I 5 f I? i i 'A < 5 7v^ s v_^ v< Q.JO^ %):H1V3:10^ ^OF-CAllFO/?x/ g! 4 ' I 4? ^ ' ' ' v 1 - giH^ ^Aavaan-T^ ^ s ^\\EUNIVER% ^. ^, _ ^g CALIFO/?^ ^VOS-ANCEtfj> %UWS01 1 1 I 'II %)J 1 I g > s g 7^ ^> i I fyMAINIHVk ^lOSANCFlfj^ ^l-UBRARYQc, ^X-UBRARY^ 3 s/^w-*^ si ir^^ si ir^ ^ 6 l^lr^S sLinil S I Jffl i i mi i MI v**n i^ g a ri. r-j 1 ? 3 S I 1 i ,_ )jnw-*' < s ^ ? tfHIBRARYQ^ ^E-UNIVERS/^ 1 1 ^ S = g fe \V\EUNIVER ^ I 1 s S 1 z< ^ i i * I 1 I' 3 a & S i g = ^E-UNIVER% ^lOSANCElfX^, ^OKALIFOfy^ I^T^Nl THE SUFFIXES /f^WAND VANT IN SANSKRIT AND AVESTAN BY HAROLD H. BENDER, PH. D. BALTIMORE J. H. FURST COMPANY 1910 CONTENTS, Stack Annex 5 033 PAGE SANSKRIT GRAMMAR REFERENCES, vii AVESTAN GRAMMAR REFERENCES, viii INTRODUCTION, 1-2 PART I. THE SUFFIXES MANT AND VANT IN SANSKRIT. CHAPTER I. THEIR PHONETIC CORRELATION. SECTION PAGE 1. Relative frequency of mant and vant, - 2. Parts of speech to which the suffixes are added, ... 3_6 3. Table of stem-finals, - 6-7 4. The suffixes in Indo-European, 7-9 5. Distinction between the suffixes in Sanskrit, - 9 6. Panini's rules, 9-10 7. Benfey's rules, 10-11 8. Sanskrit rules for use of the suffixes, - 11 9. Observance thruout the literature of rules for stems with final vowels, - - - H-23 10. Observance in the Veda, 24 11. Observance in Rig-Veda, 24-26 12. Observance in Atharva- Veda, - - - 26-28 13. The case of -ovant, ----- 28 14. Observance thruout the literature of rules for consonantal stems, 29-31 15. Observance in Rig- Veda, - - 31-33 16. Observance in Atharva- Veda, 33-34 17. Historical survey and comparisons, - ... 34-36 18. Causes of interchange between the suffixes, - 36-40 19. Forms in the Veda with both suffixes, - - - 40-42 20. Other double forms, 42 21. Final sands, 42-43 iii 2005042 IV CONTENTS. CHAPTER II. SUCTION PAGE 1-2. Internal combination, .--..... 43 3. External combination, ----.... 43-45 4. Diastole of final stem-vowel, ....... 45-52 5. Systole of final stem-vowel, - - ..... 52-53 6. Loss of final vowel, --------- 54 7. Loss of final consonant, -----... 54-56 8. Change in quality of final vowel, -.-... 56-57 9. Insertion of letter between stem and suffix, - ... 57-58 10. Miscellaneous, ... ..... . 59 CHAPTER III. MEANING. SECTION PAGE Introduction, ---------. 59-60 1. Primary signification, 60-61 2. Formation of present active participles, ----- 61-62 3. ' Containing the root or word ' , ----__ 62-63 4. No change in meaning, -------- 63 5. Formation of past passive participles, - .... 63-64 6. Formation of past active participles, - - ... 64-65 7. Suffixes used actively instead of passively, - ... 65-66 8. Suffixes express relation, -------- 66-67 9. Suffixes express accompaniment, ----.. 67 10. Meliorative or majorative use, ------- 67-68 11. Suffixes mean ' surrounded by ', 68 12. Use as noun-suffix of agency, ------- 68 13. Suffixes mean 'consisting of ', ------- 68-69 14. Suffixes have causative value, -----.. 69 15. Suffixes express contiguity, ----... gg 16. Suffixes express resemblance, 69-71 17. Suffixes equivalent to English suffix -able, - 71-72 18. Suffixes mean ' engaged in',- - - - - - . 72 19. Pejorative use, ----.-... 72 20. Miscellaneous, - 72-74 21. Conclusion, 74-76 CONTENTS. V PART II. THE SUFFIXES MANT AND VANT IN AVESTAN. CHAPTER I. THEIR PHONETIC CORRELATION. SBCTIOM PAGK 1. Relative frequency of mant and vant, ..... 77 2. Parts of speech to which the suffixes are added, - - - 77-80 3. Rules for use of the suffixes, 80 4. Observance of rules for stems with final yowels, ... 80-83 5. Observance of rules for consonantal stems, .... 83-85 6-7. Comparison of Avestan with Sanskrit observance of the rules, - 85-88 CHAPTER II. SAATOHI. SECTION PAGE 1. Diastole of final stem- vowel, ------- 88 2. Systole of final stem-vowel, 88-89 3. Loss of final vowel, 89-90 4. Loss of final consonant, -------- 90 5. Change in quality of final vowel, 90-91 6. Insertion of letter between stem and suffix, - ... 91 7. Contraction, 92 8. Metathesis, 92-93 9. Miscellaneous, 93 CHAPTER III. MEAKING. SECTJOX PAGE Introduction, 93 1. Primary signification, 93-95 2. Formation of present active participles, ----- 95 3. 'Containing the root or word', ------ 95-96 4. No change in meaning, -----.,.. 95 5. Formation of past passive participles, ----- QQ Vi CONTENTS. SECTION PACK 6. Formation of past active participles, 7. Suffixes used actively instead of passively, - ... 97 8. Suffixes express relation, 97 9. Suffixes express accompaniment, - - 10. Meliorative or majorative use, 97 11. Suffixes mean ' surrounded by ', 97-98 12. Use as noun-suffix of agency, - 98 13. Suffixes mean ' consisting of ', ------- 98 14. Suffixes have causative value, ------- 98 15. Suffixes express contiguity, ------- 98 16. Suffixes express resemblance, 98-99 17. Suffixes equivalent to English suffix -able, - ... 99 18. Suffixes mean 'engaged in', 99 19. Pejorative use, 99 20. Miscellaneous, 99-100 21. Conclusion, 100-101 INDEXES, 103-116 SANSKRIT GRAMMAR REFERENCES. Panini, I. 1:54; iv. 2:72, 85-88; V. 2:94-101, 109, 115-118; VI. 1 : 219-221 ; 3:119, 120, 131 ; vm. 2 : 7, 9-17. Benfey, Grammatik der Sanskrit-Sprache, pp. 170, 239, 243 and 464, 465, 562-565. Monier- Williams, Sanskrit Grammar (3rd edition), 45 a., 69 a., 84 (iv and v), 140. Stenzler, Elementarbuch der Sanskrit-Spraehe (3rd ed.), Grammatik, 80, 215. Biihler, Leitfadenfiir den JElementarcursus des Sanskrit, pp. 43, 68. Whitney, Sanskrit Grammar (3rd edition), 111 d., 247, 452- 458, 494, 517, 959, 960, 1107, 1233, 1235, 1307 b. Thumb, Handbuch der Sanskrit Grammatik, 327, 328, 329, 402, 618. Wackernagel, Altindische Grammatik, Vol. II, 53b. Brugmann, Grundriss, (2nd edition), n, 1, 351-357. vn AVESTAN GRAMMAR REFERENCES. Justi, Handbuch der Zendsprache, Grammatik, 372, 384, 576- 578. Hovelacque, Grammaire de la Langue Zende, pp. 77-81, 112 (2. 6d. 153, 158). Harlez, Manud de la Langue de V Avesta (2. e-d.), 74, 75, 97. Geiger, Handbuch der Awestasprache, Grammatik, 63, 70, 106. Spiegel, Vergleichende Grammatik der altiranischen Sprachen, 92, Suffix No. 84 ; 93, Suffix Nos. 47, 57 ; 132, 164. Kanga, A practical Grammar of the Avesta Language, 70, 71, 132, 135, 138. Jackson, Avesta Grammar, 291, 294, 295, 296, 297, 376, 433, 441, 442, 759, 821, 826, 851, 857. Bartholomae, Handbuch der altiranischen Dialekte, 198, 199, 272. Vlll INTRODUCTION. One of the most interesting correlations in Indo-Iranian Gram- mar is that of the twin suffixes mant and vant. A detailed study of them, however, has not heretofore been made. Panini's treatment is obviously artificial and mechanical, as is that of Benfey, who follows him in method and in most of his results. Most of the other Sanskrit grammarians, such as Monier- Williams, Biihler, Stenzler, are content with the statement of rules for declension. Thumb devotes more space to the suffixes, but adds little to our knowledge of them. Lanman (JAOS. x, 515 ff.) and Bartholomae (KZ. xxix, 490 ff.) treat in detail the suffixes for the Rig- Veda and for Indo-Iranian, respectively, but their dis- cussion is largely restricted to inflexion. The paragraphs on mant and vant in Lindner's Altindische Nominalbildung (pp. 136 and 146) give little beside lists of examples. Whitney's treatment is a general one and disproportionately based upon the older language. Wackernagel's scholarly work has not yet reached the suffixes. 1 In Avestan the situation is almost equally unsatisfactory. Bartholomae's discussions in his Handbuch der altiranischen Dialekte and in the Grundriss der iranischen Philologie are very brief. The grammars of Harlez and Geiger have practically nothing on mant and vant, while those of Hovelacque and Kanga consider only the declension of the suffixes. Justi gives a good list of examples but no discussion. Spiegel and Jackson treat the suffixes in a more satisfactory manner, but not in much detail. Hence an effort has been made in the following pages to reach, J My study of mant and vant was practically completed when Wackernagel's article appeared in KZ. XLIII, 277. Consequently I have not in the body of my work given it the consideration it deserves. I have, however, treated many of its essential points in foot-notes and have made frequent references to his views in regard to individual words. 