fyxmll WLnivmity ptag BOUGHT WITH THE INCOME FROM THE SAGE ENDOWMENT FUND THE GIFT OF ' Henry W. Sage 1891 AdUrp £/./M//f..;£ PK 677.U3Tl898 VerSi,y "^ * Tifiiii 1 ™ l Mi!ii , 1 i l §ffi!f,I'!!,P honetic s :in comp 3 1924 023 201 068 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924023201068 A. MA.NTJ.AL SANSKRIT PHONETICS. A MANUAL OP SANSKRIT PHONETICS. IN COMPARISON WITH THE INDOGERMANIC MOTHER-LANGUAGE, FOR STUDENTS OF GERMANIC AND CLASSICAL PHILOLOGY, Dr. G. G. UHLENBECK, Extraord. Professor of Sanskrit and Comparative Philology in the University of Amsterdam. ENGLISH EDITION BY THE AUTHOR. London : LUZAC & Co. Publishers to the India Office. 46, Great Russell Street. 1898. & /\> \i*bs V PRINTED BY E. J. BRILL, LEYDEN (HOLLAND). PREFACE. The idea of publishing an English edition of my Handboek der Indische Klanhleer (Leiden, Blankenberg & C°. 1894) was suggested to me by others; but for that suggestion this translation certainly would not have been undertaken. In general I have followed the Dutch text: alterations have been made only, when there were positive inaccuracies in the Dutch edition or when an alteration had become necessary because of the progress of comparative philology since 1894. I have also considerably augmented the references to the scientific literature and somewhat enlarged the part on the accent. Some emendations are due to the recensions of Kern, Speyer, Zubaty [and Johansson], of whose observations and remarks I have made a grateful use. I further need only repeat, what I said in the preface to the Dutch original : the purpose of my having composed this little handbook is to introduce the student into the compara- tive-phonetic study of Sanskrit. Though as a rule students of Classical and Germanic philology do not read Vedic texts, I have not hesitated to cite words and forms, which had already become obsolete in the epic and classical language, mentioning, of course, the fact, that such a word only oc- VI PREFACE. curs in Vedic. In general I have avoided to cite Iranic, Li- thuanian and Slavonic: I know^ it is true, that this lessens the scientific worth of my book, but I do not think it desi- rable to trouble the student with several languages, which are either quite unknown to him or which he only just begins to learn. May this edition be of some use to the students in England, America and India. Amsterdam, Nov. 1897. C. C. Uhlbnbeck. CONTENTS, Page Preface T . INTEODUCTIOK § 1. The Indogermanic family of languages. ... 1. § 2. The Aryan group 2. § 3. The Indian dialects 3. § 4. Sounds and letters 5. PAET I. SYSTEM OF VOWELS. A. The representation of the Indogermanic vowels in Sanskrit. § 5. The Indogermanic vowels in general .... 9. § 6. The Indogermanic vowel-gradation 10. § 7. Idg. a 14. § 8. Idg. a 16. § 9. Idg. e 17. § 10. Idg. e. 18. § 11. Idg. o 19. § 12. Idg. o 21. § 13. Idg. 9 22. § 14. Idg. i 26. § 15. Idg. i 27. § 16. Idg. u 27. VIII CONTENTS. Page § 17. Idg. « 28 - § 18. Idg. m y 8. § 19. Kg. 9 29 - § 20. The problem of the long sonant nasals. ... 30. § 21. Idg. r, I '31. B. The relation op the Indian towels to the Indogermanic. § 22. The Indian vowels in general 32. § 23. Skr. a 37. § 24. Skr. a 38. § 25. Skr. i 38. § 26. Skr. i 39. § 27. Skr. u 39. § 28. Skr. a • 40. § 29. Skr. e 40. § 30. Skr. o 41. § 31. Skr. r, f, i, l 41. § 32. Skr. ai 42. § 33. Skr. au 42. PAKT. II. SYSTEM OF CONSONANTS. A. The representation aw the Indogermanic consonants in Sanskrit. § 34. The Indogermanic consonants in general . . . 