahi (Bhag. Arddh. v. 1516). This may have been a feature of popular speech in the west central area of India and foreshadows Apabhramsa developments. But the use of his still comparatively rare in Jain Sauraseni and the majority of verbs, including even those based on the ninth class, add the usual second person singular ending - susu "listen" (very frequent). Jinan "overcome". - is also the normal ending of thematic verbs: labhas, gachas, etc. The use of the bare stem in thematie verbs as in Sanskrit is rare. Jain Maharastri resembles Jain Sauraseni as regards the endings of the second person singular of the imperative, except that there are occasional instances of the use of -dihi in verbs where it was justified neither by etymology nor by the analogy of the ninth conjugational class. This was due to the influence of the canonical language. The bare stem is used as an imperative quite frequently in Jain Maharastri prose and verse, of which one might take the Samardecakahd as a typical example. The ending su is also widely used. Usage in popular Maharastri texts is very similar to this, except that forms in-dhi are much rarer where not due to direct causes: e.g. in the Vasudevahindi an extensive search revealed only the one form Khamdi "endure", but frequent instances of sundhl "listen" (adapted to the ninth class) and even md bhadhi "do not fear" (based on the analogy of verbs ending in a long-d). In an even later and more popular text, the Kuvalayamdd, there has been a further change: there is a huge preponderance of forms in-s, while -di does not even occur in the few verbs where it still persisted in the Vasudevandi, e.g. "listen", md bihan "do not fear". In the causative and other verbs in the ending-chi has become extremely rare in the Kuvalayamdld, although there are some instances of a new reduplicated form de-dehi, alongside desu. There are some isolated examples in other popular Maharastri texts of the Apabhramsa tendency to change final a tou when the thematie stem was used as an imperative of the second person: e.g. bhanu "speak" and "stay" occur in the Valldaggam. These forms in - ultimately became quite frequent in Apabhrama. It is therefore evident from the texts that the ending-hihad undergone an extension in Ardhamaigadhi that was not shared by any other early Middle Indo-Aryan dialect. Magadhi, Sauraseni, Jain Sauraseni and Jain Maharastri all agree in using the ending-hi only in certain groups of verbs, while in the popular form of Maharastri it was almost non-existent, its place being taken mainly by-s, and also by the use of the bare thematic * J. Bloch, Les Terons dAsaka (Paris, 1950), p. 77. R. Pischel, op. cit., p. 327. Mixed forms india, with a short a are also found in Ardhamagadhi. Jain Maharastri and Jain Saurasni, .. m as (Karlowpakkia, . 89). Vajdaggam, ed. J. Laber(Calcutta, 1944), Fasc. I, p. 11. - 150 - - 151 -

Page #86
98 A STUDY OF SOME FEATURES OF THE IMPERATIVE IN MIDDLE INDO-ARYAN REMARQUES SUR QUELQUES CONJONCTIONS DU MOYEN INDO-ARYEN L. A. SCHWARZSCHILD stem. This aloofness of Ardhamagadhi from the other Middle Indo-Aryan dialects is evident from many features of the verbal system, c.g. the infinitives in itae. But Ardhamagadhi did not stand altogether outside the general stream of linguistic development; it exercised a profound literary influence on the non-canonical writings of the Jains, which accounts for instance for the use of some di forms in Jain Maharastri. Ardhamagadhi was a language of the east central area that had gained early literary eminence. As such it had many points of resemblance with the language which was spoken in the west central area and which was later moulded into the literary language Digambara or Southern Apabhrama. Some of these points of resemblance extend even further into the western region and into Western Apabhramia. This applies to the ending-dhi of Ardhamagadhi, which appears in Apabhramla as-ahi, the most frequent termination of the second person singular imperative. The analogical extension of the long vowel in Ardhamagadhi does not appear to have affected the area where Apabhrama was formed, and -ahi is found only with roots ending in d, c.8. dhi. The use of the short form -ahi had been foreshadowed already by isolated examples in earlier Middle Indo-Aryan, as in ni jinshi (quoted above from the Bhag. Aradh.). Other Apabhrama endings of the second person are -, based on the bare thematic stem, and also e. I borrowed from the optative, as well as probably from the second person plural. The su forms have almost completely disappeared in Western Apabhramsa, but they still existed in Southern Apabhrama. The Apabhramla forms are clear antecedents of the modern imperative, where particularly the bare stem is represented in Hindi, Panjabi, Marathi, Bengali, etc.), as well as the bare stem with a final - (Sindhi), and there are also remnants of the imperative ending -ahl, particularly in the older forms of the modern languages (e.g. Old Hindi, Old Awadhi). The imperative ending - has left an interesting survival in Old Awadhi: this is an instance of the resemblance that has been noted between Eastern Hindi and Southern Apabhramsa. Already in the Uktivyaktiprakarana, attributed to the twelfth century, there occur future imperatives of the type n anesu, padhes, which are still found in 16th-century Awadhi, for instance in the works of Jayasi and Tulsi Das (e.g. kahes "you shall say). There is also in Old Awadhi a second person plural of the future imperative in-chu. Attempts have been made to explain the endings and chu by means of the Middle Indo-Aryan causative . The meaning of the future imperative is, however, distinctly associated with the future rather than the present tense, and it would therefore seem probable that the Southern Apabhramia type of future in esami, -esahl, etc., was responsible for the use of the characteristic vowele. To this were added the personal endings of the imperative - second person singular) and -u (second person plural) to form a future imperative The development of the second person singular endings of the imperative can only give a glimpse of the changing pattern of regional distribution that is so characteristic of the development of the verbal system in Middle Indo-Aryan. S'il est vrai que le style periodique et l'usage des propositions subordonnees appartiennent surtout aux langues cultivees et raffinees, et que la paralaxe est du ressort des langues dites & primitives, on s'attendrait a un essor remarquable de la subordination en sanskrit classique. Mais il n'en est rien. Le sanskrit, si apte a exprimer les pensees les plus complexes et les plus enchevetrees, se sert surtout de la composition nominale et des constructions absolutives, et neglige les phrases subordonnees. L'usage - toujours assez restreint qu'on fait de la subordination en sanskrit depend de l'epoque et surtout du genre litteraire ; le style des brdhmana et des bhasya se signale par les propositions liees sous forme corre lative!. L. Renou a montre que les parties descriptives du kuya evitent tout a fait la subordination, tandis que le style oratoire, qui pourrait bien se rattacher au style bhastja, se sert parfois de phrases relatives, meme compliquees. Les textes prakrits qui nous sont parvenus ne sont souvent qu'un reflet du sanskrit, mais dans quelques textes moins stereotypes on entrevoit une souplesse et une richesse d'expession toutes nouvelles, et ce sont quelques-unes de ces innovations qu'on voudrait examiner ici. Nombre de textes prakrits, y compris les vers maharastris du drame, les parties descriptives du canon jaina, aussi bien que les passages descriptifs du kanya apabhramsa, se laissent traduire presque mot pour mot en sanskrit. L'influence du style descriptif du sanskrit a predomine au point que la syntaxe et la tournure des phrases sont restees assez proches du sanskrit, malgre tous les 1. AMINARD, La subordination dans la pro vedie Elode sur le Salapatha Brahmana, 1) Arnoles der Universite de Lyon, 3. Paris 1936. 2. L. RENOU, Sur la structure du klivs J.A1950, s. 1, p. 11. For examples of this in the development of the future system, se JRAS, 1953, pp. 50-52. B Saksend, The Evolution of Allahabad, 1937), pp. 269-270.

Page #87
672 L. A. SCHWARZSCHILD changements de phonetique, de grammaire et de vocabulaire. Mais dans les textes plus populaires qui ont subi l'influence de la langue parlee, dans quelques inscriptions, et surtout dans les oeuvres en prose, on peut distinguer deux courants nouveaux : d'une part il y a une tendance a la parataxe, d'autre part on peut constater des methodes nouvelles de la subordination. Ce sont toujours les propositions correlatives qui jouent un role important. Dans les inscriptions kharosthies trouvees au Turkestan il y a meme des distinctions modales dans les propositions subordonnees. T. Burrow a remarque1 que l'optatif y signale une proposition relative a valeur generale: gesa vivada siyadi rayadvarammi visajidavya ceux qui se disputeraient doivent etre envoyes a la cour du roi . Le futur sert a exprimer quelque chose de plus defini et de plus reel: go mamnusa... abomala karisyati... se isa visajidavo l'homme qui sera desobeissant, on doit l'envoyer ici >>. Le moyen indo-aryen se rattache ainsi a l'usage de la prose vedique ou l'optatif confere la nuance indefinie. Dans les phrases conditionnelles la langue des inscriptions de Niya se sert ou de l'optatif ou du futur, selon que la condition est plus ou moins probable. On ne se sert jamais de l'indicatif du present dans la phrase conditionnelle. Apres jena > (Kalhakosaprakaranam). Parfois, meme dans ces textes populaires et tardifs, on pourrait entrevoir une continuation de l'usage sanskrit, ou yad peut introduire des propositions completives: sohanam jam aham parancio imie dullhasilae e il n'y a pas de doute que cette mauvaise femme m'a trompe (Jambucariya, p. 106). Mais c'est la nuance explicative et causale qui est beaucoup plus frappante : na lajjasi lumam imanam niyapandarakesanam jam lie... neuram pariginhasi << tu n'as donc pas honte de tes cheveux blancs, que tu aies pris son ornement de pied? (Jambucariya, p. 106). jam peut alors changer de place avec jena 'pour que', et on lit dans le meme passage du Jambucariya: na lajjasi tumam... jena evam bhanasi? tu n'as done pas honte... que tu parles ainsi ?> L'extension de la conjonction subordonnante jam a enrichi la phrase moyen indo-aryenne d'une souplesse toute nouvelle. Les causes de cette extension sont multiples. La structure correlative, si importante en sanskrit classique, s'est affaiblie dans la langue populaire, et les liens entre les demonstratifs et le relatif ont ete relaches. L. Renou1 a releve dans la prose vedique des phrases qui temoignent de la simplification du demonstratif, senti comme element abstrait, et on trouve done yatha repris par laval, yasmat repris par tena, etc. En prakrit, meme dans les textes anciens, cette tendance a prevalu, et les phrases correlatives irregulieres abondent: jam est souvent suivi de la, ou de taha; jatha et ladhima 1. L. RENOU, Grammaire sanscrite, Paris 1961, p. 530. - 155

Page #88
674 1. A. SCHWARZSCHILD sont associes si souvent dans la Bhagavali Aradhana que l'on pourrait considerer cette expression comme caracteristique de la langue de ce texte. Parfois, surtout en jaina maharastri, il est difficile d'etre sur si la... jam est correlatif, ou si l'on est en presence d'une particule temporelle ta, suivie de jam, conjonction completive ou consecutive: la sukayam kayam jam calio asi e tu as donc bien fait de t'en allers. C'est ainsi sans doute que jam s'est degage de plus en plus de la construction correlative et est devenu une conjonction subordonnante. Les changements phonetiques et morphologiques ont contribue a l'affaiblissement du correlatif : etad et yad se ressemblaient bien et s'employaient tres souvent dans les phrases correlatives du sanskrit. Mais en prakrit elad est devenu e(y)am, et ne ressemble plus a gad, devenu jam, et l'on a donc de la peine a reconnaitre le schema correlatif eam... jam dans une phrase comme kaha sakka bhanium aliam eam jam porinasuie Bharaha-Ramayane ayam comment peut-elle qualifier de mensonge ce qui nous a ete transmis par les anciens textes sacres, le Mahabharata et le Ramayana? (Dhurlakhyana 3.20). Dans les textes populaires posterieurs a la Vasudevahindi, surtout dans la Kuvalayamala, l'usage de jam comme conjonction subordonnante s'est donc de plus en plus repandu. Mais jam n'est pas la seule conjonction subordonnante a sens causal et consecutif: l'ancien instrumental du pronom relatif, jena, s'emploie dans le meme sens, comme dans l'exemple du Jambucariya, cite plus haut. L'usage semble varier d'un texte a l'autre ; quelques ecrivains, comme l'auteur du Kathakosaprakaranam et l'auteur du Dharmopadesamalalika preferent jam, d'autres comme Silanka, auteur Caupannamahapurisacariyam font un usage tres etendu de jena. Parfois l'origine correlative de la conjonction jena est claire, et des expressions comme kahim...jena, et kisa...jena sont assez communes: aha so una kahim maha satta, jena se pariyanami balavisesam ou est donc mon ennemi, pour que je puisse connaitre la nature de ses forces (Caupannamahapurisacariyam, p. 224), et kisa...asabbhavini samjaya, jena mamam pi avaharasi vayanam est-ce que tu es devenue malade, que tu detournes ton visage, meme de moi? (ibid. p. 239). Mais pour la plupart, jena est une conjonction causale et consecutive, sans antecedent demonstratif: sasallo viva bhagavao tanu samuvalakkhijjai, jena pecchasu.... milanalavannam uvalakkhijjai vayanakamalam le corps de notre maltre est comme perce par une fleche, puis que vois donc le lotus de son visage semble fletri>> (ibid.). jena peut meme remplacer jam dans les propositions completives: na ya eriso susahana samayaro jena taddivasam bhikkham gaheuna puno vi geham agacchijjai ce n'est pas la coutume des religieux de merite qu'ils - 156 CONJONCTIONS DU MOYEN INDO-ARYEN reviennent a une maison ou ils ont recu l'aumone ce jour meme (ibid.). 675 Dans le dialecte jaina maharastri du Caupannamahapurisacariya l'emploi de jena etait donc tres en vogue, mais on n'y neglige pas les autres conjonctions subordonnantes, et jaha y joue un role important. jahd, comme gatha en sanskrit, y sert a introduire une citation; le tour direct est generalement precede de jaha et suivi de 'lli>, etc. Il est evident qu'il ne s'agit plus d'une pensee, d'une citation ou d'un discours direct, introduits pas jaha, mais d'une proposition subordonnee. La personne du pronom ou du verbe peut bien en servir de preuve. Comme la plupart des propositions directes ou indirectes sont a la troisieme personne, ces temoignages sont assez rares, mais il y en a dans la Kuvalayamala: bhanio ya sanunayam kumaro raina 'pulla ma evam cintesu jaha aham tumhanam satta... tumam ca mama putto 'tti le roi dit au prince avec tendresse : fils, ne pense pas que je sois ton ennemi. Je suis ton pere... (Kuv. 11.8). Si jaha servait ici a introduire un discours direct, la phrase aham tumhanam sattu n'aurait pas de sens; on s'attendrait a un pronom de la deuxieme ou de la troisieme personne, et l'on devrait done traduire fils, ne pense pas : il est mon ennemi >>, ou fils, ne pense pas vous etes mon ennemi >>. Dans quelques textes jaina sauraseni a influence populaire jaha figure dans les propositions conditionnelles: appanam pi cavantam jaha sakkadi rakkhidum surindo vi to kim chandadi saggam savvullama-bhoya-samjullam? Si le plus puissant des dieux etait en effet capable de s'empecher de tomber (du ciel), pourquoi quitte-t-il le ciel si plein de jouissances celestes? (Karlikeyanupreksa, 29). -157

Page #89
676 LA SCHWARZSCHILD Dans ce texte aha...lo (peut-etre derive de jahd... to) sert souvent a introduire des propositions conditionnelles. Il n'y a aucun doute que l'usage des conjonctions subordonnantes s'est etendu dans le dialecte jaina Sauraseni, et dans les textes en prose jaina mahariri. Cette souplesse d'expression toute nouvelle ajoute beaucoup de charme aux contes jaina. Mais cette tendance n'a pas dure, et dans la langue parlee la paralaxe devait jouer un role de plus en plus important. Par consequent toutes les conjonctions subordonnantes qu'on vient de discuter, jam, jena et jaha ont disparu. Des le temps de Saraha l'indo-aryen a comble cette lacune par l'emprunt de la conjonction subordonnante persane ki. Mais le systeme de subordination syntaxique n'a jamais evolue a un etat tres complique en indo-aryen moderne. L'usage des conjonctions subordonnantes en jnina Saurasenl et surtout dans la prose jaina mhariri represente donc un developpement interessant, mais assez ephemere et quelque peu en dehors du grand courant du developpement de l'indo-aryen. Some Interrogative Particles in Prakrit L. A. SCHWARZSCHILD Victoria INTERROGATIVE particles, such as words meaning 'why' are not subject to many of the semantic influences that bring about the loss of words, and yet such particles are very liable to change. They are constantly overshadowed by the interrogative pronoun and may often be replaced by more specific and intense expressions such as for what reason'. Such expressions are generally emphatic and may even border on slang, as for instance English 'why on earth?' and whatever for?' and they are therefore particularly prone to change with linguistic fashions and even with the taste of individual authors. This can be illustrated from Middle Indo-Aryan. In Sanskrit the sense of why?' was conveyed usually by Tasmat, the ablative singular of the interrogative pronoun; a reason was asked for more specifically by kena karanna for what reason?! A rather more vague inquiry for a cause could be introduced by the neuter of the interrogative pronoun, leim, which was often strengthened by the addition of the particles ti, kehalu etc. Of these expressions leasmat has survived occasionally as kamha, the ablative singular of the interrogative pronoun in Prakrit, but it was no longer generally used in the sense of why?'. In the Svetimbara Jain canon the other two expressions of Sanskrit, kena learned and lime maintain their popularity, but they are often used in fixed locutions peculiar to the