1 2 INTRODUCTION. from a study of all the material available, more definite and detailed conclusions in regard to the role these suffixes play in Sanskrit and Avestan than have been reached heretofore. The list of possessives in mant and vant upon which this study has been based was collected, for the Sanskrit, from Monier- Williams' Dictionary (second edition, 1899), but a number of words has been added from other sources. For the Avestan, Bartholomae's Altiranisches Worterbuch was of course used for the collection of material. PART I. THE SUFFIXES MANT AND VANT IN SANSKRIT. CHAPTER I. THEIR PHONETIC CORRELATION. 1. Of the 2200 words with mant and vant suffixes that have been collected for the purpose of this study, 1748 have vant and 452 have mant. Whitney says (Grammar, 1235) that in the older language there are only about one-third as many mant suffixes as there are vant. In the Rig- Veda this ratio holds almost exactly (74 : 221), but in the Atharva-Veda there are almost four times as many words which take vant as there are which take mant (32 : 122) and, indeed, as the figures above show, in all periods of the language taken together the mant suffixes are only about one-fourth as numerous as the vant suffixes. Or, in other words, of all the words with mant or vant suffixes four-fifths end in vant and only one-fifth in mant. 2. a. Most possessives in mant and vant are formed from nouns, which may be of any declension whatsoever. b. But a large number are formed by adding vant to the past passive participle. This suffix is so used, however, only to form new participles, viz., present active, past active, and past passive participles. In the last case there is of course no change in meaning. This function of vant is further restricted to the later language, there being only one example in Rig- Veda, and one in Atharva-Veda of the use of the possessive suffix with a past participle (cf. Chap, in, 2, 4, 5, 6, 12). c. There are a number of possessives which are derived directly from adjectives. Of course, in some of these cases, nouns may once have existed from which the possessives were formed, but 3 4 THE SUFFIXES MAXT AND VANT IX SANSKRIT. which have since been lost. Thus paravant, ' dependent,' may have been derived from a noun para, 'a superior,' instead of from the adjective para, ' superior.' A lost abstract noun posyd may have been the base of posydi-ant, ' causing prosperity,' instead of the adjective p6sya, ' prosperous.' In most instances the word belongs to Chapter in, 4 where the suffix causes no change of meaning when added to a stem. Following are examples : paravant, ' dependent,' < para, ' superior.' aparavant, 'having no superior,' < apara. 1 dgumdnt, f swift,' < c ft, id. kridumant, ' playing/ < kr'idu, id. nyubjimant, ' bent,' ' crooked,' < nyubja, id. dhunimant, ' roaring,' < dhuni, id. posyavant, 'causing prosperity,' < posya, 'prosperous.' pramattavant, ' careless/ < pramatta, id. apramattavant, ' not careless/ < apramatta, id. abaddhavant, ' unmeaning/ < abaddha, id. uttardvant, ' being above/ < uttara, ' higher.' ksdmavant, ' burned/ ' charred,' < ksarnd, ' burning/ ' burned.' ubhayavant, ' having both/ < ubhdya, ' both.' rtviyavant, ' having menstrual courses/ < rtviya, ' being in proper time/ ' regular.' vibhumdnt, ' mighty ' or ' omnipresent/ < vibhu, id. susumdnt, ' very stimulating/ < sust, id. (cf. Chap, n, 5 and note). vUthavant, ' beaming/ f shining/ < vibha, id. divitmant, derived by Mouier- Williams from an adjective divit, which he quotes for K-V. x, 76, 6 and translates ' going to the sky.' But Grassmann makes divit a masculine noun in that passage and renders it ' Glanz.' From this noun he derives divitmant (cf. Aufrecht, KZ. n, 149 and Grassmann, KZ. xi, 9). nilavant, ' blackish/ ' dark/ < nila, ' of a dark color.' 1 On this word and the preceding one cf. note to Chap, in, 16. THE SUFFIXES MANT AND VANT IN SANSKRIT. 5 isdvant, ' vigorous/ < (according to some) isd (cf. Chap, n, 9'a.). mdhindvant, l great/ ' powerful/ or (according to some) ' exhila- rated/ ' excited/ < mahina, ' great/ or ' glad.' viqvddevydvant, < vigvddevya. bhangurdvant, ' treacherous/ < bhangurd, i to which Rule (6) could possibly be of application. Furthermore 40 ft of this remainder violate Rule (6) and take mant. The 60 f c that take vant do so not because THE SUFFIXES MANT AND VANT IN SANSKRIT. 11 their stems end in a consonant, but because their penultimate vowel is a or a. Finally, an examination of the table in 3 will clearly demonstrate that final consonants are no factor whatever in the determination of choice between the two suffixes. 8. Investigation will clearly show one rule in operation thruout Sanskrit and Avestan literature as well as in Indo- Iranian, viz. : The choice between mant and vant depended upon the final vowel of the preceding vowel stem or the penultimate vowel of the preceding consonantal stem. This rule manifests itself thruout the entire body of Sanskrit literature as follows : 1. After a word ending in a, a or I, or in a consonant preceded by a, a or I, vant was used. 2. After a word ending in a vowel other than a, a or i, or in a consonant preceded by a vowel other than a, a or i, mant was used. First the usage with final vowels will be investigated and then that with penultimate vowels. Comparisons will be made with the Rig- and Atharva-Vedas. 9. The following table shows the distribution of mant and vant among 1985 words which end in vowels. The table is fol- lowed by lists of all the words which are counted therein. Final Vowel Number of words in which it occurs before vant Number of words in which it occurs before mant 1. a 1214 19 2. a 258 5 3. i 42 241 4. I 47 21 5. u 10 84 6. u 2 10 7. r 2 16 8. e 1 9. o 9 4 Totals 1585 400 12 THE SUFFIXES MAXT AND VANT IN SANSKRIT. Following are the examples for each vowel with each suffix : I. a with vant. angavant, aksavant, aksaravanf, agunavant, dgravant, agrima- vant, agriyarant, aghdvant, ankuravant, angulitravant, ajiravant, atattvdrthavant, atanavant, adrstavant, adhikdravant, adhvardvant, anantavant, dnapatyavant, andgatavant, andtmavant, andyatana- I'ant, dmkavant, aniiyajdvant, anuragavant, anuvdkyavant, anu- qayavant, anusvaravant, dntavant, antdriksai-ant, dnnavant, any- arthavant, anmyavant, dpatyavant, dpavant, aparavdnt, aparadha- vant, aparyaptavant, dpavlravant, apasavyavant, apdqrayavant, apasthdvant, apidhanavant, apupdvant, apratibodharant, apradd- navant, apramattavant, aplavavant, abaddhavant, abalavant. abhi- janavant, abhimdnavant, abhiriipavant, abMtavant, abhuktavant, abhujamgavant, abhydsavant, abhyucchrayai-ant, dmavant, aman- travant, amarsavant, amoghavant, ayavant, ayatnavant, arepha- vantj ariipavant, arkdoamt, arthavant, alpavant, alpavittavant, avakdqdvant, avagddhavant, avagunthanavant, ai-adydvant, avad- hdnavant, avarohavant, avdggamanavant, avdptavant, avikdravant, avijtidnavant, aviqesavant, dmryavant, avyabhicdravant, avyaya- vant, avratavant, aqanavant, aritdvant, dquklavant, aqmarant, agmavarsavant, aqvamedhdvant, dgvavant, asangavant, asamcaya- vant, asamtosavant, asahdyavant, asnehavant, dhiranyavant, dkdga- vant, dkdravant, dkopavant, dgamavant, dcdravant, dcdryavant, ddambaravant, dtapavant, dtithyavant, dtmavant, dddnavant, ddy- antavant, dditydvant, ddaravant, dpmavant, dpydnavant, dpydyana- vant, dmiksavant, dyatanavant, dydmavant, dyusyavant, dramb- hanavant, drudhavant, drogyavant, drseyavant, aldpavant) dlam- banavant, dlasyavant, dlingitavant, alokavant, dglrvdddbhidhdna- vant, dgr 'ayavant, dsanjanavant, dsecanavant, dstaranavant, dsvdda- vant, indravant, indriydvant, idhmdvant, indhanavant, isdvant, ihavant, irsyavant, uktavant, ukthavant, ujjesdvant, ujjhitavant, utthanavcmt, utsangavant, utsdhavant, utsdhagduryadhanasdhasa- vant, utsrstavant, udakdvant, udaravant, utlattavant, udayavant, udgrdhavant, udbhdsavant, udbhuta&pargavant, udvdsavant, un- mddavant, upakarandvant, upagatavant, upagudhavant, upacdra- THE SUFFIXES SIANT AND VANT IN SANSKRIT. 