43. Semivowels. § 35. Idg. i 45. § 36. Idg. u 46. Nasals. § 37. Idg. m 47. CONTENTS. IX Page § 38. Idg. n 48. § 39. Idg. n 50. § 40. Idg. n 50. Liquids. §41. The liquids in general 50. § 42. Idg. r 51. §. 43. Idg. 1 52. § 44. Combinations of r and I with dentals .... 53. § 45. Metathesis of r 56. Labial explosives. § 46. Idg. jp 56. § 47. Idg. b 57. § 48. Idg. bh 58. Dental explosives. § 49. Idg. t 59. § 50. Idg. d 61. § 51. Idg. dh 62. Guttural explosives. § 52. General remarks 63. § 53. The palatalization-rule 63. § 54. Idg. q, h 66. § 55. Idg. qh, hh 67. § 56. Idg. s , g 68. § 57. Idg. gh, gh 68. Palatal explosives. § 58. General remarks 69. § 59. Idg. h 70. X CONTENTS. Page § 60. Idg. § 71. § 61. Idg. gh 72. Spirants. § 62. Idg. * 73. § 63. Idg. « 76. § 64. Idg. j 77. § 65. Idg. v 78. § 66. Idg. y 78. § 67. Idg. p, $ 79. B. The relation or the Indian consonants to THE INDOGERMANIC. Semivowels. § 68. Skr. y 80. § 69. Skr. v 81. Nasals. § 70. Skr. m 81. § 71. Skr. n 81. § 72. Skr. n 82. § 73. Skr. n 83. § 74. Skr. n 83. § 75. Nasalization of vowels 83. Liquids. § 76. Skr. / 83. § 77. Skr. 1 84. Labial explosives. § 78. Skr. p , 84. § 79. Skr. ph 85. CONTENTS. XI Page § 80. Skr. b 85. § 81. Skr. bh 85. Dental explosives. § 82. Skr. t 86. § 83. Skr. th 86. § 84. Skr. d 86. § 85. Skr. dh 87. Lingual explosives. § 86. Skr. t 87. § 87. Skr. th 88. § 88. Skr. d 88. § 89. Skr. dh 89. Guttural explosives. § 90. Skr. k 89. § 91. Skr. hh 90. § 92. Skr. g 90. ^ 93. Skr. gh 91. Palatal affricates. § 94. Skr. c 91. § 95. Skr. oh 91. § 96. Skr. j 91. § 97. Skr. jh , 92. Spirants. § 98. Skr. g 92. § 99. Skr. s 93. XII CONTENTS. Page § 100. Skr. * 93. § 101. Skr. A 93. PAET III. SANDHI-SYSTEM. § 102. General remarks 95. § 103. Final and initial vowels 96. § 104. Simplification of consonant-groups 98. § 105. Final nasals 100. § 106. Voiced and voiceless explosives 101. § 107. Explosives before nasals 102. § 108. Assimilation of dental explosives to palatal af- fricates, to g, to Unguals and to I . . . 102. § 109. Treatment of -* and -s 103. § 110. Treatment of -r 106. § 111. Changes of initial consonants 107. PART IV. ACCENT. § 112. The accent of the Indogermanic period. . . 109. § 113. Accentual agreement between Sanskrit and Greek 110. § 114. Accentual agreement between Sanskrit and Ger- manic 112. § 115. Dissyllabic value of long vowels 113. § 116. Accentuation of text? 113. § 117. Enclisis 114. INTRODUCTION. § 1. The Indogermanic family of languages. The great family of languages, to which Sanskrit belongs, is called the Indogermanic, Indoceltic or Aryan. I prefer the first name, because it is the most usual, though the name Indoceltic may claim a greater accuracy. The word Indogermanic dates from a time, when it was not yet proved, that the Celtic dialects also make part of our family of languages, and indicates by the combined name of the utmost branches, Indian and Ger- manic, the whole territory of speech, to which they belong. Now that it is certain, that Celtic also is a member of our family , it would be accurate to replace the word Indogermanic by Indoceltic, because not Germanic, but Celtic is the utmost branch to the Occident. The name Indogermanic however is generally adopted and it would be impossible to supplant it by another. By the word Aryan is generally understood a certain subdivision of the Indogermanic family , viz. the Indo- Iranian, and therefore it would seem unsuitable to use this name also for the whole Indogermanic family. See G. Meyer, Idg. forschungen 2, 125 sqq. and Spiegel, Die Arische periode (Leipzig 1887) VI sq. The Indogermanic family consists of the following nine groups : 1. Aryan, see § 2. 