Page #90
SOME INTERROGATIVE PARTICLES IN PRAKRIT: 205 206 : SHRI MAHAVIRA JAINA VIDYALAYA GOLDEN JUBILEE VOLUME canon. The most striking of these locutions is the use of the slightly emphatic and adversative particle re to introduce a question. This particle has been derived by Pischell from Vedie sed, sa+id. This derivation no longer seems tenable on account of the Pali evidence, as given for instance by M. Mayrhofer, and from the evidence of Middle Indo-Aryan in general: the distribution of the particle se shows it to be quite clearly a Magadht form of the neuter singular of the pronoun sa and equivalent to the form tam Page #91
SOME INTERROGATIVE PARTICLES IN PRAKRIT : 207 208 : SHRI MAHAVIRA JAINA VIDYALAYA GOLDEN JUBILEE VOLUME samiya Page #92
SOME INTERROGATIVE PARTICLES IN PRAKRIT : 209 Brid Communication OL The more popular Jaina Maharastri texts already show signs of new developments; thus kiha "why' appears in the Vasudevahindi (92.16), kiha bihesi 'why are you afraid?! This word must probably be explained from katham how?' influenced by kih. New forms based on the interrogative pronoun, particularly the neuter plural lain, become prevalent in Apabhramsa, and a new cycle of fixed locutions begins. wedish Deutsche... Wie Desser Past is t in einigen Pul Paste for Dentale inge s of the distinctive vaba while of phone in any disetele for further invest Only a few Midden-Aryan have been quoted by Picbel showing changes between palatasaad destals, and the low are mainly derivatives of the desiderative of certain verb: pair. The story of a derivatism in the there wa s form of the mon t h the talem b ag. 18. in connect with arith metlicontaining mum p ala mainly in the works of the r i se Varsel T e m a 10,7. It is very probable that was the inted by the mentally very close word o o two, well as by die this would explain the presented the vowel in dega richat de pi p AME www. > Hant hatto be hungry Prskrid AMA, JM d e depine AM JM dupa douch AMES they wendy is by d e par. Hindi jo paie, but Kemiller. The sociation between pale pale and wurde mening was obvious that l e d for any further explanation of this contamination The derivation of drid, dois m ight from your by M arghilorward p eally and antially. The specialised m in moonlighi preval et is well as th i ng of light. The of this specialiste s b y laktaserye by the panel for A. , La Mi and Made IndoAryss only the special n ing survived Pal m eonlight moonlight whe, bright forscht of the most do l or night, moonlight Prikrita , dond, et moonlight Hindi light, Nepall moon, Sina moon It is the desire > 7 > par HE juga, Deltamala V. 49) d air d Tom. IV 194 tot chai ochop' tashaw by eating of hello par to 00068 AMJMM Aps Sched with it. This to stand may be implis the per Hem. given by condut by the stos Arydo r ash in which there is no ar , though her other very posible > klllknation of turpugate Tres iron stails" ( s wdhara Apa hare papan) bew ed. There is obsberat (Via IV. thention between doors, nicht and other wondering the on: Sank a ra, Prikrit The change the dental healt o deal mone, night-makeranaki dom in the few examples are due to a variety of * w i ng. Die La Berlin 2005 1 CONTAMINAS P. Sehr Gel o pair and dow n light are set of all , Heidelberg , 73. Ale K Hoffman phoneti interes; the change the initialatalo OMG 110, 10, 151 dental and in the words is due to contami. Turner, Comparatie Dia ry of the Indosation Aron Long , Fe, IV. p. 20, Oxford 1964 It has alwly been suggested by Paobel that the Mali Tel Society's Pala Dictionar development of ma p ato da could have been London - 332 W Adam, Early due to istion with AMG. IM, duke Nistory Co, Coba 10,3 Some a rd-change in Prakt The development of Old to Middle Indo-Aryan is bar in the course of the evolution of the Prikrit dialects. In stedsinly by profund changes in phonemediate the s on of major phoneme change in any Prakrit bution, and especially in the catering of consta dinlet od particularly in thes e of any changes There are only comparatively few major phone involving whole group of th e other changes and only a reduction in the number of the thesibilasta) one st o re prised to read in poemes it is my try that only on the great work of Pa t ie wens die phomes o marginal tit l e have been fototisch mit Labiale... An der Palstalen finden 18. M. Katre, Problems of Recoration in Argon, Simla 1958, p. 7 de R. Pachel, Grown burg 1900, p. 215. der Prakrit Sprachen, Bursa - 164 - 165

Page #93
102 'moon (lover of the night), and also Prakrit dosasijjanda, dosdragana (dood-ratna 'night jewel") 'moon. It seems therefore quite probable that dosing a clear night, moonlight' representa a contamination between poland moonlight' and doet 'night', and the presence of the initial d in this word does not represent the result of an ordinary phonetic change. Journal of the American Oriental Society, 98.1 (1978) 2. ASSIMILATION Sometimes hesitation between palatal and dental consonants is due to assimilatory influences: this is certainly the ease in vijjajjhars, a rare Apabhrapta form for thara < dyddkara. The change of I to e, ok in quechs, chuccha < tuccha 'small is also due to assimila tion: the tendency to retain the sasimilated form was aaniated by the presence in Prakrit of a number of very similar words meaning 'small' and beginning with e, ch: ella, challa, chulla, chudda. The nearly homophonous Prakrit word cumcha 'dried up' (probably not a Dest word bat connected with Sasakrit fuska 'dried up) appears to have followed the pattern set by escola 'small', hence we find an alternative form tucka 'dried up' in the Deedmandia III.15. The Sanskrit verb take (teaky-") regularly becomes tacchi "to cut' in Prakrit and occurs frequently in the Jain canon particularly in the description of various tortures, e.g., Sayagadanga 1.4.1.21: avi tepatbinde tacchiya khara-candin ya "they are rosated alive and seld is poured into their wounds. The nasimilated form cacchai < tacchai is listed by Hemaeandra IV 194, and also by Kramadivara and Markandeya, cacchai, though not found in the texts, must have been current over a wide ares, particularly in eastern and central Indis: a glance at the Comparative Dictionary of the Indo-Aryan Languages by R. L. Turner reveals the interesting fact that the distribution of forms with initial c is very similar for both take- and fuccha. The main similarities in distribution are as follows: Hindi Bhojpuri Bengali (Oriya) Panjabi (Oriya) Initial e Jake cond edchal edcha cachid, hdchild Initial : lachd 1dchild Juha chick cch AGA cucha fucc tuck The above list indicates that this very sporadic sasimilatory change (of tor e in lake, fuccka) occurred over a M. Mayrhofer, Uber Kontaminationen der indoirsnischen Sippen von si taky, teaky-, "wark Indo Iranica, Melanges presentes a Georg Morgmatierne, Wiesbaden 1964. wide area. The list also gives an interesting illustration of the complexity of the development of the Indo-Aryan languages. The basic scheme of 'Controlled Historical Reconstructions" would never account for a case of this kind where there is correspondence between the otherwise not very closely related Panjabi and Oriya languages: the network of dialectal relationships in Middle and Modern Indo-Aryan can only be accounted for by 'sets of intricate interrelations as illustrated in the important article by F. C. Southworth on Family-tree Diagram 3 POBLE TENDENCE TOWARDS DISSIMILATION The remaining examples given by Piachel to show variation between palatal and dental consonants are all, with the exception of tighati, desiderative verbs. It is highly improbable that a change which is purely phonetic, such as a 'dialectal variation between dental and palatals' should be almost entirely confined to a small and comparatively rare morphological group auch an desiderative verbe. Morphological factors most inevi tably have played some part, rather than purely phonetic developments. As is well known, velar consonants are represented by palatala in the reduplicating syllable in Sanskrit verbs. This pattern was so dominant as to lead to analogical formations, eg, cit-'to perceive' perfect ciketa, desiderstive nikita (the k in these forms is analogical"). In Middle Indo-Aryan a few desideratives have survived and they were distinctive enough to remain as a group. It was precisely in this group that the change of > d e>foccurred, and this change would appear to be one of dissimilation": the speakers had the feeling that the first syllable differed from the second syllable and they tended to differentiate the initial consonant even further. Bence cikita- AMg, tigicht etc., 'cure', Pali tikicchati fighate- AMg. diginckd ete., "hunger jugupe Pali diguecht 'avoidance' (AMg, duc probably represents a contamination between a form of this kind and jug), As dissimilation played some part in the development of this group of desideratives, it might appear as if D. P. Pattanayak, A Controlled Historical Recon atruction of Oriya, Assamese, Bengali and Hindi, The Hague, 1966. F.C. Southworth, Family-tree diagrams, Language XL 1964, p. 562. L. Renou, Grammaire Sanscrite, Paris 1961, p. 52. 8. Sen, Comparative Grammar of Middle Indo-Aryan, Poona 1900, p. 54. H. Laders, Beobachtungen uber die Sprache dee buddhistischen Urkanona, aus dem Nachlass herausgege ben von E. Waldschmidt (Abhandl. d. Deutschen Akade mie d. Wissenach. Berlin 1962/10), p. 100. -166 Brief Communications cithai < tighati were the only clear example of varia dence it est be deduced with some certainty that the tion between palatal and dental con fricative element in the affricate pronunciation of e, ch stern dialects. This is perhaps what is implied in the J. retained ita palstal quality for a longer time in the statement of Vararucit casarparpa apastal lathalapa The desideratives listed above are interesting in their (11.5), eh, J. Jk and 6 are so pronounced as to be clear. distribution. The forms with the initial dissimilatory It is probable that in the east this distinct palatal frie dental consonant cerur only in the eastern dialects of sive element was retained even longer before the high Middle Indo-Aryan (Magadht, Ardhamsgadht) or in front vowels, and that it remained even when the plosive those which underwent certain esters influences (Jain element of the affriesten (tel. [d] beesme depalatalised, Mahiristel) and to a lesser extent Pali); they do not oosa in [tgl. (di). There was thus a close similarity between eur in Baursen and only rarely in Mahiretet: thean di alects generally have juguld avoidance', dichas 'healer. This dialectal limitation makes it probable that the change of palatals to dentals in the desiderative verbo was facilitated over much of northern and easter India by close phonetic similarity between palatal + on the one hand and dental + on the other. + (pronounced (teil), (pronounced djip) on the one hand, and ti, di on the other. This phonetic similarity was accentuated by the fact that the high front vowsli would have the effect of slightly retracting the point of articulation of the preceding dental to at least an alveolar if not a post-alveolar position. The problem of the pronunciation of the palatal series of consonants in Middle Indo-Aryan was discussed by Grierson," who came to the conclusion that in Standard Mahratri Prikrit and in Baurnannt... the palatals were probably pronounced as dento-palatals, as in modern Marathi, but that in Magadht they were pronounced clearly as true palatals. This is based mainly on Markao deys's statement (xii, 21) about Magadht; cajagor upari yah ayll, 'y is prefixed to c and j. Markandeys's statement need not be taken literally, but it clearly indicates, as le suggested by Grierson, that the 'palatal' series of consonants had a more distinct palatal quality in the Eastern dialects of Middle Indo-Aryan. There is no proof that they were "pure palatale se Grierson auggests. e, oh, J.JA were still palatal plosives at the time of the Prsti Akhys literature. There are indications in Middle IndoAryan that at least in the western areas of India these palatal plosives had become affriestes and that e represented [> and j represented id> -- apatys > apsons in the Girnar version of Aloka's 5th Rock Ediet, apen in the Shabbatgarhi version, but apatiya in the Kalai and Dhauli inscrip tions. The modern languages are divided mainly between a more easterly alveo-palatal affriente, and amore westerly palato-alveolar and even denti-alveolar affrieste pronunciation of e, ck, j.jh. From this and the Akan vi G. A. Grierson, The Pronunciation of Prakrit Palatale, JRA8 1013, pp. 301. 14W.B. Allen, Phonetics in Ancient India, Oxford 1963, p. 52. 103 8rt Kundakundearys's Prassenandra, edited by A. N. Upadhye, Bombay 1935, p. CXIII. H. Jacobi, Anaprushlie Brahlungen in Maddah, Leipsig 1886, p. 59, 1.27. - 167

Page #94
104 Journal of the American Oriental Society, 98.1 (1972) homophonous clash also accounts for the comparative rasity of cethai "to strive' in Middie Indo-Aryan and its atanty survival in the modern languages. The neutralization of the opposition between palstal and dental consonants can secount for a number of other irregular forms that have come to notice since Pischel's grammar was published. A series of such forme have been quoted by Laders in his study of eastern linguistic fes tures in P ji> di in Aloka's Rock Ediet X at Kalsi palitidis (Dhaali, Jaugada politidis), the Pali name of King Prusensjit, Pasenadi Pali vidita (ita) 'conquered", Pali palissadissatiPage #95
would be feasible to establish some hypothetical di minimal pairs: e.g. faloyati SCHWARZSCHILD: Initial Retroflex Consonants in MIA "he disturbs Mlapati 'he establishes dama dama a mixed caste "self-restraint" but these would not reflect Sanskrit literary as age, still less would they reflect any dialect or stage of the Old Indo-Aryan spoken language. Sanskrit poets who excelled in word-play composed verses in which only certain consonantsand preferably the rarer consonants-are used: bhribhir-bharibhir-hirair-bhabharair abhirehire... Magha, Sisupalovatha XIX v. 66. Yet even Magha did not create similar effects with retroflex consonants, since these were not sufficiently well established in the language. Word-play connected with retroflex initials is found only on a very minor scale, as for instance In the Janakiharanam of Kumaradasa: saradindimatam abhimrajety-abhisekostsamadundubbih kaopdt. (IV.45). The large kettle-drum beaten at the coronation ceremony turns suddenly into a funeral drum. In this verse there is obviously a contrast be tween dindimatam 'the condition of being a dip dime-drum and dandubhib 'a large kettle-drum." There is some emphasis on the use of initial retroflex consonants in the formation of onomato poele words in Sanskrit and literary effects are sometimes created by the repetition of such words, eg.. the famous line fhofhamthan thathatham thathathah (Mahanafaka 111.5), but these can hardly be regarded as part of normal linguistic usage. tanka -titibha doraka dola There can be little doubt that retroflex initial plosives remained an allen feature in Old IndoAryan. There is evidence that the borrowed words with such initials tended to be assimilated to the general phonemic pattern of Old IndoAryan, and the initial consonant has sporadically become a dental. Examples are: Janka 'chisel fifita a high number "to go" -findise 'name of a plant "name of a plant 'name of a person' The changes between dental and retroflex may not always be due to Indo-Aryan. Some of these words may be explicable from hesitations in the languages from which they were borrowed; and they may have been borrowed in situations analogous to what F. B. J. Kuiper calls 'consonant variation in Munda: Munda languages originally did not have retroflex consonants as separate phonemes. It is noteworthy that in Dravidian. Initial retroflex consonants are as uncommon as they are in Sanskrit," and the origin of most of these borrowed words beginning with 1. th. d. dh remains a mystery. More importance than ever must therefore be attached to the possibility that there was yet another element involved-an element which was non-Indo-Aryan, non-Dravidian and non-Munda and which contributed towards the main characteristics of India as a linguistic area," The absence of initial retroflex consonants from the older forms of Indo-Aryan and Dravidian has many parallels, notably in Australian Aboriginal languages. The reasons for this situation are varied: it has been argued that retroflex consonants are often of a secondary nature and this may well be the case in Australia"they may be brought about by assimilation to preceding sounds. But the absence of initial apico-alveolars as well as apico-domal (retroflex) consonants in many languages might indicate that this is a more general matter of both articulation and perception: some of the most characteristic indila dindita Jota -dola Thakkana Thakkana K. Hoffmann, Indogermanische Forschungen Vol. 60, pp. 256 iffdoraka deld to fly" "a strap, "string" awing." 483 F. B. J. Kiper, "Consonant Variation in Munda," Lingus 14 (1965), pp. 54 ft. For retrofles consonants in Munds see also H. J. Pinnow, Versuch einer historischen Lautlehre der Kharia-sprache. (Wiesbaden, 1959), p. 320. 10 E. J. A. Henderson, "The topography of certain phonetic and morphological characteristics of South East Astan Languages," Lingua 15 (1965), pp. 420-422. Para contrary opinion see A. Master, op. cit. p. 3. -170 For a discussion of this problem so T. Burrow, "Sanskrit and Pre-Aryan Tribes and Languages. The Bulletin of the Ramakrishna Mission, Institute of Culture, 1956, and Indo-Asian Culture, Vel. VII, 1960. W. Oates, "Syllable Patterning and Phonetically complex Consonants in see Australian Languages," Pacific Linguistics, Series A. No. 10. (Papers in Aus tralian Linglusties no. 1) 484 Journal of the American Oriental Society 93.4 (1973) features of retroflex articulation are connected with the effect on the preceding vowel and further acoustic investigation of this is needed. THE RETROPLEX NASAL The situation with regard to the nasal con In Middle Indo-Aryan the status of retroflex sonants was not parallel to the history of the initial plosive consonants changed only very slow-plosives; it was much more complicated. ly. Already in Pali and in the earliest ArdhaMagadhl texts there occur a few inherited words in which initial d- and th- have been replaced by d. th. This may be due to assimilation as in the case of das to bite (p-p-p. daffha< data) and dah to burn' (p.p.p. doddha in Mahare tr), but the causes are often not clear as in the large-scale change that affected the derivatives of sthi to stand. In the later Apabhramia texts there are further instances of such changes: danda "stick"> danda; darbha 'bunch of grass"> dabtha; dhranka 'crow> dhamke. At the same time in Apabhramia there was also an increasing influx of further deit words beginning with retroflex plosives. Perhaps the most telling indication of the growing acceptance of initial retroflex plosives is the use made of them in Apab hramia literary works where they occur in rhymes, as in the Nayakumaracaria: ghanta-jankarem .....ali-jhamkureys (VII.1.8) or in the HarivamSapurana: ghutiyal dhaliyal (88.7.9.). dhamkeppipa... laeppinu (85.22.3), and particularly in the Prikytapaingala, where the examples are sometimes dramatic: kad fuffa phuffel manthd... the body is torn. apart and the brain is splattered... (II.183). or sometimes they occur quite normally in alliteration: ambara dambara disa a display is seen in the sky (L188) and in L101-2 where there is a play on initial th In Middle Indo-Aryan there are still only few Instances of meaningful oppositions between initial dentals and retroflex consonants, such as fola 'an insect, tolo- to weigh. One example has already been noted by Fischel "dura becomes dara in the meaning 'fear... and on the other hand the dental remains in the meaning 'some thing, a little, half." There was thus a very gradual change in Middle Indo-Aryan and particularly in Apabhramsa to wards the modern Indo-Aryan acceptance of initial retroflex plosives. It was implied by Helmer Smith that this change could be due to assimilation: Saddani V. (Lund, 1954). P. 1405 14 R. Pischel, Comparative Grammar of the Prikrit Languages. Translated by Sabhadra Jha. 2nd ed. (Be nares, 1965), p. 164. Vedic and Classical Sanskrit had only three nasal phonemes m, n and s: but a did not occur initially and was conditioned in the majority of its occurrences as any student of internal sandhi knows. A was entirely conditioned by the prox imity of palatal consonants, while the velar nasal was significant only in certain rare instances in juncture (e.g., prin-mukha 'facing east), In Pali, far from there being a reduction in the number of nasal phonemes there has been a slight increase in that i, from being an allophone has achieved phonemic status and quite clearly dif ferentiates between words. There are thus four and m nasal phonemes in Pali: A., kana kappa "ear kanna "trickling down kamma "action" and pana pappa panna endowed with knowledge leat 'fallen" Sk. kang Sk. karna Sk. kanna Sk. karma -171 Sk. praja Sk. para probably the past passive participle of pajjati. There are a few rare instances where nn occurs instead of fin (Sk. ng) and vice versa in Pal (eg, sammannatiPage #96
SCHWACHILLI Retro C Journal of the America anin MIA Oriental Society (1973) The widence of the Atoka Interpon the new instances of the change intervalle -> the complexity of the time theater in the sand in Classical Sanskrit en central Asokan iption the palatal but the etym o l some of these may su does not the dental in all be lowed . E m where the expected when the co o l all the individual examples are equally poding Sanskrit words contain and there can be doubt about the existence of a prikriti che of Inter in the G P vetoThe A riptions how Edit from Lure Nandanghand Laura Armas that o f the distinctive value of north AS, P rati wester disruher than escort This is m e nt with the way in which the cha X SEHPA change of intervale t r adusly e otherwise, where panded in Vedie and Carl Sanskrit: develop On the other hand the distinction betwem thens that are cateringin wa a r palatal and the dental m i nd the prominent in their mild increase in late Vedic wete Alo i sition and in the Niyames with the eastward expansion of Indo-Aryan dialect, and the two monteur In contrast, while developments that n orthwestem dial Gimar REVO Gimar RE eta in origin, such exercises and Il is not prstent influence on Vedle and Classical Sans In the case of the ro d ental distinction the position reversed and set out by M A Pyrit texts that of and has been Mehendat the custom Ackn iptions ma rendered complex by the important role of man Lain distinction that is lost in the northwest, Bript traditions and by the emendations on be where intervocalie denn als are generally half of modern editor in Intervocale position replaced by Pall with its fourfold maal w athe rule everywhere, but the majority of contrast m o n thus represents a more com the Jain paper manuscripts at least preserve plex and probably a more resing system of initial and geminated medial The fol nasal consonants than any of the Abkan dialects Towing questions therefore arise The phoneme value of appears to have been 1. Was there any phonem distinction between precarious even in Old Indo-Ary there WB T tendency for to our in complementary 2. Was Initial in seine optione variant distribution within So while ever W o f . Did It wally refert dental rather than med initially, there are already in Vedic examples retroflex articulation 3. OF Wait merely seribal tradition in some medial Jan u scripts to write initial for There can be little doubt that at the time of which can only be accounted for by such o t her texts in Andham andhi there was kritiet development. These are mal jewell, barely honenie distinction between pillar hand in the loved, and and this applied only to the one environ n'hem in the Atharva Veda ment, the medial eminated consonant. Even As pointed out in an important article by M. there many hesitations can be noted. One of the Maybe there are a number of further well- oldest wets of the Sudan is the thip in the Discourse on Wow Thea of this IMA M . Aktris in text has been proved by Alsdorby metrelas (Bombay. 1921. LA "Cant ons well as other considerationBut even for this to the Staty to th e test it is possible to say that and De Che n Yol 30 Sushi - -0-) are kept absolutely apart by De alle tue est, 20 the maps, the chose finds on the whole Men er tre . A ELL, DIL am etymologically correct distribution of und in the ext e ndit to the first word in The following may be noted from Alors The only excepti i pred important s of this text: 2 ) but there follows another citie particle am) ople and the whole moun phrase Isfact phonologically ) 1.24 (can compound thus hinn (CM ) LT) bu is a different environment from n ot henna ( ) LS where the i nitial and therefore dentalIt O ) 11 (Ca cine that this does not would be difficult to refect Carna) teature of actual pronunciation The evidence of the Prakrit gramma but there is some evidence of tesitation, ports that of the Dhammapadalaman writes ne L 223, para sa pad a ) 1.20 ander spawal The same situation is found in other very old nis replaced by after a vow when it is parts of the came mol inatedThis obviously exchdes initial for instance the Uttard for which a separate rule follow hyayama our optionally in the initial position ( c am) 25.4 L NID has made a critical study of chine (cchina) 25.36 the statement of the oldest grammarian, Varana disse (CO ) 27.1 ow (. 2) is substituted for but everywhere. This study shows that the word saratra was probably & Later addition. The dinne Araga). option na t ionally in the initial position It seems therefore that to all practical urposes almost certainly Mands for an original nadon not the distinction between and cannot be con in initial s on. The optional initialis civen sidered to be phonemie in Ardhamad by the other grammarians, whether of the East known from the paper manuscripts, still low or Southern school. The only exceptions known from the palm leat manuscripts. The store Canda who omits rules about and uation generally reflected in the Jain canons and Bharata, who make general reference to that of the Westem kan inscripting the land where the c o nt, pol Initial position allophone of medial between the Vindy and the sea. The fact There is evidence that at least for some time that the main on the whole make special this did representa uneophonte distinctions mention of initial intervocale). V and not simply whim of the scribal tradition support to the view that there has been an earlier This evidence stems not only from the greement allophonte distinction initial media with the northwestern Alkan Inscriptions, but The post Aco Prakrit Inscriptions also a from independent text in particular the firm the evidence of the m ans, but they Gandhari Dhammapada, where excretly the same show gradually more and more examples of the distinction is maintained. It is further show to use of the graphy for the initial at By the be more than a seribal matter by the fact that at fourth century A.D. din varably in all the beginning of entitie particles and pronouns positions. The same situation is found in the is written even in the texts which Svetmbar Jain palmeu ma ripts, in the preserve the initial dental. This can be wen Jain Sau l texts and in Apabama it was the particles indeed has been pointed out already stated by Master that the basal was by J. Browth withion of the Gandh inand Middle EndAryan dasan initial Dhammapadaan enclitie particle forms with the but raphically rather than h a lly and the preceding word a single mit which, phonological tradition was not continued. Nevertheless this ly. Is in effects compound in all it s tradition of spelling has some relevance to pro (L The United by Char l es PP 102 IN TGOMMA MG al Series, vol . 1 Pra , 19 3 3 LIG . 1 1 Vol 1