13 vant, upadcganavant, upadehavant, upamdnavant, upaydmavant, upardgavant, upavant, upacamavant, upasrtavant, updyavant, upoktavant, ubkayavant, urugdyavant, usnaxparqavant, usmavant, urdhvagamanavant, -usavant, usmavant, uhavant, rtavant, rtavya- vant, rtviyavant, rksavant, rnavant, rsabharant, ekavinqavant, ekdikavant, edhavant, di^varyavant, oghavant, odanavant, dutkan- thyavant, dulsukyavant, kahsavant, kakudavant, kantakavant, ka- nakavant, kandavant, karsanavant, kandaravant, kapdtatorana- vant, kapdlavant, kamalavant, karapaitravant, karanavant, kara- vant, kanravant, karnayrdhavant, kdrnavant, kdrnakavanf, kar- makrtavant, karmavant, kalakalavant, kalairavant, kalahavant, kalydnavant, kavant, kdndavant, kdmavaht, kdyavant, kdrakavant, kdranavant, kdritavant, kdrandavant, kdryavant, ktdavant, kdstha- vant, k&savant, kinavant, kidrgvydpdravant, kutuhalavant, kumdrd- vant, kumudavant, kulavant, kulcKjilavant, kuqavant, kugalavant, kusumavant, kusumbhavcwit, kulavant, krtavant, krtyavant, krpdyita- vant, keqavant, kfearavant, kotaravant, kopavant, kogavant, kostha- vant } kdutukavant, kduplnavant, krurdcdravihdravant, krdunca- vant, kledavant, klicitavant, ksatravant, ksamavant, Jcsinavant, kslrdvant, ksutavant, ksemavant, k-sdmavant, Icsditavant, khadavant, khadgavant, khadiravant, khdditavant, gajavant, gandvant, gand- hapdsdnavant, gandhavant, gandlianragddmavant, gatavant, ga- manavant, garanavant, garbJiavant, gahanavant, gdtravant, giin- aganavant, gunavant, guptavant, gulmavant, guhdgahanavant, grhavant, grhdgramavant, grhyavant, gotravant, gduravavant, grdbhanavanf, gralianavant, grdhavant, grdhyavant, grdmavant, gharmdvant, ghosavant, ghrtak^dudravant, ghrtapdtrastanavant, ghrldvant, cakravant, cakravdkavant, candagrdhavant, candavant, canasitavant, candrdvant, candrakavant, campakavant, caranavant, caritravant, carmavant, casalavant, cdtvdlavant, cdritravant, citta- vard, citravant, citrdstaranavant, citkdravant, clvaravant, cditra- vant, chaftravant, chandobhangavant, chandomavant, chinnavant, janavant, janmavant, jayavant, jalavant, javavant, jdgaritavant, jdtavant, jdmbavant, jdlavant, jitavant, jlrnavant, jivavant, jwana- vant, jusdnavant, jndnavant, tankavant, tamkdravant, taksavant, taddgavant, tattvavant, tatpravaravant, tatavant, tanutravant, tard- mavant, talatravant, taldngulitravant, tddrgrupavant, tdvadvirya- 14 THE SUFFIXES MAXT AND VAXT IX SANSKRIT. vcint, tigmdvant, tilaka&ant, tlrtliarant, tu/ndavant, tuviruvant, tuna- rant, tr/mrcirant, tiilavant, trnai'cmt, tejavaitt, tokdoant } toy av ant, tyaktavant, trayidhamavant, trikutavant, trajgunyavant, trydlikhitd- vant, tvaksdravyavaharavant, trayodagadvlpavant, traguttarasanya- vant, dahstrakaralavant, dandaparusyavant, dandavant, ddtravant, dantavant, darbhapinjilhtcant, darqitavant, dagapadmavant, dana- vant, dayadavant, ddfainyavant, ddhavant, d'ipavant, duhsatti-a- vant, duhkhagokavant, durdtmavant, dutavant, drstavant, drstanta- vant, devayajanavantj devdvant, devaravant, dehavant, do$o.vant, dohadavantj dduhitravant, dyumndvant, drapsdvant, dravinavant, dravyavant, drumavant, dvdravant, dviguklavant, dvihiitavant, dvitiyavant, dvipavant, dhanamadavant, dhanavant, dharmadh- vajcivant, dharmavant, dharmavirodhavant, dharmdbhijanavant, dhdrmcivant, dhdnyadhanavant, dhdnyavant, dhdmavant, dhisnya- vant } dhumavant, dhrtavant, dhdiryavant, dhydiiavant, dhvajavant, natiprcicurapadyarant, ndikadravyoccayarant, naktavant, naga- vant, namaskdravanty nayavant, nayanavant, narmavant, navaydu- vanavant, nastavant, ndgayajnopavitavant, ndgavant, ndthavant, nddavant, ndndraindkarcivant, ndrnavant, ndrakapdlakundalavant, nasikavant, niraparddhai'ant, nirdtmavantj nirdyavyayavant, nir- maldtmavant, nirvikdravant, nirvicesarant, nihsecanavant, nihsd- ranavant, nigciditavant, nigdlcivant, nigirnavant, nitambavant, niddnavant, nidhdnavant, ninrttavant, nipdnavant, nimbavant, niyamavant, nirdhavavant, niruktavant, niruddhavant, nirbhart- sitavajit, nirvedavant, nivegavant, nisadavant, nisiddhavant, nispesa- vant, nisrstavant, nistyaktavant, nihrddavant, n'iddvant, riilacolaka- vant, nilavant, nupuravant, Jiydyavant, paksapucchdvant, paksd- vant, pankajavant, pankavant, pakvavant, pancagavydpdnarant, pancada$dvant, pailcainavant, panyavant, patdkocchrdyamnt, pa- tidharmavant, padmavant, padmasdugandhikavant, paravant, pa- rdyanavant, paramapaddtmavant, pardkramavant, pardgavant, parigfhyctoant, pariyrahavant, paricayavant, parindmarant, pari- nahavant, paritosavant, pariputanavant, paripotavant, paripiir- nasahaxracandravant, paribarhavant, paribodhavant, panvdha- vant, parirarai'anf, parivestitavant, parigesavant, pariJidravant, parnai-ant, paryastavant, parydptavant, parydlocitavant, parva- vant, palitavant, pavamdnavant, pavitravant, pdrlravant, pav'ird- THE SUFFIXES MANT AND VANT IN SANSKRIT. 15 vant. pdnavant, pdkavant, pdthavant, pdnigrdhavant, pdtheyavant, pddavant, pddukavant, pdmavant, pdrijdtavant, pdvakavant, pdqa- vant, pieindavant, picchavant, pitrmdtrgurugu^r'usddhydnavant, pit- tavant, pidhdnavant, pibavant, pUavant, punsavant, pucchavant, pundarlkdvant, punyavant, putrdvant, punargarbhavant, pura- staddhomavant, puronuvdkyavcmt, piirlsavant, purukcirakavant, purusavant, puspadantavant, puspaphalavant, puspavant, purva- vant, prthupajarant, prthusattvavcint, pesanavant, pdutrikeyavant, prakdgavant, prakdganavant, prakdravant, prakarsavant, prakra- mabhangavant, prakledavant, praJcsiptavant, praJcsveditavant, pra- grahai-ant, pragrdhavant, prajtinanavant, pranatavant, pranatdt- mavant, pranayavant, pratdnavant, pratikulavant, pratigabdavant, praticdkhavant, pratyavamarsavant, pratipannavant, pratibandha- vant, pratibodhavant, pratibhdnavant, pratibhdvavant, pratiglna- vant, pratmddhavant, pratihdravant, pratndvant, prdtikavant, pratyabhijndtavant, pratyabhiyuktavant, pratyavamarsavant, pra- dandavant, prattavant, praddnavant, pradeqavant, pradharsitavant, pranrttavant, pranrtyavard, praphuUanagavant, prabalavant, pra- bodhavant, prabodhitavant, prabhdvavant, prabliiitadhanadhdnya- vant, pramattavant, pramddavant, pramdnavant, prameditavant, pramuditavant, praydjdvant, prayatnavant, prayatdtmavant, prayo- janavant, prarohavant, praldpavant, proktavant, pravanavant, prurant, prositavant, pravdlavant, pravibhdgavant, pravaravant } pravargydvant, pravrttavant, praqdkhavant, pragrayavant, pra- sungavant, prasddavant, prasecanavant, prasthavant, prasthitavant, prasveditavant, praJiitamgamavant, praharanavant, praharsavant } prdgalbhyavant, prdkgrngavant, prdggamanavant, prdnaddvant, prdnavant, prdndrthavant, prdptavant, prdrabdhavant, prdrthita- vant, priydvant, premavant, presitavant, plaksavant, plavavant, plutavant, phanavant, phaladantavant, phalapuspavant, phala- tnulavant, phdlavant, phalastanavant, phullapadmotpalavant, phul- lavant, phutkdravant, phenavant, bakavant, baddhagodhdngulitra- vant, baddhapankavant, barhavant, bdlavant, bahudugdhavant, bahupuspaprabdlavant, bdnavant, bilavant, bisavant, brjavant, buddlidvant, bodhavant, budhndvant, brahmacaryavant, brahma- vant, brahmavisnvarkavant, brdhmanavdnt, brdhmanavant, bhdga- vant, bhangavant, bhadravant, bhdvant, bhdnavant, blidgyavant, 16 THE SUFFIXES Jf.4.vr AND VANT IN SANSKRIT. bhdjanavant, bhdravant, bhdvavant, bharitavant, bhdskarai-ant, bhinnavant, bhlmagrdhavant, bhuktavant, bhutakaranavant, bhuta- vant, bhogavdnt, bhogavant, bhrdtrsamghdtavant, bhratri-yavant, makarandavant, makhdrant, maghdvant, mangalavant, maniratna- vant, matsardvant, madanavant, matdvant, mananavant, mantra- vant, manddravant, manmathavant, mayukhavant, maruddvlpa- vant, mdrtavant, marydddparvatavant, malavant, malayavant, mahdtmavant, mahdbhogavant, mahabhogavant, maharatnavant, mahdrasavant, mahdrthavant, maJidvegavant, mahavratdvant, ma- hdifitarant, mahagetavctnt, mahdhimavant, inahdlicmavant, mahoc- chrdyavant, indnsabudbudavant, mdnsdvant, mddravant, mdna-vant, mayavant, mdydbalavant, mdrarant, mdlyavant, mdsikdrthavant, mitravant, mitrdvarunavant, minavant, mukhavant, munjakecavant, muiijavant, mudgavani, muskavant, mugdhavant, mujavant, malad- vdravant, mulavant, mrgavant, mrgdnkavant, mrstavant, mrndla- vant, mrduromavant, meghavant, mehavant, moJcsavant, mohavant, yajndvant, yojnopavttavant, yatnavant, yantravant, yamavant, yamasattvavantj yavagodhumavant, ydvisthavant, ydnavant, ydma- vant, ydjy d,vant, ydbhavant, ydmyasattvavant, yavadriryavant, yuktdvant, yuddhavant, yupavant, yupakavant, yogaksemavaid, yogarddhirupavant, yogavant, yduvanavant, raksitavant, ranga- vant, rdtnavant, rdthavcmt, ratavant, rayavant, rdsavant, rasiia- vant, rdgagrahavant, rdgavant, rdjavant, rdjanydvant, rdjyavant, rephavant, rukmavant, rucitavant, rudrdvant, rupayduvanavant, ruparasagandhasparqavant, ruparasasfiarqavant, rupavant, rupa- sdubhdgyavant, rupasparqavant, rocandvant, romavant, rohita- vant, laksanavant, laghudvdravant, latdgahanavant, latdvalaya- vant, labdhavant, Idbhavant, Idlavant, laldmavant, lavavant, Idsa- vant, Idvanyavant, lingavant, liptavant, lobhitavant, lokavant, lomavant, losiavant, lohacarmavant, lohavant, lohitavant, Idulya- vant, vaneanavant, vatavant, vatsavant, vayunavant, varuthavant, varnavant, varnd$ramavant, varnitavant, vartitavant, varmavant, varsmavant, valavant, valkavant, valkalavant, vasanarant, vastra- vant, vdtagulmavant, vatavant, vdgghastavanf, rajta-ant, r'mavant, vddavant, vdditravant, vdravant, viqdkhavant, vikdravant, rikalpa- vant, vigrahavant, vighnavant, vicaksanacanasitai-ant, vicoksana- vant, mcdravant, injayavant, vijitavant, vijndnavant, litathdbhi- THE SUFFIXES MANT AND VANT IN SANSKRIT. 17 nivegavant, vitdnavant, vitarkavant, vidydbhimdnavant, vidydved- avratavant, viUavant, viddhavant, vidyunmandalavant, vinayavant, vmodavant, vindhyavant, vipandpanavant, vipuldrthabhogavant, vipratisaravant, vibhdgavant, vibhavavant, vibhusanarant, vibhra- mavant, vimalakiriiahdravant, vimarqavant, vircigavant, virodha- vant, vildfsavant, vivdsanavant, vivekavant, vivrtavant, viqesavant, vigesanavant, inqrambhavant, viyutavant, vigvddevavant, vi$varu- pavant, vigvavant, visdvant, visddavant, visayavant, visdnavant, visvaggamanavant, viskambhavant, visrstavant, vistlrnavant, vis- mayai-ant, vismayavisddavant, vismrtavant, vihitavant, vihdravant, vitavant, vmdpanavatunavant, mrdvant, virinavant, mrydvant, vlryasattvavantj vrtrcihavant, vrJcsavant, vrjinavant, vrttavant, vrttasvddliydyavant, vrsavant, vrstavant, vegavant, vejo.navant y vetntvant, vedavant, ve/j.evov (RV. I, 168, 7). This word has long been a crux. The Petersburg Lexicon, followed by Monier- Williams, makes no comment on pipisvant except to say that it is derived from the \/pi ( =p^). Grassmann adds that it is the desiderative of this root. Benfey (Grammati/c, p. 170) says that pipisvant is perhaps an anomalous perfect active participle. Whitney ( Gram- mar, 1223 g. ) classes pipisvant among words with which vant has the aspect of forming primary derivatives, but adds a (?) after the word. Sayana connects the word with \/pis, ' to crush.' pipisvant is the only derivative with vant which may be said with any great degree of certainty to be primary. 34 THE SUFFIXES MANT AND VANT IN SANSKRIT. 1. Penultimate a with vant. brdhmanvant, aksanvdnt, girsanvdnt, datvdnt, padvdnt, asthan- vdnt, udanvdnt, dtmanvdnt, dgmanvant, gcanvant, dsanvdnt, antdr- vantj svarvant, qdqvant, bhdrgasvant, vydcasvant, vdrcasvant, tejas- vant, ojasvant, urjasvant, dnupadasvant, m&dasvant, nasvdnt, dnas- vant, tdpasvant, ndbhasvant, tamasvant, ndmasvant, pdyasvant, jyayasvant, pdrasvant, sdrasvant, hdrasvant, dvasvant, vwasvant, vivdsvant, ydqasvant, mdhasvant, sdhasvant. 3. Penultimate i with mant. jyotismant, havismant. 5. Penultimate u with vant. marvtvant. Penultimate u with mant. gartitmant, dyusmant, cdksusmant. Thus martitvant, an obviously hieratic archaism, is the only violation of the rules for either suffix among the 45 possessives in Atharva-Veda whose stems end in a consonant. 17. From the above comparisons the conclusion will be reached that, while the distinction of usage between mant and vant was to a degree weakened in very early times, and while this confusion between the suffixes increased and spread by a kind of see-saw analogy, still the Hindus never to a very appreciable extent lost from their linguistic consciousness the feeling that vant belonged especially to the a-vowels and mant to the u- vowels, even when those vowels were protected by consonants. The allegiance between the suffixes and the other vowels was not so constant. In earliest times vant claimed all except the li-vowels. Later mant took all non-a vowels from vant except I until finally there was a strong tendency to assign a and a to vant and all other vowels to mant. The historical process may be sketched thus : 1. Indo-Iranian and Avestan : final or penultimate u, u took mant ; the other vowels, vant. 2. Vedic : final or penultimate a, a, I took vant ; the other vowels, mant. THE SUFFIXES MANT AND VANT IN SANSKEIT. 35 3. Classical Sanskrit tendency to use vant with a or a and mant with the other vowels. We find in the Veda that i and I, for example, show a decidedly stronger leaning toward vant than in the literature as a whole. The Veda apparently represents a period of transition from the usage of Indo-Iranian, as preserved by the Avestan, to the usage of classical Sanskrit. The Indo-Iranian rule doubtless required mant after final or penultimate u, u, and vant after all other vowels, but beginning even before the Vedas there is a tendency toward restricting vant to a and d and leaving the remaining vowels to mant. At this point the possibility suggests itself that vant was the original suffix and that mant was used in its stead after u or u for the purpose of dissimilation. The great frequency of occurrence of the a-vowels as compared with that of the other vowels may easily have tended toward the custom of limiting the use of vant to them. According to Whitney's table (Grammar, 75) a and d are nearly twice as common in Sanskrit as all other vowels and diphthongs together, or, more accurately, as 27.97 to 15.55. The transition from 1 to 3 (above) may, perhaps, be best illustrated by the statistics for i and I in the Rig- and Atharva- Vedas as compared on the one side with Avestan and on the other with Sanskrit (exclusive of the two Vedas). Avestan EV. AV. Sanskrit Final i-mant 1 19 5 217 " l- " 1 6 4 11 Penultimate i-mant 7 2 7 (( i_ a Totals 2 32 11 235 Final i-vant 20 10 3 29 I- 19 6 22 Penultimate i-vant 9 3 9 " l- " 1 1 1 Totals 30 33 9 61 36 THE SUFFIXES MANX AND VANT IN SANSKRIT. It is evident that i and I have gradually gone over from vant to mant. In the Avesta mant is used with i, I only twice out of 32 examples, or only a little more than 6 ft, whereas in classical Sanskrit mant is used 235 times out of a total of 296, or nearly 80%. In RV. and AV., however, mant and vant are about evenly divided after i f I (mant 43, vant 42). As might be expected, RV. stands closer to the Indo-Iranian usage than does AV. 1 18. Given two suffixes so similar in form and signification existing side by side thruout the entire written history of a lan- guage, it will cause no surprise to find that each one has frequently gone over into the other's territory and appropriated its functions. There is indeed occasion for surprise that one of the suffixes did not at some period of the language absorb the other, or at least that one of them did not pass away from its original sphere of use into some other field of function or meaning not already so thoroly