Page #97
187 SCHWAR SCHILD: Initial Retroflex Consonants in MIA VARIANT FORMS OF THE LOCATIVE IN MIDDLE INDO-ARYAN L. A. Schwarzschild nunciation: the use of the graphy os tiot 2 in Intervalle position, but it remains indicate initial retroflexion, but it does reflect a allophonie Variant in initial position, the loss of the allophonic distinction between northwestem dialectal feature gradually in and it is highly probable that this loss of fencing Vedle and Sanskrit). distinction was linked with an alveolar articulation . The allophonie distinction initial medial of nas found in most modern Hindi dialects. is established, northwestern Alolan inscrip The tradition of writing for this alveolar sound tont, Ardhamgadhl and early Jain Maremained inked with Apabhiramin and is found rastel texts recording to paper manuscripts. even in very late texts such as the Pro and the Gadhari Dharmapada) paign while the modern l writen. The following stages in the development of 4. The allophonie distinction is lost throughout the central area and the east except for Oriya y thus be recognized in Indo-Aryan: and parts of Bihar (note the retention of 1. and are separate phonenses, but only in the Old Kosall of the Unhaktiprakarang) occurs initially, (Vediend Sanskrit 5. Only one r ally alveolar) remains, and V. B. Arun, C rew Phanty of Mind and written as in all positions Jain texts Panjo Luna, 1961) pp. 79-80. For more detailed written on palm-lea, Jain Saurassenland de M. P. Jawwal. Lingute Study Apabrania). del xid, 1963). For comments on the devant 6. This generally alveolar, is written as distinctive features A. N. T anjan and Collo The retroflex nint is absent except in bor M . "Toward Theological Typology of the th r owed words, modern Hindi dialects). din L ustic Area in Current Trade in Lingui, % Aphonemle distinction between andre Vol. List in South A . (The Hago, 1969) mains in Old Kasalt, in Oriya, Panjabi, Sindhi, 562 Gujarati, Rajasthani and Marathi Aditional Note: The present paper was completed shartly before the appearance of T. Burrow's artide This evident from this table that the pronun "Spoleous Case in Sanskrit," SAS 1991 ciation (ol never occur initially at any stage. pp. 35-550, empent study gives further and The als thus illustrate even more clearly than previously unrecognd amples of the change of media the plosives that Indo-Aryan was generally re -1- to Sankrit. luctant to accept initial retroflex consonants, A. INTRODUCTION The concept of 'free variation in linguistics (e.g. Lyons 1969: 72) is very old. The Sanskrit grammarians were fully aware of optional rules called vibhasa (Panini) and vikalba, quite apart from the much-discussed free word-order of Sanskrit (Staal 1967). The term vibhasa is used 112 times by Papini, which is surprising in view of the well-known economy of wording, and this shows clearly the importance attached to this concept in Sanskrit grammar. But these optional rules only refer to very minor points: the majority are rules of limited application referring to certain compounds. A typical example of this restricted application is furnished by the very first rule involving vibhasa in Papini (I.1.28): Vibhasa diksamase bahuprihau 'Pronominal adjectives of the type sarua may optionally follow the pronominal declension if they occur in a bahuurihi compound designating a region. Optional rules of this kind are frequent, but there is little free variation over the major features of Sanskrit phonology and morphology, among the most prominent examples one could quote are the oblique endings of the neuter of adjectives in -i, -, both in the singular and the dual. In Middle Indo-Aryan the position seems to be totally reversed, and there appear to be numerous free variants, particularly in morphology. Some of these forms are not really in free variation with one another, they might at least originally have been regionally and chronologically separate. But often, even within one single text there seem to be - 175

Page #98
L. A. SCHWARZSCHILD a number of variant morphs. Scribal and metrical features may be involved, but these do not account for all cases. It is possible that such forms were not in completely free variation with one another, but were syntactically and semantically conditioned, however subtly, An attempt is made here to illustrate this from the evidence of the locative singular masculine-neuter endings of nouns and adjectives in -a in Jaina Prakrit. 78 Two main types of endings are involved: i. corresponding to the Sanskrit - ii. -msi, -mhi. -mmi and -mmi derived from the Sanskrit pronominal ending -smin. B. REGIONAL VARIATION OF THE ENDING -smin The different forms derived from -smin are without doubt dialectally conditioned. The clusters sibilant + nasal, as in -smin, have undergone changes in Middle Indo-Aryan, but owing to the intricate pattern of dialectal diffusion it is difficult to assess the exact regional distribution of the various developments. That such widely different forms cannot belong to one and the same dialect has already been stressed in the case of -sn- by Luders (1952: 130): 'Ich halte es fur ausgeschlossen, dass die Verbindung des Zischlautes und in demselben Dialekte und noch dazu in demselben Worte bald zu -sin- bald zu-nk-entwickelt haben sollte.' Luders comes to the conclusion that the forms retaining the sibilant are eastern in origin, and those showing the change-sn>-nh- are western. The situation with regard to -sm is basically similar: the change of s to h in a sibilant + nasal cluster is characteristically western in origin. This is shown by the Asokan inscriptions (Mehendale 1948: 26): -176 VARIANT FORMS OF THE LOCATIVE IN MIDDLE INDO-ARTAN 79 northwestern western elsewhere -smin But in the literary Middle Indo-Aryan dialects the situation was more complex, and there was even some differentiation in the development of -smin according to whether it occurred in the nominal declension system on the one hand, or in combination with monosyllabic pronominal stems on the other. The situation can be summarised as follows: Magadhi Sauraseni Jain Sauraseni Pali Nominal Declension -ahim (-e) mmi, -mhi (rare) Ardha-Magadhi Maharastri, Jain Maharastri Apabhramla -smi -mhi -si Pronouns ja, ta-, ka -ssim -mmi, -mhi -mhi, -smim (Sanskrit borrowing) -msi, -mmi, -mmi -177 -mmi, -mi nhim, (-) but note the pronoun assim C. HISTORICAL DEVELOPMENT OF THE ENDING -smin 1. The Sauraseni and Magadhi locatives of the type tassim, tassim, like the Asokan -si reflect 'regular' phonetic changes with the assimilations of -m- to the preceding sibilant (Pischel: 6 65, Sen: 69). The corresponding Magadhi nominal ending-ahim shows the further development of -sm- to -->-->-- This has been explained by the phonetic weakness of terminational elements in Indo-Aryan' (Turner 1927: 230). But the status of terminational element is not in itself connected with phonetic weakness: after all-smin is a terminational element in the locative pronominal form ta-smin. It is mainly a matter of accentuation. In tasmin the consonant cluster -sm-occurred at the beginning of the second syllable where it was in the immediate vicinity of the stress accent (Pischel 46), hence the

Page #99
L. A. SCHWARZSCHILD 'regular' phonetic change to -ss-, -- in Sauraseni and Magadhi; but in a word like gharasmin 'in a house the same consonantal group-sm- occurred in a less accented environment, hence. gharasmin> gharahim in Magadhi. 80 Lack of accentuation also accounts for the loss of the sibilant in the most common forms of the locative in Middle Indo-Aryan: -mmi and mmi. In Maharastri and Jain Maharasti the more accented Sanskrit pronominal forms of the type tasmin show exactly the same development as is found in -smin after nouns, hence tammi, jammi, kammi. This may be accounted for by analogical extension. The locative singular tassim is found only very rarely in Maharastri (e.g. Lilavaikaha 244, 281) alongside the more common tammi. There remains only one general exception and that is asmin> assim, the locative singular of the demonstrative pronoun of vicinity, e.g. in Jain Maharaistri: assim ceva desakale 'at this place and time' (Ausgewahlte Erzahlungen 67.8). This form assim thus contrasts with the entire declension system of the central dialects which is characterised by the locative singular masculine-neuter endings -mmi, mmi. A form ammi, ammi in this is never found, there are however two new forms of equivalent meaning, aammi and iammi (Pischel 429). A locative singular demonstrative pronoun ammi would have been ambiguous and mistaken as an ending in many contexts, and this probably accounts for the exceptional retention of assim as the one phonetically 'regular' form. Owing to the analogical extension of -mmi there is therefore uniformity in the locative endings derived from -smin in Jain Maharastri with only the very minor hesitation between -mmi and -mmi. Another very minor variant is -mmi (with compensatory lengthening of i associated with the loss of final -n). This occurred in the Maharastri of the Paumacariya of Vimalasuri, the Dhurtakhyana and the Nanapancamikaha. There is thus little free option with regard to this particular ending in the most important literary Prakrits belonging to the Central region. -178 VARIANT FORMS OF THE LOCATIVE IN MIDDLE INDO-ARYAN 81 2. In the Jain Sauraseni dialect of the Digambara texts locative singular forms in -mki are not uncommon, they occur particularly in the Bhagavati Aradhana and in the Kattigeanupekkha their occurrence is now so well attested that they can no longer be dismissed as a mistake, as was done by Pischel ( 366). They show a survival of the western traditions of the Girnar inscriptions and of Pali, and they continued to exist alongside the forms in -mmi which had spread from the central dialects. The Jain Sauraseni texts thus show optional variants in the locative forms due to regional literary influences. 3. The regional variants of the locative singular ending as listed above fall into two main groups: a. Those forms of the old ending -smin in which the sibilant (or --) has remained the dominant initial member of the consonant cluster-sm-, namely, -ssim,-ssim, Pali-smim, Magadhi -ahim, Apabhramsa-him. b. those forms of the old ending -smin in which the nasal consonant has become the initial and dominant member: -mi, mmi, mmi, -mmi, -mhi. It can be seen clearly that the final nasal has invariably disappeared by dissimilation in this second group of endings which all begin with a nasal. This evidence is confirmed by Buddhist Hybrid Sanskrit where we find the same two groups of endings: a. -asmin, -asmim and probably esmin. M.M.-11 Arguments in favour of the existence of esmin have been put forward by Roth (1966: 44) and there is also a probable occurrence of a similar ending in Maharastri: eesim ceya ciyanalammi 'in this fire which had been heaped up' (Kuvalayamala 48.3: an interpretation of eesim as etesam is however just possilbe here). b. amse, -amhi - 179

Page #100
VARIANT FORMS OF THR LOCATIVE IN MIDDLE INDO-ARYAN 83 L 1. SCHIWAR SCHILD established in the locative case of nouns of the- and - declension (aggimni, bahummi). They gradually spread to nouns of the - declension at the expense of the original - ending. particularly in the cast, as is evident from the Asokan inscriptions (Bloch 1950: 19). In verse, owing to the contingencies of metre, usage seems to be very free; -smin and can occur in Pali verse, and they occur quite indiscriminately even in the most archaic Jain verses: Ayaramga-sulta 9.1.1. sisiramsi addha-padiranne tam vsaja sattham 'when the cold season has half begun he should abandon his outer garment' here si and are used side by side. These Buddhist Hybrid Sanskrit endings confirm the following general rule for Middle Indo-Aryan: In locative singular endings beginning with a nasal conso nant the final nasal consonant is lost by dissimilation. 4. The Apabhrama ending -ahim is of interest in that it clearly belongs to group a) and has retained the final nasal. It must be derived from amin by means of a development through assim > asim akim > alim. This is already foreshadowed by the pronominal forms of the type tassim, the adverbially used and relatively unaccented Muhuristri tahim Page #101
04 VARIANT FORMS OF THE LOCATIVE IN MIDDLE INDO-ARYAN 85 LA SCHWARZSCHILD (148.26). In the few verse sections of this text locatives in mi are found without any special emphasis: bhavanaramuddama jhamm *in the middle of this ocean of existence' (322.26). The same situation prevails in the Miharastri texts edited by Jacobi (1886). The distribution of locative endings in the Vasudevahindi is thus not a random one, the use of as opposed to shows the difference between early Maharustri prose and verse style. In later narrative texts mi is more prevalent but the usage is still not indiscriminate. The locative in prevails in the following conditions: But this differentiation between nouns and adjectives is by no mcans an absolute rule, even in this text. The ending - mi does occur with nouns, though rarely, and it seems to convey greater emphasis and urgency: gakio kumaro kamrammi the prince was seized by the neck (137.20). There is abundant evidence throughout narrative Jain literature of the other closely linked tendencies, the survival of the ending in the immediate environment of the verb and in fixed locutions, thus antie 'in the vicinity of usually occurs before the verb and is very common, while antianmi is hardly, if ever, found. In fixed locutions such as place-names and times of day or of the year, the ending is used almost exclusively, e.g. paose 'in the evening, Mayanamalusave 'at the time of the great springfestival'. That this usage was basic to Middle Indo-Aryan prose is proved further by the Niya inscriptions, where - occurs mainly in samvatsare, mase, divase used in dating formulas (Burrow 1937: 24). (a) in nouns rather than in adjectives (b) in formulaic expressions and fixed locutions (c) in the immediate environment of the verb. This is evident as a general tendency, though not as an absolute rule in Maharlistri, Jain Maharastri and Jain Sauraseni, though there are numerous stylistic differences between the various texts. There is evidence of this tendency sometimes even in verse, e.g, in the Dharmopadesamala-vinarana where the distinction in the locative ending may serve as a means of differentiating a noun from an adjective: patte pottammi (p. 2, v. 3) which is rendered in the commentary by patre (noun, locative), prapte (adjective, locative). This distinction is most noticeable in the prose of the Kuvalayamala: ruddammi bhana-samudde flagga-laddhammi kaha i manyatte (adj.) (noun) (adj.) (noun) 'in this most dreadful ocean of existence where birth as a human being is obtained only by the rarest chance (p. 2.1.12). erisammi ja samaye 'and at such a time' (195.1) (adj.) (noun) diharammi samsare 'in this long circuit of mundane existence'. (adj.) (noun) E. CONCLUSION A glance at the paradigms listed in Prikrit grammars may give the impression that there are a variety of different endings which could be used indiscriminately in Middle Indo-Aryan. The texts, however, show that many of these are regional and chronological variants possibly indicating scribal traditions, and some of them are of stylistic significance. The cases of optional usage emphasise the fact that the Jain texts reflects a living and evolving language, and have subtlcties of expression that may often escape us. REFERENCES Texts quoted: Acardiga-Satra, ed. W. Schubring, Leipzig 1910 (Abhandlungen fur die Kunde des Mergenlandes, Vol. XII, No. 1). Ausgewahlte Erihlung in Maharastri, ed. Hermann Jacobi, Leipzig 1886. -182 -183

Page #102
VARIANT FORMS OF THE LOCATIVE IN MIDDLE-INDO-ARIAN 87 L.A. SCHWARZSCHILD Burrow, T. 1937 Th Language of the Koroshi Dauments from Chinese Turkson, Cambridge Arayaka-Erzahlunge, ed. E. Leumann, Leipzig 1897 Abhandlungen fur die Kinde der Morgenlandes Vol. X, pt. 2). Hinuber, O. von isri Anantakerti Digambar Jain Granthamala, 1968 Studes Kantex del Pali, besonders des Vinaya-Plaka, Munich. Bhaganat! Aradhand, in M Bombay 1993. Luders, H. Chappaluda of Kundakunda, ed. Pannil Son. Bombay 1920 (Manikand Digambar Jain Series, no. 17) 1954 Beobachtungen der die Sprache des buddhistischen Urkanons, aus dem Nachlass herausgegeben von Ernst Waldschmidt, Berlin Dharmapadesomala-miner of Jayasimhari, ed. L. B. Gandhi, Bombay 1949 (Singhi Fain Series, vol. 28). Lyons, J. 1969 Introduction to Theoretical Linguistics, Cambridge. Mehendale, M. A. 1948 Historical Grammar of Inscriptional Prakrits, Poona. Dhirtakkyldna of Haribhadrasari, ed. A. N. Upadhye, Bombay 1944 (Singhi Jain Series, vol. 19) Pischel, R. 1900 Grammatik der Prakritspracher, Strassburg. Roth, G. Karttienpresa by Svami-Kumara, ed. A. N. Upadhye, Agas 1960 (Srimad Rajandro-Faina Sastramala) 1966 "Blik sun Ivinaya and Bhiksu-Prakiraka and Notes on the Language', Journal of the Bihar Research Society, vol. LII. Kunalandla by Uddyotana-Suri, ed. A. N. Upadhye, Bombay 1969 (Singhi Jain Series, vol. 45) Sen, S. 1960 A Comparative Grammar of Middle Indo-Aryan, Poona. Staal, J. F. Liloa by Kouhala, ed. A. N. Upadhye, Bombay 1966 (Singhi Jain Series, vol. 31) 1967 Word Order in Sanskrit and Uniwsal Grammar, Dordrecht Tumer, R. L. 1927 "The Phonetic Weakness of Terminational Elements in Indo-Aryan', fournal of the Royal Asiatic Society. Nanopancami kahdo by Mahesvarasuri, ed. A. S. Gopani, Bombay 1940 (Singhi Jain Series, vol. 25) Paumscariyam by Vimalasuri, ed. H. Jacobi, Bhavnagar 1914 (Jain Dharma Prasaraka Sabha) Vasudevahindi by Sanghadasa, ed. Caturvijaya and Punyavijaya, Bhavnagar 1930 A nand Jain Grantlamala, vol. 80) OTHER REFERENCES Bernhard, F. 1964 "Gab es einen Lokativ auf mich im buddhi stischen Sanskrit?' (Nachrichten der Akademie der Wurzuchaften in Gottingen, 1 Philol.-Hist. Klasse No. 4) pp. 199-209 Bloch, J. 1950 Let Inscriptions d'Asoka, Paris. Bohtlingk, o. 1887 Panini's Grammatik, Leipzig. - 185 - -184

Page #103
330 LA SCHWACHILD DISTINCTION AND CONFUSION: A STUDY OF NEUTER PLURAL ENDINGS IN MIDDLE INDO-ARYAN By L. A. SCHWARZSCHILD I Introductory note From the time of Vararulon, most scholars, with the exception of some Gujarat Jain writers, have tended to view the Prakrit dialects in terms of Sanskrit and the divergencies from regular Sanskrit derivation have dominated discussions. T. Burrow's studies on the language of the Kharosti documents from Chinese Turkestan (1931) brought a new perspective to Middle IndoAryan, and showed how much Prakrit was of intrinse interest in its grammatical structure. One of the most striking characteristics of Prakrit is an internal and not a Sansit-dominated feature, namely the bewildering profusion of declensional endings and the apparently haphazard manner in which they can be used. Thus in the case of the nominative-scusative plural of the common neuter nouns in the grammars state that and are all used in Pukrit (Pischel 1900:25). It is well known that some of the endings are dialectal variants as will be discussed below, but this does not account for the entire situation. Such a profusion of alterant, and free variation of the kind postulated by Prakrit grammers would be unthinkable in any natural Innguage. It would seem obvious that syntactie and stylistic variation must be involved. An attempt has been made in ocent paper (Schwarzschild (1977)) to show this in connexion with the locative endings and a similar situation in discussed here in relation to the neuter plural endings all creatures, which recurs even in the Mahabharata (Michelson (1904), 103) and in trt paid modhen paddni adj. adj. noun * the three footsteps filled with mead (1.154.4) ugd no virpothuanian adj. adj. noun may all paths be good and easy to cross for ta' (VIL.63.6) minal priyd bhojanni primo adj. noun may you not take away from us our well-liked nourishments (1.104.8). A definite development can be seen within the Reveda: in the later hymns the proportion of uni endings is on the increase and even adjectives are in isolated instances affected by this change itfararsi adma adj. noun higher abodes' (X.57.10) halmsi priyatani barkisi adj. the offerings laid out on the straw'(X.115.11). But the ending-dini only very gradually became cominon in adjectives. This delay may be attributed to two son: (1) - originated from the nominal declension in the first place (from nouns in -a) (ii) within the unit of the noun-phrase, the noun is the dominating constituent. The noun, rather than the adjective therefore takes more readily the fuller and more distinctive ending in in the neuter plural The short form in was also favoured in Vedio when several nouns and particularly adjectives in the neuter plural were listed, and this was evidently for the same reason, the absence of need for distinctiveness. This trend was noticeable particularly when the longer ending and already occurred once within the same noun phrase It seems that there was still some survival of the Vedie situation in the earlier phases of Middle Indo-Aryan. The ending is found in the Asokan inscriptions everywhere except at Gimnar (soe Bloch (1950), 59) (J.R.E.II) widhini haldita e lopitil noun adj. adj. medicinal herbs have been caused to be imported and planted' There are also instances of this sage in early Pali texta (Weller (1915). 45) and in Ardhamgadh (Pischel (1900). $367), mainly in adjectives and in enumerations where one might expect the short forms sponding to the Vedic tradition : jhanaas con lakhand, catturi alam nyandar puchand of that meditation there are four outward signs,...four support, words, and questions (The 4.1). But this is rare, and can be regarded as an archaising stylistic variant. IIDerivation and distribution of the neuter plural endings There are two main problems 1. What principles govern the distribution of the neuter plural ending 2. What principles govern the distribution of the neuter plural ending ? tim 1. (w) The ending - The neuter plural ending in - oceurs consonally in the older Pili texta, and in Andhamaigadh and Jain Moharistul, though only rarely. It has been thought by Geiger (1916), 50 and by Pichel (1900). $367 to represents survival of the old Vedie ending of the neuter plural The situation is, however, wlightly more intricate than it may appear at first clano, is indeed the older etymological form of the neuter plural of nouns in just as and are the older neuter plural endings of nouns in and respectively. The ending-dini (just like wine and -in) is an innovation based on the influence of the group of noun that have a stem-final, such as Luronation plural Law . In the eveda (see Wackernagel (1930), 103) forms in-doutnumber those in-dini by 3 to 2. The distribution is not arbitrary, there is stylistic and syntactie variation: the older ending in His favoured in formal, fixed locations; it also survives most readily in adjectives and particularly in past participles while in ours mainly in nouns (Renou (1952). 995). The two forms are frequently in juxtapositions in the formula saayhl h`ngeriiyn adj noun - 186 -187

Page #104
A STUDY OF SEUTER PLURAL ENDINGS IN MIDDLE INDO-ARYAN The situation as discussed above can be summarized in the following manner: Vedic neuter plural -a (archaic style, adjectives and enumerations) early Pali, Amg., J.M., Asokan 4 (as before)" usage then discontinued, -d supplanted by the long form -dni. 1. (6) The late MIA neuter plural - There is a totally different situation where neuter plurals in - are found in later Pali texts, in southern Indian inscriptions (Mehendale (1948), 241) and particularly in Buddhist Hybrid Sanskrit. These forms are not associated with adjectives or enumerations, they occur readily in all kinds of nouns, sometimes with complete confusion of gender: sabbe te rupa all these appearances. This different usage does not appear to have arisen from the Vedio tradition described above, and an explanation must be sought elsewhere. Such an explanation was already indicated by Hemacandra (Pischel (1877-80), 1, 33) and was discussed further particularly by Edgerton (1953), 58: the use of -d for ani in the neuter plural shows the influence of the old nominative masculine ending ->ah. The main argument in favour of Edgerton's theory is the early evidence of confusion between the neuter and masculine endings in the plural. Although the distinction between masculines and neuters was always weak (Renou (1961), 278), the confusion seems to have begun in the East and arose from the similarity of the neuter plural nominative, vocative and accusative ending -dni and the masculine accusative plural ending -an. This confusion and the subsequent use of -ini in the masculine has been discussed in detail by Luders (1913), 988 ff., and more recently by Regamey (1954), 526, and by Bechert (1958), 310. The use of -a in the neuter plural, based on this confusion of genders, became more widespread in later Middle Indo-Aryan and was particularly common in the eastern Apabhrama of the Dohakosas (Tagare (1948), 138). This is in agreement with the growing lack of distinction between genders which was especially pronounced in the east where all distinction between genders has disappeared in the modern languages (Bloch (1963), map 5). Gender distinction has also disappeared in the Niya inscriptions, and -dni is only used in Sanskritizing formulae (Burrow (1937), 25). The situation can be summarized in the following table: Nom. Acc. Sanskrit Neuter pl. -ani Nom. Acc. Middle Indo-Aryan (Eastern) (Ardhamagadhi, Jain Maharagtet, the eastern dialect underlying Pali, and Apabhrama) Neuter pl. -, -, - Mase. -ah -188 -an 331 Masc. - (-dini, traces in Pali) e (Western influence) 332 L. A. SCHWARZSCHILD 2. The ending-d There has been some discussion of the origin of this ending. Gray (1935), 566 thought of it as coming from Vedic - with the addition of -ini borrowed from nouns with a stem-final in. This explanation is not altogether satisfactory, particularly in view of the existence of parallel forms in - and - for the - and -u stems. There is no other evidence of any such composite ending. -aim is much more likely to be a derivative by metathesis and weakening from -ins (and -im, -dim from ini, ni). This change was made possible by the well-known phonetic weakness of terminational elements in Middle IndoAryan (Turner (1927), 230). The isolated examples of a neuter plural ending -mim, e.g. dhanamim, quoted by Kramadievara in his grammar, are of interest in that they may well represent intermediate forms. The distribution of -dim and -dni is as follows: -dni only only - (and minor variants) -dim and -dni Ardhamagadhi Sauraseni Magadhi Jain Maharastri It is obvious that -ani is the older ending, and - with its minor variants is more recent. This again involves a stylistic difference: -ani belongs to a more elevated and formal style, and it is therefore not surprising that it is the only ending found in inscriptional Prakrits. It is equally obvious that -ani survived longer in the west and that -a like many other innovations began in the eastern dialects. But it is those dialects in which both -ni and -im are found which are most interesting in this respect: it is here in Ardhamagadhi, Magadhi, Sauraseni and Jain Maharastri that the stylistic and syntactic differences between the two endings can be seen most clearly. -Tai is both the older and also the more emphatic ending: it is used in situations where' phonetic weakening of terminational elements' is least likely to occur. Thus -dai is the prevailing form before enclitic and emphatic particles of any kind; some examples of this were already listed by Pischel (1900), $367): ruppa-payani va, suvanna-payani va encl. Pali Gandhari Dhammapada Inscriptional MIA Jain Sauraseni Maharastri Apabhrama encl. silver or golden vessels' (Oravaiyasuttans, 37) annani vi jai loe dukkhaim and those other misfortunes that are in this world' (Kunalayamala, 135.24). The ending ani is particularly common with anna Page #105
A STUDY OF NEUTRE PLURAL ENDINGS IN MIDDLE INDO ARYAN eydin panca mala-sippani demonstr. noun pronoun these five basic skills' (Avasyakacuri). Sometimes both an enclitic particle and the adjective-noun contrast are involved: bhavanani toranani ya, attalaya-viviha-citta-tungim noun noun encl. adj. dwellings and gates, high and bright with various upper stories' (Paumacariyam, 28.87). Although poetic works in particular show a certain amount of liberty in the use of the two different endings of the neuter plural, the basic theme of emphatic (-ai) versus less emphatic (-ai) prevails. This contrast is so strong that it should probably be reflected in translations: tulani ya mandalaggam adj. encl. noun and their scimitars were completely shattered. 333 In the Vasudevahindi the ending -dni prevails, but aim is used occasionally in conversational style and in quotations, and clearly the use of the two endings is governed by finer nuances of style. The distribution of the endings of the neuter plural thus shows us that there is still much to learn about Middle Indo-Aryan style and details of syntax: the most important texts from this point of view are the vivid prose-stories in Jain Maharastri. BIBLIOGRAPHY Bechert, H. 1968. Grammatisches aus dem Apadina-Buch', Zeitschrift der Deutschen Morgen. landischen Gesellschaft, ev, 308 ff. Bloch, J. 1950, Les inscriptions d'Asola. Paris. Bloch, J. 1963, Application de la cartographie a l'histoire de l'Indo-Aryen. Paris. Barrow, T. 1937. The language of the Kharoghi documents from Chinese Turkestan. Cambridge. Edgerton, F. 1953. Buddhist Hybrid Sanskrit grammar and dictionary, 1. Grammar. New Haven. Geiger, W. 1916. Pili, Literatur und Sprache. (Grundriss der Indo-arischen Philologie und Altertumskunde, 1, 7.) Strasburg. Gray, L. H. 1935, Observations on Middle Indian morphology, BSOS, vm, 563 ff. Luders, H. 1913. Epigraphische Beitrage. III. Das vierte Saulen-Edikt des Aloka, Sitzunge berichte der Freuarischen Akademie der Wissenschaften, 988-1028 Mehendale, M. A. 1948. Historical grammar of inscriptional Prakrits. Poons. Michelson, T. 1904. Linguistic Archaisms of the Ramkyana', Journal of the American Oriental Society, xxv, 89-145. Pischel, R. (ed.). 1877-80. Hemacandra's Grasematik der Prakritsprachen. Halle. 1900. Grammatik der Regamey, C. 1954. Randbemerkungen zur Sprache und Textuberlieferung des Karandavyuha Asiatica, Festschrift Friedrich Weller, 514-27. Leipzig. Renou, L. 1952. Grammaire de la langue vidique. Paris. Renou, L. 1961. Grammaire sansorile (2nd ed.). Paris. Schwarzschild, L. A. 1977. Variant forms of the locative in Middle Indo-Aryan, Mahavira and his teachings, 77-87. Bombay: Bhagavan Mahavira 2500th Nirvana Mahotsava Samiti. Tagare, G. V. 1948. Historical grammar of Apabhrasia. Poona Turner, R. I. 1927. The phonetic weakness of terminational elements in Indo-Aryan', Journal of the Royal Asiatic Society, 1927, 2, 227-39. Wackernagel, J. 1930. Altindische Grammatik, III. Gottingen. Weller, F. 1915. Zum Lalita Vistara. Uber die Prosa des Lalitaristara. Leipzig. -190 INDEXES