preempted. The immediate occasions for the encroachment of the suffixes upon each other are in a general way not difficult to determine. Some of them are, broadly : 1. The existence of two forms of the same word which require 1 Wackernagel ( KZ. xiJii, 277) makes a rule for the Avesta and for Indo- Iranian that mant is used with stems in M, us, and ao, and vant with all other stems. For the Avesta alone this rule holds very well. But, especially in the light of the Veda, it seems to me to be evident that, both in Indo-Iranian and Avestan, only the last vowel in the stem is the determinative factor as to which suffix follows it. Wackernagel makes a point of the fact that mant follows us, but vant follows is, ah, &c. But it must be noticed that in Avestan, whatever the final consonant may be, vant is always (58 cases) used with a stem in which the penulti- mate vowel is a non-u vowel. The Veda looks in the same direction. Of the 9 examples of penultimate u in RV. 2 have vant and 7 have mant. Of these 7 only 2 are stems ending in -us. Beside mldhusmant, caksusmant we find virukmant, garutmant, vihiitmant, kakudmant, vidyunmant. AV. has garutmant as well as ayuymant and cdkxusmant. I cannot avoid the conclusion that the final consonant of the stem has no influence on the form of the suffix but that the preceding vowel determines which suffix is to be used, regardless of the consonant. Avestan ao is Indo-Iranian au. With these modifications Wackernagel' s rule for Avestan and Indo-Iranian is then equivalent to mine above that final or penultimate u, u took mant and the remaining vowels vant. Cf. 21. THE SUFFIXES MANX AND VANT IN SANSKRIT. 37 different suffixes. The more common one of the two will attract the suffix of the other into agreement with its own suffix. Examples are : f&jasvant, tejovant. dhrtimant, dhrtamant. jdmbavant, jdmbuvant. 2. Assonance of a word or line, alliteration or dissimilation may demand one suffix or the other in violation of the gram- matical rules. Possible examples are : kdnvamant. The mant may be due, as Schulze suggests (KZ. xxxix, 612), to disinclination against two successive ra's, but the weight of evidence of other words is decidedly against such an explanation (cf. under ydvamant, 5, below). The word is a aira^ \ey6fjbevov, and there is a bare possibility that mant was used for the sake of m-alliteration. The text is RV. vm, 2, 22 : a in sinca kdnvamantam nd ghd vidma cavasdndt yacdstaram catdmiiteh. marutvant. vant was perhaps used for the sake of dissimilation from the initial ra ; cf. garutmant. (Cf. also vdnmant, vibhava- mant, mramant, etc.) 3. Wrong readings. Many exceptions to the rules exist in only a single example, or in unauthoritative literature, and are frequently due to wrong or doubtful readings of the text. Some such examples are : riipamati, name of a queen. Found only in a Nepal inscrip- tion (cf. Indian Antiquary, ix, 188. 9loka 7). rdjamatl was the name of another wife of the same king as jRupamatl, above. Cf. the same inscription, Indian Antiquary, ix, 188, 189. This word is quoted neither in the Petersburg Lexicon nor in Momer-Williams' Dictionary. cubhramati, name of a river. Probably a wrong reading for cabhravati, which is also the name of a river. Cf. Lassen's Indische Alterthumskunde, n, 802. 38 THE SUFFIXES MANT AND VANT IN SANSKRIT. stanayitnuvant is a wrong reading for stanayitnumant in the Calcutta edition of the Mahabharata, 6, 731. tarasmant is in the Harivanca (i, 7, 87) a varia lectio for the familiar tdrasvant. It does not appear elsewhere. dasmdnt, a problematic a-Traf \e^6^evov (RV. I, 74, 4). akdramant is found only in the Raja-Taranginl. dkdravant occurs in the Katha-sarit-sagara and elsewhere. dhrtamati, name of a river, makes its only appearance in the Visnu-Pura a, 184. The Petersburg Lexicon calls it a "falsche Form fur dhrtimafi." The latter is the name of a river in the Mahabharata. suryamafi occurs only as the name of a princess in the Raja- Tarangini (7, 152 and 179). suryavant is common from the Veda on, and suryavafi is the name of a princess in the Vasava- datta. prdnamdnt appears only in the Maitrayani-Samhita (1, 6, 12), whereas prdnavant, the normal form, is found in Katyayana- Qrauta-Sutra, Qakuntala, Harivaii9a, and Su9ruta-Samhita. lalamafi is quotable only for a Nepal inscription where it is the name of a queen. Cf. Indian Antiquary, ix, 185 (Petersburg Lexicon wrongly has 158), yloka 9. This word is, doubtless, like Idlavant derived from laid. The latter possessive appears in Su9ruta-Samhita. 4. The tendency to use vant instead of mant in the meaning " containing the root or word " (cf. Chap, in, 3). The examples of vant in this sense with z-stems may be con- sidered violations of the existing phonetic rules instead of sur- vivals from the time when i properly took vant. The very meaning of the words indicates that they belong to the post-Vedic period when mant had replaced vant after i. Some examples are : hdrivant, ' containing the word hari ' ; cf. harimant. trivant, 'containing the word tri*. bhidvant, ' containing the y bhid '. pretivant, ' containing the word preti '. vivant, ' containing the word vi\ THE SUFFIXES MANT AND VANT IN SANSKRIT. 39 ehivant, 'containing the word ehi'. abhivant, ' containing the word abhi'. vahnivant, ' containing the word vahni' cf. the normal form with normal meaning, vahnimant, ' containing fire '. jaghnivant, l containing an intensive form of j/ han '. namovrktivant, l containing the word namovrkti '. parivant, ' containing the word pari '. etivant, ' containing a form of the j/i (etiy. 5. The whole broad sweep of analogy with all its possibilities of chance resemblance and remote suggestion, so easy in its operation and so difficult in the tracing thereof. A couple examples follow : ydvamant. Schulze (KZ. xxxix, 612) thinks that mant was used instead of vant in order to avoid the cacophony of vava. The same may be said of kdnvamant, vibhavamant and alpasva- mant, but these exhaust the list. On the other hand the number of words in which vavant is found is large. Besides devdvant, which Schulze mentions, are dgvavant (also in RV.), gduravavant, javavant, jlvavant, tattvavant, pakvavant, parvavant, purvavant, plavavant, bhavavant, lavavant, vibhavavant, viqvavant, sattvavant, sarvavant, svatvavant, svdvant, havavant. In such cases, Schulze thinks, analogy has been stronger than euphony. Or, he enquires, could the cacophony have been considered more objectionable in unaccented vava than in accented vdva? Thus ydvamant, Jcdnva- mant y but devdvant. Against this suggestion stand dgvavant, vigvddevavant, vigvavant. ydvamant, ' containing barley/ ' abundance of grain,' very probably takes mant thru the influence of other words expressing possession of wealth, such as gomant, paqumdnt, vdsumant, rdyi- mant, pustimdnt, all of which are good Rig-Vedic words, ydva- mant is found four times in Rig-Veda and in three of them is in immediate connection with such words. The phrase gomad ydva- mat is found twice and ydvamad gomat once. The passages are : RV. vin, 93, 3. sd na indrah givdh sdkhagvavad gomad ydvamat urudhdreva dohate 40 THE SUFFIXES MANT AND VANT IN SANSKRIT. RV. IX, 69, 8. a nah pavasva vdsumad dhiranyavad dqvdvad gomad ydvamat suviryam yuydm hi soma pitdro mdma sthdna divo murdhdnah prdsthitd vayaskftah RV. x, 42, 7. drdc chdtrum dpa bddhasva durdm ugro ydh gdmbah puruhuta tena asme dhehi ydvamad gomad indra krdhi dhiyam jaritr& vdjaratndm The fourth passage is RV. x, 131, 2. kuvid angd ydvamanto ydvam aid ydthd ddnty anupurvdm viyuya ihehdisdm krnuhi bhojandni y& barhiso ndmovrJdim ndjagmiih Here the ydvam has doubtless operated with regressive verbal assimilation upon the preceding ydvamantas resulting in the form with mant instead of a form with vant. visnuvant occurs only once in Rig- Veda (vin, 35, 14). In this passage it stands between dngirasvant and maridvant. The former adds its suffix regularly and is a good Vedic word. The latter, while it adds its suffix irregularly, is a very widely used word, appearing in Rig- Veda alone 40 times. In all three words vant expresses accompaniment, and it was probably the influence of the two other possessives which caused the word visnu to take the suffix vant in this one passage. 