Page #106
Grammatical Index by Colin Mayrhofer absolutive: 7.22, 37-41, 119, 143; allied to the infinitive 22; distinct from the infinitive in Ardhamagadhi; used as infinitive 25; frequent in MIA 37; -Ona common in MIA, does not survive in NIA 39; -t, tanam) not found in Apabhrama 39; -dani said to be an ending in Magadhi 40; -pa-ti in Asokan inscriptions 40; labial forms-eppi, -eppinu and weakened forms thereof 40-41, uncertain remnants in NIA 41; forms derived from -ia andya, -i in Apabhrarsa, -ia in Sauraseni 40-41; see "conjugation" accents no heavy stress in MIA 137, 140; subject pronoun with imperative not heavily accented 147, 148; accent and terminational elements 177-178 adverbs: pronominal: of place 28-36, of time 57-68, 116; manner and place clearly distinguished 30; place and direction confused 35-36; manner and time confused 66; distribution of dialect forms later than Apabhrarsa 35; influence declension of pronouns 49-50 Apabhramsa traits innovations rather than survivals 60, 63; Apabhrama endings do not survive in NIA 65; intrusive consonant from relative pronoun 108; *itra 29, 36 etta, etto 34, 36; ja 66-67; tai(y)a 58-64: tahim 180 a 66-67; lahe 49; saifm) 59-60; endings: -am 131,-a 111, -Ith 3033, 36, -him 116, -hinto 35,-he 35, 36, -ho 35, of genitive plural 64-65, 131, of locative singular feminine 131; see "pronouns" affixes and suffixes common to nominal and verbal stems 52-56; dative suffixes formed by locative endings on genitive endings 97; comparative replaces superlative 128-9; -ana in verbal adjectives 122; -ara 129 -alla 138; -ala - ara 14; -ira 117, 123-126, illa or-ila replaces - ima 129; -isa, -ka svarthe 12, 64, 73, 74,95; kr in onomatopoeic verbs 73: -kka 70-75; -kya in nouns 74; je 25: -du 21: -tama superlative in numerals 128, -ti a quasi-suffix 114, does not appear in NIA 1165 -na 70, 71: -maand-ima 128-129; - 191 -

Page #107
- Ila 55, 70, 72, simultaneous extension of verbs and nouns/adjectives in Apabhrama 55 analogy: morphological 2, 7, 8, 9, 15 16, 18, 20-21, 24, 27, 29, 33-35 39-40, 43-45, 48-49, 50, 54-57, 6167, 71-73, 102, 108-109, 113, 116, 119-128, 130-133, 136, 142, 144, 149-150, 163, 166, 178; semantic 53, 59, 85-87. 114-116, 125, 143, 150, 165-166, 168 analytic: state of NIA and Romance languages 28 anaptyctic: see "svarabhakti under vowels" anusvira final tends to disappear 46, sandhi consonant after 106-7 aorist: 3 aspiration of consonants not due to 32; of j 38: in derivative of sakala 102 assimilation of consonants 37. 39-40, 80, 118, 123, 166, 167, 168, 171; of vowels 8, 10, 11, 19, 134; of aya to ae 60: of phonemic features in loan words 170 back formation: 80, 122, 132 causative: 7 compounds: see "prefixes". "postpositions" conjugation in Prakrit, with the connecting vowel- in Apabhrama, with the connecting vowel -- 7. 26, -- replacing -- in the absolutive 26: fourth Sanskrit class and Prakrit mm-118: of imperative tends to be defective 146; see "imperative consonants simplification of double 4, 101, 163: loss of intervocalic 15, 71, 140, of intervocalie - 24, 41, 101, of before a consonant 122: development of intervocalic 80: reduplication after a shortened vowel 78: reduplication of 134, 61, of c 71: enclitic tha keeps its dental initial 76; prefix causes reduplication of initial consonant of word 80, 81; sandhi consonants y and h 106-107; weakening of initial in an enclitic 107: geminated consonants in passive 120; evolution of the group -5-10- 10 -- in the future tense 2, 4, 6, 49, of the group tv 37-40, of the groups dl., d ry and off-to-I. 53-54, of the group sibilant plus - or. 100, sibilant plus nasal 176, -sm176-178, of initial y 108; labials become vowels 139, pronunciation of initial v 141, 145; changes from palatals to dentals 165-168, most examples are desiderative verbs 166-167; resemblance between palatal + and dentali 167; initial retroflex 169-174; nasal phonemes in old and middle IA 171174, as phoneme or as allophone of n 172-174, alveolar articulation of n 174: see aspiration", "assimilation" "dissimilation","loanwords". "palatalization","samprasarana "sporadic change", "vowels contamination see semantic" under analogy", "influence contraction: 11, 15, 16, 19, 87-88; involving a remnant of the reduced vowel 18, 19 correlation: see "subordination declension: feminine: 42-51, 95, origin of oblique singular-e 43-46, confusion of oblique singular forms due to phonetic causes 46, origin of Apabhramsa singular ending-he. hi-him 47-48, 50, vocative in-c 109: evolution of nouns in - and -9596; nominal stems not declined 100102: variant forms of locative 176-184, derived from -smin 176-183, derived from e 180-183, the ending.mmi more emphatic 181, 183, the ending -e commoner in nouns than in adjectives, common in formulas and in the environment of the verb 182183: neuter plural (nominative/accusative) endings 186-190, the ending -186-188, confusion between plural masculine and neuter endings 188, the endingam 189-190, the ending-ini more elevated and emphatic 189, 190, cases: vocative 148; -tar becomes locative 32; dative rare in MIA 45: other cases substituted for dative 46 genitive plural endings in adverbs 64-65: adverbial uses accusative 159-160, 162-163, instrumental 111, 162; ablative 159; see "affixes and suffixes denominative: 120 desiderative: see "phonemic change" dialects in MIA: features peculiar to a dialect or distinguishing one dialect from another regions (mainly Asokan inscriptions) south 100, 11, 37, 39; north 11, 44, 46; cast 2, 10, 19, 20, 32, 37, 39, 40, 45, 46, 48, 93, 94, 107, 108, 160. 167, 168, 171, 172, 181, 188, 189: west 4, 5, 9, 10, 11, 18, 19, 20, 37, 40, 93, 94, 167, 172, 173, 176, 177, 179, 181, 189; north-west 4, 23, 40, 44, 45, 46, 96, 172, 177; centre 11, 45, 48, 93, 172, 179, 180, 181; west-centre 11; east-centre 8, 9, 11, 41; post-Asokan inscriptions 189; Niya 172: Pali 167, 171; Ardhamaigadhi and the Jaina canon 2, 3, 23, 24, 25, 39, 43, 108, 150, 151, 152, 160, 167, 177: Dhakki 39; Jaina Prakrit 3, 5, 9; Magadhi and Magadhi of dramas 25, 39, 40, 108, 150, 160, 167, 177; Maharastri and Jain Maharastri 3.4. 9, 25, 39, 43, 45, 46, 58-60, 62, 80, 104, 125, 126, 151, 152, 160, 177, the two contrasted 167; Sauraseni and Sauraseni of dramas 32, 33, 126, 150, 151, 177; Sauraseni and Digambara Sauraseni 3, 4, 9, 10, 11, 25, 33, 39, 48, 50, 138, 151, 157, 167, 177, 179; southem or Digambara Apabhrama contrasted with westemor Svetambara or Gurjara Apabhrama 4, 5, 7, 9, 12, 13, 17, 18, 20, 25, 26, 35, 50, 65, 93, 101, 152: eastern Apabhiramsa 26, 39, 101: -192 - 193

Page #108
earlier and later Apabhramsa 54, 123, 139, 145; literary Apabhramsas did not derive from particular literary Prakrits 50, in morphology Apabhramsa differed totally from the literary traditions of the central dialects 180; sets of interrelations needed to describe dialectical relationships in MIA 166; distribution of features in NIA, see "future tense", "infinitive", "adverbs" differentiation: of vowels 134-135, 1378, 140; of ae to ai 60 dissimilation: of vowels 134-137, 140, rare in MIA 135, 137, 140, common in Romance 135, 137; of consonants 166-168, causing loss of final nasal 179-180 elision: of first syllable of kera and its derivatives 98; of syllables in particles 109 emphasis: see "declension", "particles", "prefixes", "reinforcement" enclitics: 15, 18, 25, 75-76, 87, 105, 107, 108, 109, 111, 173, 190 epenthesis: of vowels 134 foreign words: see "loan words" frequency: influences phonetic evolution 2, 108: resistance to morphological analogy future tense: 1-11; -s-type and -h-type have the same origin 1; distribution of forms belongs to the period of the modem vernaculars 8; connecting vowel 6-8; ending vowel 8-11; types of future: -s 1, 4, 5, 20, 152; -194 -h- 1, 3, 4, 5, 8; -ss-3, 8, 11; -hi- 8, 9, 10; -ha-9, 10; sigmatic 11; velar of roots ending in c71; expresses will and or futurity 3; association with imperative 146; in subordinated clauses 154; see "participle" gender: distinction tends to disappear in eastern dialects 188 gerund: see "absolutive" hasruti: 26 homonyms: and homophones, avoidance of 31, 32, 39, 40, 41, 61, 74, 76, 84, 96, 100, 167-168, 178 imperative: 146-152; with pronoun subject 146-149, in Apabhramsa and Medieval French 149; second-person singular forms 149. 152: bare stem of thematic verbs 149, becomes -u in Apabhramsa 151, 152, mixture of imperative and optative forms 150, -ahi, -chi, -su 149-152, -hi 149-151,-ahi especially after n 150, -i, -e, -hu in Apabhramsa 152 infinitive: 22-27, 144; allied to noun of action, participle, absolutive 22; derivation of the Prakrit infinitive 23-24, 26-27: distribution in NIA does not reflect MIA dialects 27; in Ardhamagadhi 24-25, 104, 152; used as absolutive 25, 106; possible vestige of Apabhramsa absolutive 41; association with the particle je 104110; see "nouns of action" influence: of one dialect or language on another: morphological 4-5, 7, 910, 19, 24-25, 39-40, 48, 50, 107, 126; semantic 31-32, 37, 38, 179-180; see "analogy". "loan words" injunctive: 3 labialization: of consonants 40, 55, 93; affecting vowel 144, 145 lengthening of vowel: compensatory after simplification of consonant group 6, 7, 80, 100, 101, 122, 163; of samprasarana vowel 39; provokes disappearance of anusvara 46, before imperative ending-hi common, before-su rare 150; associated with loss of final -n 178 literary: see "popular speech" loan words: Sanskrit words from Prakrit 55; Prakrit words from Sanskrit 74; exceptional phonetic developments in loan words from Persian 124, non-IA 122, 127, 136, 142, 143, 169, unknown 144; non-Sanskritic features: initial retroflex plosives 169, intramorphemic nasal-plosive clusters 169; see "sporadic change" metathesis: rarely a factor in MIA 59, 67; of vowels 134; of vowel and nasal consonant 189 nouns of action: accusative infinitive 22; dative distinguished from infinitive 23: -h-in the oblique cases in Apabhramsa 27 -195 number of the noun and pronoun: plural of respect 20; second person singular ablative confused with plural nominative 21; numerals: 128-133 onomatopoeia: 38, 73, 75, 112, 113, 114, 135, 170 optative: 146, 150, 154, 155 palatalization: 37-38 parataxis: see "subordination" participle: used as future tense 1, 3, 10; allied to the infinitive 22; past: 122, 142, 144, in -kka 71-73, new formations in -i(t)a 72, origin of present tense 123; obligatory participle 27, 144, used as infinitive 27; -ira as a kind of participial suffix 126 particles: negative 82-88; emphatic u 159, khalu 159, je 104110, 160-161, mostly final 105, nam with imperative 147, -tti not used for emphasis 112, nu 159, se developed from neuter singular pronoun 160; temporal ta 156; interrogative 159-164, replaced by pronominal expressions 159-160, 162-163, kaim 164, kissa, kisa 163, kiha 164, nam 161-162; see "vowels" passive: -mm- morpheme in Prakrit 119, 120 person of verb: frequency of third singular 2, 5; first more conservative than third 2, 6; first in future tense can express both will and futurity 3; third plural forms adopted from a neighbouring dialect 9; see "frequency", "imperative"

Page #109
res(a)immi 97, compounded or used with instrumental-locative ending 97 phonemic changes: 164-168: see "assimilation","consonants", "contamination", "dissimilation" popular etymology: 136; see "contamination" popular speech: contrasted with written or literary language 4, 5, 6, 9, 21, 29, 35, 36, 37, 39, 40, 50, 62, 81, 86, 88, 108, 109, 126, 132, 146, 147, 151, 154-158, 162, 164, 171: invoked, as in Romance philology, to explain features absent from documents 42: documents indirectly reveal condition of spoken languages 51, 129; responsible for new words in MIA 69, 113: see "postpositions" possessive adjectives: 12-21; continuous evolution from Apabhramsa forms 12, 13, 14; modern forms in -r-all derived from -kera 20 postpositions: preferred to endings in popular speech 65, connection with the decline of case system 91.98; properly fom compounds with preceding noun 90, 96, 98; loss of initial vowel 93: kara 13, 18: kera: 13, 14, 15, 18, 20, 90, 92, used with genitive of pronouns and stem of nouns 14, enclitic 15, 18, origin of resi, resammi 89-90,95 sporadic change: in consonants: from palatal to dental 168, from retroflex to dental 170; in vowels: 134, from u to a 136-7: in loan words 136, 170 subjunctive: 146, 150 subordination: 153-158; contrasted with parataxis 153-4, 158; contrasted with correlative structures using demonstrative and relative 154-156, less common in popular language 155, with modal verbs 154: development of subordination with jam consecutive, completive, causal 154-156, 158, with jena causal and consecutive 156-158, with jaha of indirect speech and conditional 157 Prakrit grammarians: 3, 5, 7, 12, 14-15, 17, 20-21, 33, 37-40, 49, 53, 58, 59, 66, 71-72, 74, 86-87, 90, 97, 100, 104, 106, 109, 115, 118, 120, 123, 129, 131-132, 137, 141, 165 167, 173, 186, 188-189 prefixes: u- from Sanskrit ape, ava- 77 81, ud-79, 143; phonetic changes determined by recognition of compounds 80, 81, these changes simpler than in Romance languages 80; new creation 0-80, 81; vo-141-145, more emphatic than Vi- 143; see "consonants" present indicative: 8, 71, 119, 122, 146, 154 pronouns: system in IA 28, 57, 64, 66; keep case distinctions longer than nouns, cp. Romance 91: personal: first person singular replaced by asmi 21; second person singular ablative confused with plural nominative 21; phonetic evolution of second person analogous to absolutive 39; second person singular nominative ending-hu 101: demonstrative: in NIA a recreation by analogy not a survival of early OIA 42-43; influence the declension of nominal Stems 48, 97, 101-2; origin of feminine oblique singular tise 49: confusion in MIA 61; evolution of ayam 61, 68, exceptional locative singular 178, of etad 61, of esa 101-2; new stems - 62-63, 48-50, relative: initial y influences emphatic particle 108; position in imperative clause 147-8; case-forms of interrogative as interrogative particles: singular ablative 159, neuter (nominative/accusative) 159, 162, Instrumental 162, genitive 163: see "adverbs","imperative". "particles reduction: see "contraction" reinforcement: for emphasis and distinctiveness, of negative particle 84-85, by repetition 85-86, 116; of pronouns 91; of particles 106, 162, cp. Romance; of exclamations and forms of address 109: by anaya and resim), 98; of imperative by subject pronouns 147: see "affixes and suffixes", declension", "postpositions" "prefixes" remodelling: 26, 64, 66, 113, 116, 128 129, 137, 138, 149 rhyme 171 Romance: see "analytic" "completeness", "dissimilation" "imperative","popular speech", "prefixes". "pronouns". "reinforcement", "semantic change" samprasarana: 37, 39, 44, 60, 67 scribal tradition: 172, 173, 176, 183 secondary endings: in future tense 3 semantic change: parallels in French 114 shortening of final vowel after an accented syllable 44, 46; of final a in adverbs spoken language: see "popular speech" disappearance of subordinating conjunctions 158 suffixes: see "affixes and suffixes svarabhakth see "vowels" synonyms: 111, 116 systems: disorder followed by the formation of new systems 28, 81. 91, 98, 116, 145, 164: see "remodelling terminational elements: weak 1, 2, 4, 6, 18, 24, 40, 41, 49, 85, 177, 189: see "accent" variation: the concept of free variation 175, 186, the numerous morphological variants that are a feature of MIA may be conditioned 175-183, 186 vowels: svarabhakti 124, 143, 145; in hiatus 139, 140; change of final -am to -a in particles 161: a and u 55, 134, 144, 151; i and e7, 30, 33-35, 49, 63, 106; 98; tanau 19, 89-94, 98, postpositional adjective with genitive of noun or pronoun 90. true postposition in compounds 90, causal 90: nau 19: resi dative indicating purpose 97, addition of locative ending gives - 196 - 197