1 The text reads dngirasvantd utd vfanuvantd mariitvantd jarittir gachatho hdvam sayosasd usdsd suryena cdditydir 19. In Sanskrit 55 words are at one time or another found 1 Wackernagel ( KZ. XLIII, 280) makes a similar suggestion. THE SUFFIXES MANX AND VANT IN SANSKRIT. 41 with both mant and vant. 15 of these appear either in RV. or AV. with one suffix or the other and 4 appear with both suffixes. The remaining 40 instances belong almost entirely to classical Sanskrit. The Vedic examples are : arcimdnt (RV.) beside arcivdnt (RV.) agnimdnt (AV.) " agnivdnt (RV.) mahismant " mdhisvant (RV.) 1 marutmant " marutvant (RV., AV.) patimaM " pdtivatl (RV.) janimant " jdnivant (RV.) haritmant " hdritvant (RV.) 2 dhimant " dhivant (RV.) visnumant " visnuvant (RV.) cf. 18-5. tdvislmant (RV.) " tdvisivant (RV.) harimant " hdrivant (RV., AV.) gafdimant " gdktivant (RV.) rayimdnt (RV.) " rayivdnt (RV.) 3 1 This word is a tfira \ey6/j.vov (RV. vii, 68, 5). Its form is mafasrantarn. Grassmann makes it a possessive in zrani from mafa's ( -j/ I. maA) . Monier-Wil- liams (Diet., s. v. ) makes the word an accusative of mdhisvanta which he ques- tioningly derives from mahis (j/ I. mah). Monier- Williams (loc. cit.) follows the Petersburg Lexicon in considering mahisa the base of mahismant. I follow Whitney ( 1233e., 1235f.) in deriving both words from mahi with an inserted s (as in indrasvant). Cf. Chap. II, 9. a haritvant, 'gold-colored,' (a <5bra| \ey6fjievov, RV. x, 112, 3) has heretofore been derived from harit, id. But if that were the case we would expect the accent to be thrown forward upon the suffix, and we should have *haritvdnt (cf. Whit- ney, Grammar, 1233a. ). It is much more likely that hdritvant is derived from hdrita, with loss of final stem-vowel (cf. Chap, n, 6). hdritais common in RV. and has the same meaning as harit. The accents of hdrita and hdritvant agree perfectly. 3 revdnt (Avestan raevant), according to the assumption of the Petersburg Lexicon, is a contraction of rayivdnt. With this view agree Grassmann, Bar- tholomae and Monier- Williams. It is, nevertheless, open to question. Rig- Veda uses revdnt 58 times, rayivdnt four times and the elsewhere more common form rayimdnt twice. It is worthy of mention that three of these four instances of rayivdnl occur in the sixth book (vi, 6, 7; VI, 44, 1 ; VI, 68, 5). The other case is I, 129, 7. For Atharva-Veda revdnt is quotable 9 times, but rayivdnt and rayimdnt do not appear. 42 THE SUFFIXES MANT AND VANT IN SANSKRIT. nrmant besides nrvdnt (KV., AV.) rgmant " fkvant 20. Following are the other words that are at any time found with both mant and vant, making 55 such words in all. dkdra bandhu uddn (cf. Chap, n, 9f.) bdhu kakud brhaspati ksudh 1 bhd jaya mani 2 tdras mati darl mukti dhrta yagas 3 pitf ragmi prpii rdjan prdna rupa vasu Idla vahni sdmrddhi saptarsi sarpis 4 vac 5 sddhu vidyut 6 sdbhra vibhava surya mra stanayitnu ?ubhra (cf. Chap. 11, 4) sruc 7 21. The operation of the rules governing the use of the suffixes after final consonants is well illustrated by the figures for s and s in the table of 3. It will be seen that after final s mant is nearly always used (28 out of 31 cases), and that after final s vant is customary (57 out of 59 cases). The explanation of 1 kqudvant, ksunmant. 'manimant, manlrati (cf. Chap, n, 4). s ydfasvant, ycqovant, yafomali. 4 sdrpismant, sarpkvant. 5 vagvant, vdnmant. 6 vidyutvant, vidyudvant, vidytinmant. 1 srdugmata, nom. prop., srugvant. THE SUFFIXES MANT AND VANT IN SANSKRIT. 43 course is that both finals were originally a, and that in the first case the s was changed to s by the influence of the "alterant" (cf. Whitney, 180c.) vowels (usually i or it) which preceded it and which, according to the rule, demanded the mant suffix. In the latter case the s was retained because preceded by a or a, and vant was employed for the same reason (cf. Whitney, Gram/mar, 180 and 184b.). CHAPTER II. SAMDHI. 1. As Whitney points out (Grammar, Hid., 1233L, 1235f.), while possessive mant and vant are usually added to noun-stems according to the rules of internal combination, they very frequently, especially in the later language, follow the rules of external combination. This latter fact is of course due to the native conception of these suffixes as more or less independent entities of fixed and well-known value, and consequently of their right to be treated somewhat after the manner of final elements of compounds. 2. The usual rules of internal combination are followed. Thus: aksanvant < aksan, pusanvant < pusan, brdhmanvant < brahman, dyusmant < dyus, cdksusmant < caksus, kdmvant < kam, qdmvant < gam (Qatapatha-Brahmana and Katyayana-Qrauta- Sutra have kdmvant and qdmvant. Cf. Whit- ney, Gram., 212b.), kimvant < 1dm. 3. In the following words the suffixes are added to the stem 44 THE SUFFIXES MANT AND VANT IN SANSKRIT. according to the rules of external combination. Some of these cases are found even in the Veda, but they become much more numerous in the later literature. angirovant (cf. dngirasvant}, cetomant < cetas, chandovatl (cf. clidndasvanf), tapovant (cf. tdpasvant, with different meaning), tamovant (cf. tdmasvant, with same meaning), tejovant (cf. t&jasvant), manovatl < manas, medovafi (cf. medasvati), yagomatl (cf. next and yd$asvan(), yagovant (cf. preceding and ydgasvanf), rodhovatl (cf. r6dhasvant\ fkvant < re (cf. next), rgmant < re, (cf. preceding), rukmant < ntc, runmant < id., vdgvant < vac, vanmatl < id., srugvant < snc, *srugmant < id. (sraugmata, nom. prop.), tvagvant < toac, purorunmant < puroruc, svarddvant, havyavddvant, vimrdvant < vimrdh, virukmant < vzVwc, bhisagvant < bhisaj, sragvant < srq/, dvasvadvant < avasvat (avasvanfy, jagadvant < jagat, jdnadvant < jdnat, dhdyadvant < dhayati (cf. 6), dhdrdyadvant < dhdrayat, pfsadvant < pfsat, THE SUFFIXES MANT AND VANT IN SANSKRIT. 45 brhddvant < brhat, bhuvadvant < bhuvat, mahadvant < mahat, sadvant < sat, vidyudvant (" wrong reading " Mon.-Wms., cf. vidyutvanfy, vidytinmant (" with irregular assimilation of t " Whitney, Gram., 1235b.), parisrunmant < parisrut, kaktinmant (cf. kaktidmant), ksudvant < ksudh, ksunmant < id., samidvant < samidh, tristubvant < tristubh, kakummatl (cf. kakubhvanfy, gudalinmant < gudalih. In KV. x, 106, 5 Grassraann boldly substitutes a form of $dtavant for " das unverstandliche qatapanta " of the text. This gdtavant, if it existed, would then be a derivative of gdta, a vrddhi form of qata which is found only in composition, and would be equivalent to gatdvant. Whitney (Gram., 1233i.) includes agwvant in his list of examples of external combination. He must then consider the word derived from a form ending in s (cf. Whitney, 174) ; probably not however from agis, ( prayer ', ' blessing ', which appears in composition as aglr, for in 1233c. he translates the possessive 'mixed with milk'. But dqirvant is beyond doubt derived from a$ir, ) and translated ' helping a hundred-fold.' The discussion centers round EV. vi, 47, 9. THE SUFFIXES MANT AND VAN! IN SANSKKIT. 51 vetrdvatl (cf. vetravant'), gdktivant (cf. gaktimanfy, gardvaft (cf. garavanfy, qundvant (cf. gundvanfy, qrutdvatl (cf. grutavanf), sdptwant, sarvdvant (cf. sarvavant), salildvatl (cf. salilavanf), sdhdvant, sahdvant (cf. 7, end), sidhrakdvant, sutdvant (cf. sutavanfy, sukhdvant (cf. sukhavant), srkdvant (cf. srkavanf), srgdvant (cf. srgavanf), somdvatl (cf. somavant and Panini, vi, 3, 131), sphurjdvant, svddhitlvant, hansdvatl (cf. hansavant), hanumant (cf. hanumanf), hitdvant (cf. hitavant and Chap, in, 6), himdvati (cf. himavanf), hirlmant < Mri = Adri (cf. harimant and hdrivanfy, hfslvant < Arst = harsa, hlddikdvant, hlddukdvant, bhogdvati (cf. bhogavdnt], vlrandvatl (cf. Panini, vi, 3, 119), mamvatl (cf. manimant and Panini, vm, 2, 11), ahwatl (cf. ahimant and Panini, vm, 2, 11). To these examples must be added those formed in very early times from pronomimal stems (cf. Chap, in, 16 and Whitney, Gram., 457, 494, 517, 1233f.). They are: mdvanty tvdvant, 52 THE SUFFIXES MANT AND VANT IN SANSKRIT. yuvavant, tavant, etdvant, ydvant, want, kivant. 