Page #110
Indexes of Old, Middle and New Indo-Aryan words! u and o 53, 79, before double consonants 78; after the semivowels y and v especially in initial position 141-2; effect of a following retroflex 171; see "assimilation", "differentiation", "dissimilation", "epenthesis, "lengthening", "metathesis", "postpositions", "shortening". "sporadic change" word-play: 170 yasrutr. 17,87 by Royce Wiles Old and Middle Indo-Aryan a 87 - 68 aammi 61, 178 abhavati 83 abhijfia 38 acarya 16 acera 16 adhas 32 adhastat 32 adhestat 32 adhestha 32 adhi 47 adima 128 agalu 136 agaru 136 aggimmi 181 aguru 136, 136n11 aguna 139 aha 158 aha 108 aham 101n ahariham 108 ahavai 83 ahe 32 ahi 47 ahi 82 "ahi 85 ahijja 38 ahila 100 aim 108 ainna 173 ainne 173 ajira 124 ajju 95 ajra 124 -aka 94 -akara 17 akara 17 akara 18 akhila 99, 100 akima 173 -ala 14, 138 alla 53 amailla 129 ambara 138 ambe 109 ambhara 20-21 amha 16, 18, 20 amhaara 18 amhaera(a) 16 amhaha 16 amhakera 14 amhakera(ka) 16 amhanam 16 amhanam 16, 18 amhara 12, 16, 18, 21 amhara 16 amharaya 13 amhariya 12 amhe 21 amhi 21 *ammi 178 *ammi 178 anantaram 112 anayam 27 andolina 125 anhaya 24 anna 189 annam 173 annatto 34 anna(y) 58 annesim 97 antiammi 183 antie 181, 183 anupariyafteum 24 anya 189 anyada 58 anyam 173 anyatra 172 ao 34 apa- 77-81 apabhavana 77 apakrs. 79 apakrsta 78 apannam 173 apasaratha 77 apasarati 81 apatya 167 api 59, 84, 116, 173 appa 40 appana 93 apta 32 -ara 14 (a)rall 18 ardha 132 ardra 53, 54 artta 32 arya 95 as- 82, 85 -asa 136 1. The main purpose of these indexes is to provide access to the discussions and citations of Old. Middle and New Indo-Aryan "words" in the preceding articles. The entries are for the most part listed as they are found in the articles. The indexes possibly er on the side of being too inclusive but I have preferred to shick the responsibility of deciding which words potential users would want information about and so I have excluded only the most extraneous words - 199 - - 198 -

Page #111
*asati 82 asit 83 asmat-santaka 94 *asme 21 asmi 21 asmin 178 asrava 24 assim 178 asti 82, 84 -at 66 atas 29 atman 32, 39, 39, 40, 93-94, 95 atmanaka 90, 93, 94 atra 29, 32, 33 atta 32, 93 atta 32 attabhava 32 attanaya 94 attha 33 auna 139n16 alina 139 alinattisa 139 ava 108 ava- 42, 43, 77-81 avagirate 78 avagraha 78 avakaham 108 avakirati 78 avakranta 81n avakrs- 79, 80 avakrs- 78 avalagna 78 avalambita 77,81 avalokana 77 avannam 173 avarugna 78 avasana 131 avasarpini 79 avaskand- 78 avastabh- 79 avastambh- 78 avatamsa 78 avatar- 78 avatarami 80 avatarati 81 avatarita 77 avatima 77 avatra- 79 avaviddha 81 avaviddhani 77 avayar- 78 avayariyavvam 78 avihimsra 124 avisii 5 avocat 144 avos 42 aya 60, 61, 62 -aya 92, 95 *ayads 63 ayae 23 ayam 61 ayampira 125 ayara 17 #ayatya 33, 61 bahavah 94n bahi(m) 60 bahistat 32 bahula 13 bahummi 181 bahuve 94n baisai 80, 81 baissa 40 bara 40 basiau 10 *basihahu 10 baju 144 be 40 besa 40 bhaahi 151 bhaddhidum je 105 Bhadhaya 46 bhadra 53 bhalla 53, 53n bhalo 530 -200 bhamai 119 bhamaya 137 bhamira 125 bhammai 119 bhamsai 55, 72 bhamuha 137 bhan- 172 bhanahi 150 bhanasu 150 bhanihamo 2, 6 bhanijjasu 150 bhanira 125 bhanjiham (?) 6 bhanu 148, 151 bhareum je 104 bhasira 125 bhavai 85 bhavati 83 bhavihamo 6 bhavihinti 6 bhavissihidi 11 bhavisyati 4 bhayaye 45 bhid- 54 bhidra 124 *bhillei 54 bhillia 54 bhinna 173 bho 109 bhoha 137 bhottae 23, 24 bhottum 25 bhram- 118 bhramati 54, 118 bhrams- 55 bhramyati 118 bhras- 72 bhru 137 bhrumukha 137 bhu- 83, 85 bhukka 72 bhulla 55 bhullai 55 bhumaya 137 bhumsai 55 bhunje 148 bihasu 151 bindhidum je 105 bodraha 142 bollai 54 bollisu 6 brahmana 31 buddhva 38 budhyate 38 bujjha 38 bujjhahi 5 bujjhai 38 ca 59, 87, 107, 110n13 caccara 38 cacchai 165, 166 cadesahim 5 cal- 54 callai 54 camatkara 73 cankamai 119 cankammai 119 cankramate 119 cankramyate 119 carama 64, 128 carejjasi 150 carima 64, 128 cattari 38 catuska 73 catvara 38 catvari 38 caukka 73 calisara 131 ccia 107 cciya 112 cest- 167 cettha- 167 cettaha 167 cethai 168 ceva 107 chad- 53, 54 chailla 139 chakka 73 challi 53, 54 challih 53n chei(1)la 139 chekila 139 chid- 54, 55, 142 chidira 124 chidra 53, 124 chilla 53, 54 *chillei 54 chillia 54 chinnah 173 chinne 173 chubuka 136 chuccha 156, 166 chudda 166 chulla 166 chutta 166 ci- 120, 143 cibua 136 cibuka 136 cid 86 clicchaa 167 ciketa 166 cikits- 165, 166 cikka 72 cimmai 119, 120 cinai 120 cintchi 151 cintesu 151 cirantana 89,91 cit 59 cit. 166 citteur 24 citha 147 citha- 167 cithai 167 cinhasu 147 cihihisi 5 ciyate 119, 120 corayati 54 cubuka 136 -201 cuccha 165, 166 culla 53, 166 cumbati 123 cumbihami 6 cumcha 166 cummai 121, 123 cunna 173 carna 173 da 70 dabbha 171 daddha 171 dah- 171 dahamo 2 dajjhihasi 9 dakka 70, 72 *dakna 70 dallai 54 dama 170 dama 170 damai 119 dambara 169 damina 119 dammai 118, 119 dams- 70, 72 damyati 118 danda 38, 171 danda 171 danim 58 dankhai 72 danksyati 72 dara 171 dara 171 dara 40 darbha 171 das- 171 dasai 2n dasami 2 dasamo 2,49 *dasna 70 dasta 171 datta 70 daha 171

Page #112
evahim 58 eval-khutto 63, 68 evalya 63, 68 evam 29. 310, 62, 63 cyla 61 olyam/ely)ar 61, 156 eyammi 181 daun 25 -ddara 168 de 109, 140 de-dehi 151 deha 95 dehi 149 desai 7 desu 7, 151 deula 140n17 deur in 140 deva 109, 140n17 deve 43,45 devaih 44 devakula 140n17 devara 139 devarani 140 devaya 43, 45 devebhih 44 devehim 44 dhammai 119, 120 dhamsai 113, 114 dhamyate 119 dhanamim 189 dharchi 146 dhasa 113 dhasai 112 dhasakka 113 dhasakkei 113 dhasatte 113 dhasatti/dhasa-ti 112, 113, 114, 115, 116 dhalika 171 dhehi 149 dhe ka 169 dhrum 5on dhumdhallai 54 di- 170 di- 170 digicha 165, 166 diharammi 182 dimba 169 dindima 170 dindimatam 170 dindisa 170 dinna 70 disai 2n disihasi 9 ditara 169 divasa 140 divase 183 divira/divira 124 do 132, 165 do-sara 130, 131, 132 docca 129 dogamchi 165 dogga 165 doha 132, 133 doha 132 dohara 132 dola 170 dola 170 dolira 125 doraka 170 goraka 170 dosa 132 dosa 165, 166 dosa 165 dos-ratna 166 dosakara 165 dosani janta 166 dosara(ka) 132 dosramana 165-166 dosirayana 166 dosayara 165 dosina 165 dosini 165 dota 170 dr 60 drum Son du 17,21 du(gamcha 138 du- 122 dualla 138 duga 165 dugamcha 165 dugulla 138 dugumocham 165 dugumch/dugurcha 138, 166 dukula 138 dumai 122 dimana 122 damia 122 dummal 121, 122, 123 dummana 122 dummia 122-123 dummijjai 122 dina 123 dundubha 169 dundubhi 170 dunnia 122, 123 durmanas 122 dubla 138 dvara 40 dvau 165 dve 40 dvesya 40 dvi- 132 dvidha 132 dvidha 133 dvika 165 dvitlya 129, 131, 132 *dvitya 129 egasara 130 egavisa 130 egaly)a 58 eguna 139n16 egina 139 chai 9 chi 149, 150 chi 61 chimi 2 chu 101-2 cina 61 cka 74, 128, 129, 131 cka-saraka 130 eka-sarga 130 ekada 58, 131 ekadasa 130 ckakin 130 ckanauim 130 ckasar(a) 132 ckasara 130 ckasas 131 ekka 74 ekkasaa 130 ekkasara(k)a 130-132 ekkasara(ka) 130, 131 ckkasari 131 ckkasariam 131-132 ekkasariyae 131 ekkasariyai 131 ckkasariyanam 131 ckkasi(m) 131 ekkasia 131 ekona 139 eksar(a) 130 emahalaya 62, 63, 68 emahiddhiya 62 emahaliya 62 emvahim 66 ena 62 eshim 58, 116 -era 15, 17, 18 erisa 61, 62, 63, 68, 136 erisammi 182 es 61 esa 101 esai 7 csam 61 esamt 2 esasi 6 esim 62 eso 61 essamti 2 esu 61, 62 esuhuma 62 cla- 34n etad 34, 36, 61, 62, 156 etades(a) 29 etarhi 29, 34, 35, 36, 49 elasmat 33 ctavat 29 etesam 61, 179 ctra 36 etrisa 63 etta 33, 36 etta- 34 ette 33 eltahe 33, 34, 35, 36 cuahe 35, 49 ettahi 33, 34 etishi 49 ettamha 33, 34 eutha 31, 33, 34, 35, 36 ettha- 34 ettha 31 ettia 33 ettie 58 ettika 33 ettila 33 ettiya 33, 34, 61 etto 34, 35, 61 ettula 33 euna 139 eva 6, 106-108, 110 eva(nu) 62 evadda 62 gabhira 124 gaccha 146 gacchaha 146 gacchasu 151 gadua 39 gahira 124 GIAIT - 120 gamane 23 gamara 17 gambhira 16 gambhirya 16 gamere du 17 gami(s)si 11 gamiai 119 gamihl 6 gamijai 119 gamira 125 gammai 119-120 gamyate 119 gantavva 120 ganchilla 129 ganthima 129 gantuna 120 garima 136 gariman 136 garisha 135-136 garitva 136 gariyas 135-136 garu 135n8 garuka) 135-136 garua 135, 136 genhahi 150 genhasu 149, 150 genhia 39 ghara 14 lldhurr m dhuni 38 dhvaja 38 dhvarsati 113 dhvani 38 dhvasta 113 dhvaksa 171 e 62, 63 cam 156 editada 63 odrla 68 cena 61, 62 besim 179 eesu 62 egada 58 - 202 - 203

Page #113
jeva 106-8, 110 jevada 62 jhad- 116 jhada 114 jhadai 114 jhadajhadajhadajhadanta 114 gharshim 178 gharasmin 178 ghattiham 6 ghaukama 25 ghettona 39 ghol- 123, 126n3 gholai 118 gholira 117-127 ghonai 126n2 ghulai 126n2 ghumm- 117-123 ghummai 118, 121, 122, 123, 126n2 ghummia 122-123 ghummira 117-127 ghummiya 118, 123 ghun-1263 ghunn- 118 ghunn- 118 ghunnanta 118 ghunnia 122 ghom. 117, 122 ghimayamana 117, 126 ghirita 122 ginhahi 146 ginheum-je 25 ginhium-je 25 gonikka 74 *gonikya 74 gramakara 17 gunellai 54 guha 124 guhira 124 gulugulai 135 guna 172 gufullai 54 guru 135, 135, 136 gul jollai 54 hallira 125 ham 109-110 hamira 19 hambho 109 hamera 15 hamho 109 hamje 109 *hamma 120 hammai 119-120 hammara 19 hammaui 12704 han 120, 121 hanai 120 handa 109 handaha 109 handi 109 hallje 109-110 hanman 119, 120 "hannai 121 hanta 109 hantavva 120 hantona 120 hanyate 119 harahi 27 harahi 27 harahu 27 harahu 27 harai 26, 27 harana 122 haranti 27 harau 13, 14 harad 27 haraum 26 hareha 26 harei 26 hari 143 haremi 26 haremo 26 harenti 26 haresi 26 hareum 26 haria 122 harihai 9 harisa 136 hariu 25 harium 26 have jja 154 havihi 5 havihunti 6 hegha 32 heshilla 129 heshima 129 heum 162-163 heva 107 hi 65.87 hims- 92 -hin 66 hinsaka 92 himsaya 92 himsira 124 hindira 125 ho 109 hohadi 9 hohammi 5 hohanti 9 hohl 9 hohimmi 5 hol 83 hosai 4 hp-143 hu 100, 102n hui 83 huvai 83 hohi 148 io 34 ira 117, 123-126 irisa 63 is. 124 -isa 136 isira 124 itas 29, 34, 36, 50 iti 109, 111, 112, 157 itra 30 *itra 29, 33, 34n, 36 itta/itta 33 ittae 33 ittha 31, 33, 34, 35 ittha 29, 30, 31, 33 ittham 29, 30, 31 itto 34, 35 iyam 61 iyanim 58 iyat 29, 63 iyat 34 "iyatika 63 jalana 122 jalia 122 jam 154-158 jamma 120 jammai 119, 120 jammi 178 jampira 125 jampiu 25 jan. 120 jan(t)u 150 janahi 150, 151 junasu 150 janera 17 janeri 17 jani 87 janihasi 9 janihinti 9 janihisi 6 janisii 5 janman 119, 120 jannam 173 jangam 161 Jan 87 jasate 38 jasayati 38 jatha 155 jau 33 jattha 33, 34 jatihami 2 jattiya 34 jatto 34 ja(t)u 150 java 66, 67 jaya 58 jalya 58 jayate 119, 120 jayaya 46 je 104-110, 160-161 jena 154-156, 158 jettahe 35 jetthali 35 jetthu 35 jettiya 34 jhadappai 114 jhadatti 112-113, 114, 115, 116 jhadavi 116 jhaditti 112, 114, 116 jhagiti 112 jhaja 38 jhalajhalai 135 jhalakkia 73 jhallika 136 jhasa 38 jhasa 38 jhasia 38 jhassa 38 "hap 112, 114 jhatiti 112-113, 114 jhana 116 jhatti 12, 114, 116 jhattihim 116 jhaya 38 jhillika 136 jhulukkiya 73 jhunai 38 jhuni 38 jhorai 38 jhDrasu 147 jhvaja 38 ji 106, 107, 160-161 jia 24 jighats- 165, 166 jimai 127n13 jimmai 121, 127n13 jinasu 151 jittha 34 jitto 34, 35 jiva 24 ja 62 ja 66-67 ja 59 jada 58 jaevi 148 jaha 157-158 jaham 31 jabe 35 jahe 49 jai 150, 151, 152 jahim 49 jai 65, 131 jal 58, 59, 64 jala 58 jaim 108 jalu 64 jaiya 65 jalya 65 jal(y)a 58, 59, 63, 64 jaiya(m) 131 jaiyahum 58 iammi 61, 178 id 160 idam 61 idanim 29, 58 id(a) 29 Idesa) 29, 63 iha 29, 31n -illa 129 ima- 61, 62 -ima 128, 129 ingala 53 ha 107, 109 hakkai 730 hakkarai 73 hallai 54 - 204 - 205 -

Page #114
jivihamo 6 jivium je 105 ijeva 106 jati 172 jnayate 86n jugga 165 juguccha 167 jugumcha 166 jugups. 38, 165, 166 jugupsa 138 jumma 165 junha 165 juryati 38 jvara 168n19 jyotsna 165, 166 kalviha 64 Kaiya 67n karya 64,67n kalya-laggi 65 kai ya 58, 59, 61, 64 kaiyaham 65 karyahum 58.68 kajja 16 kale 65 kalyana 172 kambala 138 kamha 159 kammi 178 kanha 19 kanika 74 kanikka 74 kanna 173 kannera 17 kanniarl 17 kanthammi 183 kanya 171 kao 34 kara 13, 17, 18, 19 kard 16 *kara 15, 16 kara 15, 16-17 "kara 14.16 -kara 13 *kira 13, 14 karaka 18 karana 122 karai 18 kare 148 karci 80 karesahim 5 kareum je 104, 105 karhi 29,35 kari 148 karia 122 karihai 5 karibami 6 karihi 9 karihimi 3 -ka 64 kada 29, 58, 59 kadacit 85 kadapi 116 kadua 39 kahadi 9 kaham 3 kaham 31 kahe 35 kahci 154 kahejja 154 kahim 156 kahimi 3 kahin 88 kahiu 25 kahium-pi 25 kar 59, 64 kaza 64 kara 58 kali 64 kaim 164 kaima 64 "katra 64 *kairisa 61 karttha 64 kaivalya 64 Kalvaya 64 karihinti 9 karihisi 6 karisu 7 karma 171 kama 171, 173 *karnasmin 19 karya 13, 16, 92, 97 kasmat 159 kassa 162, 163 kassaheum 162-163 katama 64 katara 64 katha 101n katham 31. 164 katham 29.31 kati 29,64 katicit 64 katipaya 64 katividha 64 kattha 34, 35 katto 34, 35 kattohimto 35 *kauhala 138 klum je 104, 106 kauthala 138 kaya 58, 60, 64 kaya 95 ka(y)a 58 *kayada 59 *kayalds 63 kuyai 59 kayaim 85 kayattia 33 kayavi 116 "kayida 60 kangani 96 kangu 96 "ke-vaddhaya 62 kecciram 63 kehim 90 kemahalaya 62 kemahiddhiya 62 kena 159, 161, 162 kena karanena 159 kera 13, 14, 17, 18, 19, 90, 92, 97, 98 -kera 13-14 kera(ka) 16, 18 keraka 18 kerau 98 keraya 15 keraya 15 keresim 97-98 kerisa 61, 62, 63 kesam 61 kesim 97 kettaho 35 ketthu 35 kettiya 33-34 keura 137 kevacciram 63 kevalya 63, 64 *ke ya 61 keyura 137 khadakki 74 khajjihisi 9 khalu 84, 100, 101, 102n, 109, 159 khamahi 151 khamai 119 khamcijasu 150 khammai 119-120, 127n12 khan- 120 khanahi 150 khanai 120 khanavettae 24 khanaveum 24 khantavva 120 khantina 120 khanyate 119 kharaya 143 khayate 119 khedajtlah 173 khellai 54 khelnu 54n kheyanna 173 khida 31 khu 109 khummai 121 khummi(y)a 123 khunna 123 -khutto 63 ki 48 kicca 37, 38 kidha 101n kidikidiabhte 135 kids(a) 29, 63, 163 kie 163 Iciha 101n, 164 kilammai 118, 119, 127n13 kilanja 136 kilihimi 5,9 kilikilai 135 kilifija 136 kim 159 kim 160-164 kim khu 162 kim nu 162 kim (ng) khu 162 kim punah 162 kim una 162 kimna 162 kinna 161-162 kinnam 161-162 kirikiria 135 kirisa 63 *-kis 131 kisa 156, 163 kise 163 kissa 163 kitikitika- 135 kittha 34 kiyat 29, 34 -kka 129 klam- 119 klamyati 118 koduhala 138 kohala 138 kokka 72 kokkhai 72 kouhala 138 kouhalla 138 kr. 73, 92 krid- 54 krida 31 kroksyati 72 kr. 79 krtva 37 krus- 72 ksam- 118 ksamate 119 ksevelati 31 ksip- 53 ksud- 123 ksudra 53 ksunna 123 ksvel 54 *ksvelyati 54 kuha 29 kukka 72 kunijjasu 150 kuo 34 kurkura 33 kutas 29, 34 kutra 29, 30, 33 kulta 33 kutthala 138 labhasu 146, 151 laddbu 25 -laggi 65 lahesu 7 lahihisi 9 lahinti 8 Ini 54 lambai 123 lambate 123 lammai 123 langheum je 105 lavira 125 - 206 - - 207 -