5. Systole. In a number of well-authenticated words the final long stem-vowel of the primitive is shortened before the possessive suffix. Whitney calls this shortening a " special irregularity." It appears however in a score of words. In most cases it is either a wrong reading or the vowel has been shortened metri causa. Or the shortening may be due in a case or two to analogy with the original stem of the noun. Thus irsyavant, 1 envious ', seems to be from irsyd, ' envy '. The vowel was perhaps shortened thru the influence of irsya, mfn., ' envious'. Examples are : anuvdkyavant, dmiksavant, pddukavant, puronuvdkydvant, pratiqdkhavant, gdurimant, prdnaddvant (cf. prdnadavanfy, tuvlrdvant, 1 mdydvant (cf. mdyavanfy, Idlavant (perhaps also Idlamafi), garydnavant (if for qaryanavant), sarasvativant < sarasvatl (cf. sdrasvaftvanf), 2 sdukhavati (wrong reading for sukhdvat'i), susumdnt, 3 1 According to Grassmann <^ turn -f- irdvant = irdvant ( <^ ira) . Petersburg Lex- icon considers the word a contraction of *tuvlravavant. 2 Another example of double possessive suffix is dvasvadmnt. 8 This word appears only once in KV. (x, 3, 1). Grassmann derives it from su6. Whitney ( 1235d.) says it appears to be primary. THE SUFFIXES MANX AND VANT IN SANSKRIT. 53 vattimant < valfa, ydjydvant, Irsyavant < wsyd, (fipravant < gipra. It is possible that in several of the above examples the short- ening was due to word-rhythm, but the evidence is not conclusive. In the actual forms of several of the above words in the texts the shortened vowel is either preceded or followed, or both preceded and followed, by a syllable containing a long vowel. Thus tuw- rdvant appears only twice (RV. x, 64, 4 and x, 64, 16), both times in the form tuvlravdn. puronuvdkydvant shows the forms puronuvdkyavantam (Qatapatha-Brahmana, 9, 3, 1, 16) and puro- nuvdkyavantas (Katyayana-Qrauta-Sutra, i ? 2, 6). pratiqdkhavant appears in the Mahabharata (14, 955) as pratigdkhavdn. prdna- ddvdn is a manuscript reading for prdnaddvdn in Atharva-Veda, 4, 35, 5. The atapatha-Brahmapa (13, 5, 4, 12) has mdyd- vattara (< mdydvant, which the Petersburg Lexicon says is metri causa for mdydvanfy. The only form of susumdnt is susuman (RV. x, 3, 1). The only form of qipravant is gipravdn (RV. VI, 17, 2). ydjy&vant occurs only in the atapatha-Brahmana. Its forms are: ydjyavdn (4, 1, 1, 26), ydjyavantam (9, 3, 1, 16), and ydjyavat (11, 2, 1, 6). anuvdkyavant has the form anuvdkyavdn in the Qatapatha-Brahmana. Here the shortened vowel (ya) is preceded and followed by a syllable with a long vowel (ata- patha-Brahmana, 4, 1, 1, 26). In Ayvalayana-Qrauta-Sutra, however, the possessive appears with a long yd (1, 5, 30), but that syllable is followed by one with a short vowel. In Idlavant, Idlamatl, irsyavant and qarydnavant, the shortened vowel is pre- ceded by a syllable with a long vowel. The following words oppose the theory of word-rhythm as applied to this shortening, sarasvativant appears only in Katya- yana-Qrauta-Sutra (9 ? \ f 20). It there takes the form sarasvati- vati. vattimant appears in the Gita-govinda as vallimat. The only form of pddukavant is pddukavatim (Hemadri's Caturvarga- Cintamani, i, 711, 10). dmiksavant appears in the genitive dmikmvatdm in the Taittiriya-Brahmana (2, 7, 16, 4). 54 THE SUFFIXES JW.vr AND VAST IN SANSKRIT. 6. Loss of final vowel. There are a half-dozen or more instances of the loss of a final vowel before the possessive suffixes. Whitney mentions only one of the cases and dismisses it as a " special irregularity." Panini (iv, 2, 87) notices three of the examples, but attempts to formulate no rule. The examples are : nadvant < nada (cf. Panini, iv, 2, 87). vdnanvant, probably for *vananavant (cf. vandnd and Whit- ney, 1233e. and Grassmann, s. v.). vdasvant, probably < vetasa (cf. Panini, iv, 2, 87). qaradvatl, probably < garadd, with loss of the a thru the influence of garddvant. hiranvant, nom. prop., probably < hiranya (cf. hiranyavanf) . hiran-vant and hiran-mdya may seem to be sufficient evidence for the assumption of an original stem kiran. But it is much more likely that hiranmdya is for hiranyamdya, which occurs in the Qatapatha-Brahmana. The analogy of hiranmdya may have caused the loss of the ya in hiranvant. kumudvant, probably < kumuda (cf. kumudavafi and Panini, iv, 2, 87). mahismant (if < mahisa, cf. Chap. I, 19, note). dhdyadvant < dhayati. hdritvant < hdrita (cf. Chap. I, 19, note). kaksivant, from kaksyavant, according to the Kacika Vrtti, followed by Grassmann and Monier-Williams. This derivation is, however, by no means absolutely convincing. 7. Loss of final consonant. There are nearly 50 cases in which the n of final an or in of the stem is lost before the pos- sessive suffix. This loss of n is the usual one in the middle forms of the stem or before a case-ending or suffix beginning with a consonant (cf. Whitney, Gram., 256, 421, 439). Whitney's general statement in regard to vant ( 1233b.) that nouns in an more often retain the ?i, while true for the older language, does not hold good for the entire literature, since the n is lost as often as it is retained. The retention of the n is thoroly Vedic. Rig- THE SUFFIXES MANT AND VANT IN SANSKRIT. 55 Veda retains the n in 13 possessives and loses it in 2. Atharva- Veda retains it in 9 and loses it in 2. In several of the examples lengthening of the remaining a has taken place. The same word may thus be found with two or even three parallel possessive forms : 1. With final n retained ; 2. With simple loss of the n ; 3. With lengthening of the remain- ing vowel. Stems in in always lose the n before the possessive suffix, and very rarely is there compensatory lengthening of the i. Examples of the loss of final n follow : aqmavant (cf. aqmanvanfy, agvimant < agvin, dtmavant (cf. atmanvdnt\ ksamimant < ksdmin, ganimant < ganin, carmavant < carman (cf. carmanvanf), janmavant < janman, tardmavant < tardman, dandimant < dandin, dhdrmavant << dhdrman, dhdmavant < dhaman, narmavatl < narman, ndmavant < ndman, parvavant < parvan, pdmavant < pdman, prdnimant < prdnin, premavant huvant < sawhu, f. *xra0want (only in superlatives xraOwista, as-xraOwasizma) < Indo-Iranian *kratuant << *kratuuant. Cf. xratumant and .B.B. xvn, 341 ; ZDMG. L, 690 ; .KZ. XLIII, 278-9. anupoiOwant < *anu-paetav. Cf. &B. xvn, 341 and ZDMG. L, 690 ; ^Z. XLIII, 279. It is noticeable that with the exception of one, or possibly two, cases of the loss of i, all the above examples show the loss of w or w before 4. There is in Avestan but one certain example of the loss of a final consonant before the possessive suffix. In this case the n of final an disappears, as in nearly 50 Sanskrit possessives (cf. Part I, Chap, n, 7). paemavant < paeman. For the suggestion that gaomavant may be derived from *gao- man-vant and belongs in this paragraph, see Chap. I, 4. 5. As in Sanskrit, the final vowel of the stem is several times changed in quality before mant and vant. These changes are, for the most part, mere irregularities, but several of the changes are of I or u to a before v, a change that is shared by a small number of other Avestan words. Cf. Bartholomae, Iran. Grundriss, I, 1, 268, 12, 17 ; BB., vm, 230, and Geldner, KZ., xxvn, 245. Following are the examples : cavant, for *cuvant or *ewant < cu (< kav}, or << cl (< kay). Cf. Bartholomae, Iran. Grundriss, i, 1, 181, a, 2, and 268, 12, 17. Istavant (< istay, f.) for *istivant. vohunavant < *vohuna = vohuni = vohuni. yaoxstavant = yaoxstivant, < yaoxstay. THE SUFFIXES MANT AND VANT IN AVESTAN. 