Page #115
Leochai 33 ledhu 75 ledhukka 74, 75 lesai 7 lestu 74 levi 148 Thikka 72 Licchavi 33 lihiu 25 loe 181 logamsi 181 luhahu 148 lukka 71 lul- 54 lullai 54 "lummai 123 luncai 71 ma 19 mahu 19,91 mahu-Lanai 90, 91 mahu-tanaya 91 maila 129 mailla 129 maim 86 maim 101n majjhe 181 makim 1010 makura 137 malaa 43 malae 47 malae 43,45 malahe 47 malai 43, 44 malay 60 malo 43 malaya 43 malaya 60 malayah 43 malayai 43 malayam 43 malayah 60 mama 19 mamakera 14 mamnai 41 mamnavi 41 mamsa 53 manas 122 mandiram 168 mandulka 74 mandoka 74 mani 172 manikka 73 manikya 73 majiram 168 mannchi 7 marihai 5 marihami 6 maro 13 mase 183 matra 31 matta 31 mauda 137 mukka 74 maula 137 maunda 137 maura 137 mayi 141 mala 12 -maya 137 me 15, 16 mellai 55,72 mera 14, 15, 91, 97 mera 13, 15-16 mera 16 meradu 21 meram 21 meraya 13 meri 13 mero 20 metta 31 metthi 31 mharau 20 -mi 173 mil. 54 millai 54, 55 misimisemane 135 mluc- 71 moharam 21 mohora 19 mokkala 55 mokkalla 55 mokkallai 55 mor 15 mora 19, 20 morau 20 mottum 55 maduka 74 mrdutva 74 mus- 55, 56 muai 71 muc. 55, 56, 70, 71, 72, 125 muccahidi 9 muccihasi 9 mucilanda 136 mucilinda 136 mucira 124, 125 muijara 15 mujjhare 15 mukka 55, 70, 71, 72, 75 *mukka 75 mukkai 72 mukkala 55 mukkia 72 *mukna 55,70 mukta 55,70 mukula 137 mukunda 137 mukura 137 mukuta 137 mullai 55 muncai 71 munijjasu 150n2 murta 71 musti 75 matra 71 mutta 55, 70, 71 muthi 75 muula 137 munc 73 mi 13, 14, 84, 86 -ma 128-129 *ma- 14 mabh(y)a 13 madhura 13 madiya 13.20 madiyena 19 maha 13, 17, 18 maha(a)rau 18 maha- 13, 14 mahs-kera 14, 90, 91, 97 mahaera(a) 16 mahacra 15 mahakaro 13 mahakera/ka) 16 mahakero 13 mahara 12, 13, 14, 17, na u 84 na ya 87 na yalvi 87 nacira 124 Nigadevaya 46 nahi 83, 84, 85, 88 nahi 84, 85, 86, 88 nahi 85, 86, 87, 88 mahi(m) 88 nahim 83 nahu 84 nai 86 naf 19 paikka 73 naim/naim 85, 86 nim 101n naim/naim 85, 86, 86n, 101n naiva 84 najjai 87 nakhu 84 nakka 73 nakim 1010 naliara 17 nalieri 17 nalikera 17 tial 84, 86, 86, 87, 87n, 147, 161, 163 nam 87n nima 162 namamsihami 2 namira 124 namra 124 nanu 84, 87n nas 84 nas- 92 nasit 85 naska 73 natthi 84 nau 19,87 nadi 139 nauna 85 naundim 85 navati 139 navatima 128 navatitama 128 pavi 84 Daya 88 nayaka 74 nayanikaya 46 *nayikya 74 ne 84 pea 84 nesai 7 neura 136 neva 84 ni(s)sammai 121 nicchallai 54 nijjinahi 151, 152 nikka 74 niksad 54 nilijjejja 150 nisad- 121 nisam- 121 nisamay. 120 nisammai 119, 120, 121, 126n10 nisamyate 119 nisanna 121 nisannai 121 nisevehi 151 nisram 121 nissammai 120, 121 nit 169 niura 136 nivvellai 55 niyattaha 147 no 84 20 na 82, 83, 84, 85, 86, 87,87n, 173 na 84, 87, 169 na 87 88 -na 173 naa 87 na api 84 na asti 84 na ca 87, 88 na cid 86 na khalu 84 na punar 85 na tu 87 DO 84 maharau 12, 14, 16, 18, nollei 54 nollia 54 nu 159 nua 55 maharal 13 *mahera 15 maherala) 16 mahisikka 74 nud- 54.55 -nuda 54 nonam 84n - 208 * 209 -

Page #116
nupura 136 nutta 55 nuttam 54 nyayena 860 nyul 169 o- 77-81 odu 144 oggarana 79 ohu 101 oiro 78 ojjala 79 olambio 81 olla 53, 54, 55 ollei 54, 55 olliya 54 olugga 78 osappini 79 osarai 81 "ossar- 81n otam 144 otu 144 ovia 144 oviddha 81 oyar- 78 oyarai 81 pabhasei 80 pabhii(m) 60 pac- 70, 71, 73, 75 pacai 71 pacati 71 paccapinahi 147 paccayahii 2 paccayaissai 2 pacchi 31 padati 74 *padatikya 74 padhama 128 padhimilla 129 padihisi 9 pahallira 125 pahanao 81 pahasei 80 pahi 150 pahila 128, 129 pahilaraya 129 pahilari 129 pahilira 125 pahilla 128-129 pahillira 125 paholira 125 pahuppai 40 paikka 74 pajjati 171 pajjchi 146 pajjuvasejjahi 150 pakampira 125 pakarci 80 pakka 71 pakva 70, 71 palissajissati 168 palivei 53 pallanka 54 pamisu 7 pani 172 panna 173 parakka 73 parakya 73 parastat 32 paratra 32 paratta 32 parattha 32, 33 paribhacum 25 paricintira 126 paridhavira 126 parigamira 126 parikampira 126 parikilira 126 parimillira 126 parisakkira 125 parisvaj- 168 paritapp-80 parituttira 126 parivasium 24 parivilasira 126 parivvacijasi 150 -210 pariyapp- 80 parna 171 parusa 136 paryanka 54 pasammai 120, 121 pasannai 121 pasaya 96 pasca 38 pascima 128, 129 pasu 96 pasu 96 parin 31 pattammi 182 patte 182 patthi 31 pauhana 139 paukama 25 pavajissiti 11 pavihasi 9 pavisu 7 payacchahi 150 paya 23 payarci 80 payatta 24 pai 40 pekkhihimi 6 pellai 54, 55 phanillai 54 pharasa 136 phasa 32 phasuya 32 phirakka 74 pibati 54 piccha 38 picchi 37, 38 picchihimi 5,9 pikka 71 pikkhivi 40 pindevi 40 pitthi 38 *pitthi 37, 38 posaha 80 prabhas- 80 "prabhutvati 40 pradipayati 53 prahnetana 89 prajna 171 prajna 173 prakr- 80 prasad- 121 prasanna 121 prasram 121 prasuka 32 prasuka 32 prathama 128, 129, 145 *prathi(1)la 128, 129 pravahana 139 pravrtta 24 preryate 54 Prstha 38 Prsti 37 prthvi 37, 38 prus 32 pucchahi 147 pucchihamo 6 pudhama 145 pudhavi 37 Pudikaya 46 pukkai 73 pukkarei 73 pukkia 73 pummai 121, 123 puna 85 punaim 85 punar 85 punch- 132122 puracchimilla 129 puras 129 *purascima 129 purassara 130 purastat 32 puratana 89,91 purattha 33 purattha 32 purahsara 130 purisa 136 pina 100 purusa 136 purusa 136 puthavi 37 pukara 73 putkrta 73 putta 66 putta 66 *putthi 37 puvvi(m) 60 -ra 14 -ra 124, 132, 133 *raHis 141 raikka 73 raja 74 *rajikya 74 ramijjai 119 rammai 119, 120 ramyate 119 rankholira 125 rasala 14 rasira 125 rau 19 -rad 18 ravala 14 rayis 141 rellai 55 relli 55 resammi 89, 97-98 resi 89-98 resi(m)/resim 90, 97-98 resimmi 97 ri- 55 riddhilla 129 riddhima 129 rikka 71 rincai 71 riyai 55 ruc- 125 rucira 124, 125 rudh- 72 rudhira 125 -211 rumbhai 72 rundhai 72 rusesu 7 russai 71 sa 160 saa 59 sabha 102 sacca 37 sad- 121 sada 58-59, 60 sadatana 91 saddhasa 38 sadhila 136 sadhvasa 38 sadrs 131 sadrsa 132 sagala 100, 102 sah- 101 sahas 111 sahasa 111, 112 sahasa 111-112 sahasa-cciya 112 sahasa/sahasa 111, 114 sahasatti/sahasa-ti 1111112, 115, 116 *sahaya 101n sahijjasu 150 sahinti 8 sahira 125 sahu 100, 100, 101, 102, 103 sahu 100, 101, 101n, 102, 103 sal 59, 60 sai(m) 131 sal(m) 59 *sal(y) 59 sajjhasa 38 Sak- 71 sakala 99, 100, 102 Sakata 74 sakka 71

Page #117
sakt 131 sakta 71 Sakya 71 salakkia 74 sam- 121 sama 99, 100 samagga 100 samaggamissai 3 samagra 99, 100 samanavi 40 samapta 100 samasta 99, 100 samatta 100 samattha 100 samayte 23 sambha 102 samcara 38 samiya 161-162 samiyam 161-162 sammai 118, 119, 127n13 sammanyate 171 sampunina 100 sampulma 99, 100 samsarijasu 150 samuthie 23 samvatsare 183 samvellai 55 samvestate 55 saiyak 162 samyati 118 sana 172 sankhae 23 santa 76 Santa 76 sapatnl 139 sara 130 -sara 130 sara 102 sarai 142 sari 131 sari(a) 131-132 sarira 95 sarira 95 Sarika 74 sarl-vanno 131 sarisa 132 sarp- 72 saru 102 sarva 99, 100, 101, 102, 102n sarvada 58 sarve 100, 102 sas. 101 *sas- 101 sasa 100 sisa 101 sasaya 101n sassa 101n sasvat/sassvat- 99, 100, 101, 101n, 102 sasvala 1010 *Sataka 74 saakka 74 saakya 74 sapka 73 satpurusa 138 Salta 71 satya 37 sau 101, 102 saudamani 115 saulya 53 saumana 138 saumara 138 saumya 138 saurisa 138 sautti 139 sava 100, 101 sava 101 Savatti 139 savu 100n savu 101, 102 savva 100, 101, 102, Sunasu 151 supurusa 138 surisa 139 sus. 71, 72, 75 Suska 71, 72, 166 suumala 138 Suumara 138 suurisa 139 suvai 127n1l *suvai 119 Suvanad 26 svaka 74 svap- 127n6, 127n11 svapiti 127011 svapnakam 26 svaskna 72 svasra 95 svasti 31 svastika 31 ST-130,142, 143 savve 100, 102 savvesim 97 savvi 100, 102 stya 59, 60 *saya 60 sayala 100 sayalu 100 se 160-161, 162 sed 160 seuya 140 sevaya 139 seveka 140 si 160-161 sidhila 136 sijjhahidi 9 sijjhihim 4 sikka 71 simantammi 181 simasimaya- 135 simisimiya 135 sine 73 sincai 71 sithila 136 sithira 123 skand-143 skanna 171 slaghya 102 slis- 72 so je 106 so ji 106 sokkhai 72 soksyati 71 soll. 53 solla 52, 53, 54, 55 solla- 53 sollei 54, 55 solliya 53 soma 138 somala 138 sotthi 31 sparsa 32 sparsuka 32 sprs 31-32 ST. 130 sram- 76, 121 sramati 127n5 sramyati 118 sranta 76 sruth- 136 sri- 143 Srira 13 stj- 142, 143 srotriya 31 spta 130 srth- 124 *sthila 136 sru- 120, 122, 127n11 sruyate 119 stabh- 79 *staq-ne 69 stha 75 sth]- 30, 69, 75, 76, 167, 171 "sthakyati 69 stheya 76 sthita 75 sthuna 172 stim- 127n13 stimita 127n13 stimyati 127n13 stokya 73 Sual 127n11 sud 53, 54 suda 54 sudita 53 *Odna 53 sukka 71, 72 sukumara 138 Sula 53 Sulya 53 *sumai 119 summai 119-120, 122, 126n10 sunahi 150, 151 sunai 120 Sunasu 151 ta 20,66-67. 155-156 ta- 13 * 14 to- 62 taa 59 tacca 37, 129 tacca-vaya 37 tacchai 166 taccu 37 tad 90, 92, 160 tad. 115, 116 tada 29. 58-59 tada hi 65 tadatti 115 tadad-iti 112 tadakka 115 tadakkia 73 tadatadatudanta 115 tadatadento 115 tadatti 114-115, 116 tadhima 155 tadisa 68 tadit 114-115 tadittadiya 115 taditti 116 tadria) 29 tadru 30n, 50n tae 48 taha 83, 155 taha 20 tnham 31 taham 93 tahara 19,20 lahe 48, 49-50 tahe 35,48 *tahialai 65 tahim 49, 180 tahim 65, 66, 116 lai 40 tar 58, 59, 64 tara 58 taisa 68 Laisa 68 tal(y) 58, 59, 60, 63, 64, 65, 68 talyahan 65 taiyahim 65, 66 talyahim 66 talyaho 65 talyahum 58,65 taks- 166 taksati 165 talayati 170 talayati 170 taliantira 125 tam 39, 160, 163 lam- 127n13 amma 147 tammai 118, 119. 127n13 tammi 178, 181 timyad 121 tamyati 118 tan 94 tan- 89,92 103 savva(y) 58 savvatto 34 - 212 - - 213

Page #118
-tana 89,91 tanai 19 Lanai 90, 91 Lanailla 129 tanaim 90, 91 tanam 93 tanau 19, 90, 91 tanaya 90, 92, 96. tanaya 89-98 *tanaya 92, 93. tanena 90 tani 96 tanka 170 janka 170 tanu 94-96 tanu 95, 96 tanu-bhedin 95n tanua 96 taquli 96 taquei 96 tanuhei 95n tanuvaga 94 tanuvaka 94, 96 tanuvaya 96 tao 48 tapta 33 tarhi 29, 35, 49 tarisa 68 tasa 48 tase 49, 50 tasiu 25 Lasmat 67 tasmin 65, 177, 178, 180 tassakera 14 tassim 177, 178, 180 tassim 177 Lasu 48 tasyah 47, 48, 49 tasyai 48 tasyam 48, 49 tat 66, 67 tat-stha 30 tatas 29, 34, 50 tatha 29, 30, 31, 83 tathya 37 tathya-vadin 37 tati 29, 64 tatra 29-30, 30n, 33, 35n, 50, 50n tatta 30, 33, 35n tattha 33, 34, 35 tattiya 34 tatto 34, 35 tattu (?) 35n tattva 22, 37 tattva-vada 37 tava 19, 62 tava 66 tavat 29, 66, 67, 68, 154 ta(y) 58 taya 48, 58 taya 48 "ayada 59 *tayah 48 taye 48 *Layida 60 to 15 leaya 96 teddaha 63 tehim 90 telecha 165 tejas 96 teka 169 Jeka 169 telokka 74 tena 155 tera 14, 15 tera 16 tera 16 teram 21 teraya 13 teri 13 tesam 61 tesim 93,97 -214 tesu 61 tettahe 35, 35n, 36 tettha 36 tetthaho 35 tetthu 35 tettiya 34 teu 96 tevada 62 tha 75, 76 thai 167 thakka 69-76 thakkai 69-76 Thakkana 170 Thakkana 170 thapillai 54 thathatharp 170 thashantham 170 "theakka 76 the(y)a 76 thia 75, 76 *thiakka 75 thimmai 121, 127n13 thokka 73 thutkara 73 - 48, 50 Li-sara 130, 131 tia 48 tie 48, 50 tigiccha 165, 166 tiha 31n tiha 133 tiha 133 fik- 170 tikka 74 tilaka 74 tim- 127n13 timbaru 1361 timira 125 timita 127n13 timmai 120, 121, 127n13 timyati 127n13 tindisa 170 -tindisa 170 tise 48, 49, 50 tissa 47, 48, 49 tisthati 165, 166, 167 *usyai 49 *tisyah 47, 49 *Hisyam 49 titha- 167 titibha 170 titila 170 itra 169 tittha 34, 35, 36 titto 34, 35 Liya 48 tman 93 to 19, 157 tohara 19 toharam 21 tohauri 19 tohora 19 tohori 19 *tokka 75 tola- 171 jola 171 tor 15 tora 19, 20 joja 170 trailokya 74 tram 50n tridha 133 *utya 129 trut 75n. tti 111 tu 87 tubble 146, 147 tubh(y)a 13 tubhyam 39 tuccha 165, 166 tuddha 50n tuddhra 50n tuha 13, 14, 17, 50n tuhaera(a) 15 tuhakera(ka) 16 tuham 148 tuhara 12, 14, 17, 19, 201 tuharaya 13 tuhari 13 tuhera(a) 16 tuhu 148 tuhu(m) 101, 148 tujjha 19,91 tujjhanau (JMah, Ap) 19, 91, 91n. *Jukka 75 tumam 39 tumam 146, 147 tumbaru 136 tumbha 21 tumbhara 20 tumburu 136 tumha 16 tumhaha 16 tumhakera(ka) 16 tumham 21 tumhanam 16 tumhara 12 tumhara 16 tumharaya 13 tumbe 21 tumhera 15 tumhera(a) 16 tuphikka 74 tura 16 turai 39 rya 16 *tusmat 21 tusnika 74 tuttavi 40 tutti 40 tuvam 39, 39n Iva 14 tvadiya 20 tvaks 166 tvaksati 165 tvam 39 -215 tvayi 141 ta- 79 tuya 129, 131 -tti 112,114, 157. u 84, 159. u- 77-81, 143 ulya 80 ubbhe 40. ucchilla 53 ucchvas- 80 uccinai 143 ucciniya 143 ud- 53, 54 ud- 79, 80, 81, 143 uda(ka) 54 udkrs- 79 udgarana 79 udirana 80 *udla 53 *udra 53 udstabh- 79 udia- 79 udvestate 55 uggaha 78 uhavana 77 uhu 101 uinna 77 uirana 80 uitha 78 ujjagaraya 124 ujjagira 124, 125 ujjal 79 ujjhium 24 ujjvala 79 ukkas- 78, 80 ukkhand- 78 ukkhanda 143 ulagga 78 ulambia 77 ulambio 81 ulla 53, 124 ullatta 143

Page #119
ullira 124, 125 uloyana 77 ulugga 78 upa- 80, 143 upakranta 81 upanahau 80 upavas- 81 upavasatha 80 *upavasta 81 upavisati 80, 81 upaya 80 uravia 143 Urdhva 40 -usa 136 usaraha 77 usarai 81 Usas- 80 ussappini 79 ussarai 81 ussarai 81n ussas- 80 ut- 143 ut-chidra 53 utci- 143 utlas 54 utsarati 81 utsarga 79n utsarpini 79 uttamsa 78 uttarai 81 uttarati 81 uttarilla 129 uttarium 24 uttariya 81, 81n utthambh- 78 uvanimantejjahi 150 uvari(m) 60 uvarilla 129 uvarima 129 uvasaggeum 24 uviddha 81 uviddhani 77 uvvellai 55 uvvillai 55 uyar-77 uyaraha 78 uyarai 81 uyarami 77 uyarami 80 uyariya 77 vac- 54, 72, 142, 144 vaccaha 147 vaccasa 144 vaccattha 144 vacciamo 6 vad- 54 vadanya 171 -vadda 62 vah- 142, 144 vahi 150 vahitra 144 vahya 144 vai 106 vajjejja 150 vakkanta 81, 81n. val 54 vallai 54 valli 55n vallira 125 vamisu 7 van- 141 vandejjahi 150 vandihami 2 vap- 144 vaplavyam 144 varaha 144 vartitum 39 *vartitina 39 *vasisyatha 10. vasu 151 validuna 39 valihami 2 vauuha 81 vava- 143 vavahara 143 -216 vavaharaya 143 vavasaya 143 Vay s ve- 144 vea- 55, 56 veccei 73 vel- 55 velae 131 velayam 131 vell- 55, 55m vella 55 vellai 55 vellati 55 velli 55, 55n vellita 55n vep- 55 vepaka 140 vevaya 140 vevira 125 vi- 143 via- 55, 56 vicalira 125 viccei 73 vidvan 38 vidya 38 vidyadhara 165, 166 vidyut 115 vigna 1441 vihallai 54 vihrad- 54 vij- 55, 56, 144 vij. 144 vijanahi 151 vijita 168 vijja 38 vijjahara 166 vijjai 144 vijjajjhara 165, 166 vijjam 38 vijju 115 vijjula 115 vijna 38 vikkei 73 vikkeum 73 vikkia 73 vikkinci 73 vikri- 73 vil- 55 vilna 55 *vilvati 55 *vilyati 55 vilina 143 vimuccisam 6 vimunca 147 vinasaka 92 vip 55 vip- 55 viparita 18 visa 100 visa 100 visagga 143 visai 80 visajidavo 154 visajidavya (Kharoshi inscription) 154 visarai 143 Visasena 100 visate 80 visaltai 143 vist- 143 Vissabhui 100 Vissasena 100 visva 99, 100 visvam 100n vitigimcha 165 vivareraya 18 viyosidavo 142 viyosita 142 vo(1)lai 143n10 vo- 141-145 voccattha 144 vocchami 54 vocchami 144 vocchindai 142 vocchinna 142 vocciniya 143 vodala 142 vodana 142 vodha 144 vodhum 144 vodra 142 vodraha 142, 144 voharei 143 vohittha 144 vojjai 144 vojjha 144 vojjhai 144 vojjhara 144 vojjira 144 vokka 72 -vokka 72 vokkai 72 vokkanta 143 vokkasei 142 vokkhai 72 vokkhandaya 143 vokkhariya 143 vola 142 volatai 143 volia 144 volina 143, 143n10, 144 vollai 54, 72, 144 Vollasaka 142 vollhaha 142 voma 142 vonta 142 Vopadeva 142 vora 140 voraccha 142, 144 vorahiu 144 voraka 142 voramana 142 vorata 142 vorava 142 voravia 143 vorukhana 142 vosagga 143 vosamiya 142 -217 vosarai 142-143 Vosaltai 143 vosajha 143 *vosaya 143 vosido 142 vosijja 143 vosirai 142-143 vosiritta 143 votavi 144 vottum 54, 144 vraj 54, 144 vrasc- 71 viddha 62 vikna 71 vinta 142 *vrtivana 39 vucca- 54 vujjai 144 vukka 71 vunna 144 vunna 144 vutta 54 vuttam 54 vyapa- 143 vyapaharayati 143 vyapakarsayati 142 vyapakranta 143 vyaparopita 143 vyath- 55 vyaty lisa 144 vyatyasta 144 vyava- 142, 143 vyavachinna 142 vyavadana 142 vyavahara 143 vyavaharaka 143 vyavasamita 142 vyavasaya 143 vyavasita 142 vyavasrayati 142, 143 vyavasri- 143 vyavasrita 142 vyavasrj- 142

Page #120
yeva 110n6 vyavasrjya 143 vyavasrsta 143 vyoman 142 vyuparamana 142 ten 93 tja 129 to 12 1or 12 Bundeli mo 14 choch 166 dekhl 9 do-sar 132 ethain 35 ethan 35 jethan 35 kethan 35 tehan 35 ti-sar 132 Chattisgarhi dekhahi 6,9 ya 87. 110n13 yad 45, 108, 154-156, 161 nisanaya 23n Osadhani 187 palakkamatu 150 palitiditu 168 palitiditu 168 paricajitp 40 poja 44 se 160 srunaru 150 sussusatu 150 tata 30 tista- 167 titha- 167 tlanaya 39 tulaye 39 vadhis siti 10 vadhiya 46 ye 45 yeva 106, 108 Avestan bi-sarm 130n11 hama 99 haurva 99 rosa 165 Pri-sarim 130n11 vohu 141 yezi 141 Yima 141 Bihari kerau 15 Dardic tam (Maiya) 95 tan (Khowar) 95 tani (Garvi) 95 tanik (Pashai) 95 tanu (Torwali) 95 tanu (Tirahi) 95 yada 29, 58, 59, 131 yada hi 65 yadi 59, 65, 141 yadri Son yadrsia) 29, 108 yahi 150 yajfiam 173 Yama 141 yani 108 yarhi 29, 35, 49 yasindt 67,155 yat 87 yat 66, 67 yat-stha 30 yatas 29 yatha 29, 31, 108, 155. Dhakki tuham 10in Asokan inscriptions abhivadetnam 39 adisa 108 alocetpa 40 am 108 anata 172 apaca 167 apacca 167 apatiya 167 arabhitpa 40 asvasanaye 23n ata 108 atha 108 atta 32 ayam 68 ava 108 chamitave 23 citha- 167 dahamti 1 dasayitpa 40 e 108 edisa 63 chatha 1 ena 108 eta 29.33 ctra 29, 33, 34n suar 61 gananayam 46 halapita 187 hemeva 106 hela 29, 33 hevam 106 hohamti 1 isaya 44 khamitave 23 ksamanaye 23, 23 lopapita 187 nasti 172 nati- 172 fiitika 172 Braj itai 35 Tai 35 jita 35 kerau 15 me 14 nahi 82 tita 35 uta 35 wah 42, 43 wo 42, 43 wuh 42, 43 yah 43 157 yathairham 108 yati 29,64 yatna 33 yatra 29, 108 yra 33 yavat 29 yavat 66, 67, 108 yavatkatham 108 yavi 87 "yayada 59 "yayida 60 ye 108 yena 108, 154 ySim 61 yugala 165 yugma 165 yusmakam 19 Avadhi calai 41 dja 129 etan 63 -hau 10 -ihahu 10 -ihasi 10 jetan 63 kahesu 152 kara 18 karihahi 10 karihi 10 karihi 10 kera 15, 18 kere 15 marihai 10 marihar 10 mo 12 mor 12 niwantesu 152 padhesu 152 pujihi 10 Sunai 41 Bengali apa 94 -be 66 cacha 166 cauhira 132 chuch 166 dohara 132, 133 doja 129 dosra 129 ekahara 132 ekasara 130 eta 63 eto 63 hetha 35 jetha 35 jai 65, 66 (kerau 15 nay 86 sara 102 tai 65 tar-tar 116 tehara 132, 133 teja 129 tesra 129 tia ja 129 Buddhist Hybrid Sanskrit bharyaya 44 hetu 162 jhlsayati 38 katima 64 katira 64 makuta 137 manasikara 146 osarati 142 osirati 142 osirati 142 procha 1463 tuva 146 tva 146 ukirati 78 Gujarati andhanum 17 chel 139 ekkasara 132 halvu 54 hraim 13 hrau 13 - 106 jhadakka 114 jhadapvu 114 jharmar 135 jhat-jha 116 jhirimiri 135 kahia(m) 88 kara 18 kera 15 kerau 15 kisa 163 kisa 163 ma 20 maharau 13, 20 mahi(m) 82 maro 12 melvu 55 Bhojpuri cachal 16 - 219 - - 218

Page #121
mokalvu 55 nahia(m) 82, 88 nahi(m) 88 nathi 84 nau 19 ne 93, 96 osarvu 81n ovi 144 paheln 128 pote 39n rahaim 13, 14 rimijhimi 135 rumjhum 135 sahu 100 sahu 102 sau 102 sav 102 taharau 13 tahi 88 tan 95 taniya 96 taro 12 tiham 31n utarvu 81 Hindi agunis 139n16. aksar 130, 132 andhera 17 becna 73 bijli 115 bikna 73 bohra 143 bojh 144 bolna 54 buhama 143 cachna 166 cauhra 132, 133 causar 131, 132, 133 chich 166 cumna 123 dhalakna 75 dhalna 75 dhasakna 113 dhasna 113 dhuni 38 dohra 132, 133 duj 129 dumna 123, 126 dusra 129, 131, 132, 133 ekahra 132, 133 eksar 132 ghumna 123, 126 haina 54 iksar 130, 132, 133 is 28 jab 57 jahan 28 jani 87 jevna 12713 jhanda 38 jhat 116 jhat-jhat 116 jhat-pat 116 jhunjhuna 38 jhumma 38 jin 87 jis 28 jola 165 jula 165 junhai 165 jyon 28 kab 57 kahan 28 khelna 31 kis 28 kyon 28 lamakna 123 lumna 123, 126 mero 20 mero 12 mukka 75 muth 75 nahin 82-88 naim 86 -220 ne 93,96 pahila 128 picha 38 pish 37 rokna 72 sab/sabh 102 sara 102, 103 sarak (ie sarak) 74 sari 103 tab 57, 66 tahan 28 taisa 68 tan 95 tehra 132, 133 tero 12 thek 76 thik 76 uj 129 tin 93 tis 28 tisra 129, 131, 132, 133 jokra 75 juk 75 Lurant 39 tyon 28 us 28 usama 81 utama 81 vah 102 vahan 28 van 28 yahan 28 yon 28 Kashmiri dula 165 pan 39n tel 65 yeli 65 Kumaoni sariko 132 tanana 93 tanh 93 Lahnda maresiu 7 mareso 7 marese 7 maresi 7 maresa 7 marsaha 7 sara 102 Magadhi jethavam 35 Maithili elay 35 jatay 35 katay 35 Otay 35 tatay 35 tate 35 Marathi ahnem 83 -c 107 ci 107 evham 66 jm 65 jhat-jhap 116 kalim 65 kevdha 62 mokalnem 55 nahim 83 nahi 82 nem 93 nim 93 osamga[?] 79n ovl 144 pahila 128 sara 102n tm 65 vovi 144 Marwari dujo 129 ekama 129 karasum 8 labhasi 8 sisasi 8 ujo 129 vijo 129 Mewari sara 102n tanam 93 tani 93 Nepali dhuni 38 diuso 140 eksaro 130 halinnu 54 jhaskai 38 jhaskanu 38 jhulkanu 75 jhulnu 75 jun 165 kahi 88 kaile 65 kailhe 65 khumcinu 123 nahi 88 par 39n sari 132 tahile 65 taile 65 taille 65 tan 95 tara-tar 116 uira 78 Oriya cachiba 166 chachiba 166 cucha 166 tachiba 166 -221 tar-tar 116 tucha 166 Pali agaru 135n8 ambala 138 annatra 172 ayam 68 daddallati 168n19 devaya 43 diguccha 166 dosapagata 165 dosina 165 etto 34 handa 109 hetu 162, 163 junha 165 kahati 8 kamma 171 kanna 171 kanna 171 kanna 171 kassa 163 katima 64 katipatha 64 khida 31 kilanja 136 kissa 163 malaya 43 nati 172 natthi 84 palissadissati 168 panna 171 panna 171 panna 171 Pasenadi 168 phoseti 31 porisa 136 sabba 100 sammannati 171 sarivanno 131 sotthi 31 southiya 31

Page #122
CORRIGENDA saghala 102 suhanau 26 teha 102 tharo 20 Page For Read Shina yun 165 ittha tassa 47.48 tassam 48 taya 48 tayam 48 tidha 133 likicchati 166 timbaru 136 tissa 48 tissam 48 vadannu 171 vidita 168 vissa 100 vissam 100n yeva 106-108, 110 98 131 Sindhi tan- 93 176 3rd line from top 1st line 2nd paragraph note 15 12th line from bottom 2nd line from botton 9th line from top Ittha geater Upartye 1952 possilbe 1931 greater Upadhye 1954 possible 1937 179 Torwali tanu 94 Panjabi andhera 17 cummna 123 ne 93, 96 oh 42 sara 102 tacchna 166 tucch 166 uh 42 Persian ki 158 Rajasthani cha 102 huim 83 jeha 102 karasu 8 karasi 8 karisa 8 keha 102 kerau 15 merad 14 mero 20 morau 20 moral 14 ranau 19 rau 98 222 -223

Page #123
PL 1420 1947 1991 THE AUSTRALIAN NATIONAL UNIVERSITY FACULTY OF ASIAN STUDIES PUBLICATIONS ON SOUTH ASIA Barz, R.K. and Y. Yadav. An Introduction to Hindi and Urdu (4th edition in preparation) Proudfoot, Ian. Ahimsa and a Mahabharata story: the development of the story of Tuladhara in the Mahabharata in connection with non-violence, cow protection and sacrifice (1987, xiii + 269 pp.) $20.00 Yadav, Y. Advanced Aural Exercises in Hindi (1985, xi + 134 pp. 6 audio-cassettes) book $12.00, cassettes $36.00, book and cassettes $43.00 Kennedy, Kenneth, A.R. The physical anthropology of the megalith builders of South India and Sri Lanka (1975, 93 pp.) $6.95 Bailey, H.W. Sad-dharma-pundarika sutra: the summary in Khotan Saka (1971, 57 pp.) $5.00 A.L. BASHAM LECTURES 1990 Jordens, J.T.F. Gandhi, conscience of Hinduism and scourge of orthodoxy $4.50 1989 Schmithausen, Lambert. Plants as sentient beings in earliest Buddhism $4.50 1988 Ray, Ajit Kumar. Rabindranath Tagore: a poet of the earth $4.50 1987 Mukherjee, S.N. Calcutta: a city of splendid palaces and dingy streets: fiction as history $4.50 1985 Thiel-Horstmann, Monika. Symbiotic antinomy: the social organisation of a North Indian sect $4.50 1981 McGregor, R.S. A new voice for new times: the development of modern Hindi literature $3.00 2017-0940 All prices in Australian dollars. Distributed by Bibliotech, ANUTECH Pty Ltd, GPO Box 4, Canberra ACT 2601, Australia Telephone (06) 249 2479 Fax (616/06) 257 5008

Page #124
Faculty of Asian Studies Australian National University Canberra 1991 ISBN 0 7315 12200