91 dmavant, nom. prop., is a aTrag keydpevov, and is found in Yt. 13, 125. Its derivation is in doubt and Bartholomae attempts no explanation. It is perhaps from drdzav, m. { the right way ' (cf. drdzav, m. nom. prop.). If the adjective is derived from wdzav, we should expect *9rdzumant ) but the form drazavant is, possibly, due to the analogy of bzrdzavant, which is also the name of a believer and which shortly precedes the only occurrence of araza- vant, being found in Yt. 13, 119 and Yt. 13, 124. raoxsndmant < raoxsna, nt. Cf. Bartholomae, Iran. Grundriss, I, 1, p. 173. Spiegel (Grammatik, p. 221) derives the adjective from raoxsnav, nt. afrasimant, perhaps for *afrasumant ) < a -f/ra-sav, f. The reading is not certain. 6. In a few words a vowel has been inserted between the stem and the possessive suffix. There are, however, no cases of inserted consonant as in Sanskrit. The examples follow : a. a in vayavant, which is, according to Bartholomae ( Worterbuch, s. v.) from vay, m. It may however be from *vaya. Cf. Chap. I, 4, note. vWusavant < *vWus. bdrdzavant, probably from 6ara2, f. Cf. ZDMG. L, 724. tizinavant. Justi (Grammatik, 384) thinks that an a has been inserted in this word. It is evidently from some noun derived from tizi. Bartholomae compares Sanskrit tejasvant, but the resemblance apparently consists only in containing the same root (Sanskrit ~[/tij}. Cf. Avestan tizvant. Why not compare Sanskrit tiksnd? b. a in cyavant. Caland (KZ. xxxm, 463) says : " cydvant ist also nebenform zu cvant und hat sein a wohl in anlenung an formen wie aetavant, avant, tavant bekommen, welche urspriinglich langes a hatten." c. i in rdmanivant ma, ' power ' ; arditivant, < avaricious ', < ardtay, ' avarice ' ; arsnavant, 'having a stallion ', < arsan, ( stallion ' ; dOravant, ' having fire ', < dtar, ' fire ' ; Istavant, ( wealthy ', < Istay, ' wealth ' ; 94 THE SUFFIXES MANT AND VANT IN AVESTAN. ustanavant, ' having life ', < iistana, ' life ' ; gaonavant, ' having hair ', < gaona, ' hair ' xratumant, ' wise ', < xratav, ' wisdom ' ; pourumant, ' crowded', i. e. ' having many', <^parav, 'many' ; brdtravant, ' having brothers ', < bratar, ' brother ' ; yaoxstivant, yaoxstavant, ( active ', < yaoxstay, ' activity ' ; vanaitivant, ' victorious ', < vanatay, ' victory ' ; vardcahvant, ( powerful ', < vardcah, ' power ' ; Vdrz9ravant, i victorious ', < varddra, ' victory ' ; visavant, ' poisonous ', < vlsi, ' poison ' ; raocahvant, ' light ' (adjective), < raoeah, ' light' (noun), < rao&, ( light ' (verb) ; spanalivant, ( sanctus ', << spanah, ' sanctitas ' ; aWyejahvant, ' having no danger ' ; afnahvant, ' rich in property ' ; afraddrdsvant, i having no outlook ' ; astaiOivant, 'eighty-fold', i. e. ' having eighty'; ahunavant, ( containing the Ahuna ' ; ahumant, 'one who posseses an Ahu'; diOivant, 'full of harm', 'danger'; ustdnavant, 'having life' ; gaomant, ' having cattle ' ; gaomavant, ' having milk ' ; xsvastivant, ' sixty-fold ', i. e. ' having sixty ' ; ca6war9sa9want, ' forty-fold ' ; Orisa0want, ' thirty-fold ' ; pancasaOwant, ' fifty-fold ' ; vlsaitivant, ' twenty-fold ' ; satavant, ' hundred-fold ' ; sp^r^zvant ) ' having jealousy ', ' a rival ' ; haptaidivant, ' seventy-fold '. Further, afrourvisvant, anupoiOwant, anusavant, ardjahvant, araOamant, arsavant(ty, astvant, azinavant, asavant, asivant, dzantivant, drdzvant (?), *xra6want, xrvant, xstdvant, xsviptavant, caxravant, cazdahvant, ci9ravant y cistivant, ta/nahva r nt ) fomahvant, daxstavant, dasaOavant, drafsakavant, 0amnahvant, Qwayahvant, THE SUFFIXES MANT AND VANT IN AVESTAN. 95 iOyejahvant, tbaesahvant, tbisvant, daibisvant, pardnahvant, pasvant, pasnvant, puOravant, baoSahvant, bardzavant (?), fracardOwant, frazaintivant, frasumant, fsumant, nairyam-ham-vdrstivant, ndiri- vant, ndmahvant, maSumant, marzdikavant, mazisvant, mazgavant, mdyavant (?), mvsdavant, myazdavant, ydtamant, vayavant, vastra- vant, vastravant, V9ryzvant, vohumant, vohvdvant, vixrumant,* v~ibwd0want } vlvahvant, raevant, raoynavant, raoxsn9mant } raocina- vant, ratumant, rdmanivant, irimant, urvistravant, saokdntavant, savavant, spdnavant, zaenahvant, zaranydvant, zairimydvant, zardnu- mant, zaranumant, saetavant, haMumant, haomavant, haSdnaepata- vant, hara h uvatl, harax v aitl, hym-urvisvant, Jiunaravant, hvyOwavant, x v anvant, x v 9nvant, x v ar&nahvant, x v ddravant. 2. In more than 7 ^ of the possess! ves the suffixes convert the word to which they are added into a present active participle, the percentage being the same as in Sanskrit. But in Avestan vant is never added to the past passive participle to form present participles as was frequently done in post-Vedic Sanskrit. Following are the examples : aenahvant, c practising evil ' ; afrasdhvant, ' granting ' ; afrasimant, ' not moving forward ' (cf. further KZ. xxvii, 228, and ZDMG. L, 136) ; ustanavant, i living ', < ustana, ' life ' ; pavant, ' protecting ' ; mwryOwa/nt, ' thinking of ; vydvant, 'beaming' ; saocinavant, ( flaming ' ; and isvant, bdnumantj bdnvant, fracarsdwant, frasumant, ydtumant, stdrdOwant. 3. In Sanskrit nearly 150 possessives have the meaning ' containing the root or word, or a derivative of the root or word '. In Avestan this number is reduced to 4. In each language this meaning was an independent development. In Sanskrit it is post-Vedic and none of the four Avestan examples is found in Sanskrit. 96 THE SUFFIXES MANT AND VANT IN AVESTAN. The Avestan words are : ahumant, ' containing the word ahu ' ; ustavant, ' containing the word usta ' ; ratumant, ' containing the word ratav ' ; x r aetumant, 'containing the word x v aetus'. 4. Class 4, in which the suffixes cause no change in meaning, is entirely unrepresented in Avestau. The reason for this is doubtless a double one. First, the entire number of mant's and vant's in Avestan is so small that the number of examples in one of the smaller Sanskrit classes may by chance be reduced to no representation in Avestan. Second, most of the Sanskrit pos- sessives in Class 4 are derived either from ordinary adjectives or from past passive participles, whereas in Avestan the suffixes are added to only four adjectives and to but one past passive parti- ciple. 5. Class 5, in which mant and vant convert a word into a past passive participle, has a few examples in Avestan. The number is limited, as in Class 4, by the fact that the suffixes are not (with one exception) joined to past passive participles in Avestan. Examples are : tdmakvant, ' deluded ', < tzmah, l darkness ' ; nasumant, ' rendered unclean by corpses ' ; nt, nom. prop., literally, 'darkened', 'deluded'. 6. There is but one example in Avestan of what Whitney rather inaccurately calls the past active participle in tavant, and which is so common in classical Sanskrit (cf. Whitney, Gram., 959, 960. For further references see Part I, Chap, in, 6.). The ta is the ending of the past passive participle to which vant is added, producing a derivative which has the meaning and construction of a past active participle. The single Avestan example is vlvardzdavant, ' one who has become great or mighty ', < varzzda, past passive participle, ' increased ', ' grown great ', < vardd, 'increase'. THE SUFFIXES MANT AND VANT IN AVESTAN. 97 7. In Avestan, as in Sanskrit, in 3 c / of the examples the suffixes are used actively instead of passively and possessively, and have the meaning ' giving, bestowing, granting, offering, yielding, paying, inflicting, &c.' Thus : pavant, ' protecting ', 'giving protection', <; *pd(y), nt., 'pro- tection ' ; paOravant, ' granting protection ', < *pa&ra, ' protection ' ; vldusavant, ' furnishing proof of guilt ', < mdiis, ' proof ; saokavant, 'yielding profit', -if I S ^ s - ^ ^ ^ ( QLOCT141996 ^- S3 I ?5; S5 I fUNIVERS/A .vvlOS-ANCHftu txri f/op-l 1 1 UDNV-SO e= %a3AiN-3\\v \ a 5 9 I I ^^ ^ * ^E-UNIVER^ ^ ^ ^ 'S s v-