2 INTRODUCTION. 2. Armenian. 3. Phrygiam,-Thracian , only known from proper names , glos- ses and inscriptions. 4. Albanian. 5. Greek. 6. Italic, which comprises not only ancient languages of Italy (Latin, Oscian, Umbrian, Samnitic), but also the mo- dern dialects, which have sprung from popular Latin. 7. Celtic, which is divided into Gallic (the extinct lang'uage of ancient Gallia), Britannia:: (Cymric or Welsh, Cornish, Bas-Breton) and Gaelic (Scotch, Irish, Manx). 8. Germanic, which is commonly divided into an eastern and a western group. The eastern comprehends Gothic and Scan- dinavian (Icelandic, Danish, Norwegian, Swedish); to the western belong English (in its oldest literary period called Anglosaxon) , Frisian and German (High-German and Low-German) 9. Balto-Slavonic , which consists of two groups, the Baltic and the Slavonic. The former is divided into Old-Prussian, Lithuanian and Lettish; the latter comprehends Southern- Slavonic (Ecclesiastical-Slavonic , Slovenian , Servo-Croatian , Bulgarian) , Bussian (Great-Russian , White-Russian , Little- Russian') and Western-Slavonic (Polish, Wendish, Bohemian and the extinct Polabic). About the mutual relations of these groups see J. Schmidt, Die verwandtschaftsverhaltnisse der Indogermanischen sprachen (Weimar 1872) and cf. Brugmann, Zur frage nach den ver- wandtschaftsverhaltnissen der Indogermanischen sprachen (Tech- mer's Internationale Zeitschr. fur allgemeine sprachwissenschaft 1, 226 sqq.). § 2. The Aryan group. The Aryan group is divided into two subdivisions : INTRODUCTION. O 1. Indian, see § 3. 2. Iranian. Ancient Iranian is handed down to us in two dialects, viz. Avestian and Old-Persian. Avestian is the language of the Avesta, the sacred book of the Parsis. The oldest parts of this Zoroastrian bible date back to many centuries before Christ, and their language is more archaic than the classical dialect of the younger Avesta. This oldest Avestian is called the fta^a-dialect. In Old-Persian we have only the cuneiform in- scriptions of the Achaemenidian kings, the oldest of which is that of Darius Hystaspes at Behistan and dates from + 520 before Christ. Numerous inscriptions are found on and near the ruins of Persepolis : the youngest of them is of the middle of the fourth century before Christ. See Bartholomae, Hand- buch der Altiranischen dialekte (Leipzig 1883), Vorgeschichte der Iranischen sprachen (Grundriss der Iranischen philologie I) ; Williams Jackson, An Avesta grammar (Part I, Stuttgart 1892); Spiegel, Die Altpersischen Keilinschriften 2 (Leip- zig 1881). The language of the Sassanidian period is called Middle- Persian, Pahlavi or Huzvaresh. The modern Iranian dialects are New-Persian, Kurdish, Ossetic, Afghan, Balucl &c. Some of the principal works on these languages are : Horn , Grundriss der Neupersischen etymologie (Strassburg 1893); Hubschmann, Persische studien (Strassburg 1895), Etymologie und laut- lehre der Ossetischen sprache (Strassburg 1887); Justi, Kur- dische grammatik (Petersburg 1880); Trumpp, Grammar of the Pasto (= Afghan, London 1873) ; Geiger, Etymologie des Balucil (Miinchen 1890), Lautlehre des Balucl (Miinchen 1891); Tomaschek, die Pamir-dialekte (Wien 1880). § 3. The Indian dialects. In the first place is to be 4 INTRODUCTION. mentioned the Vedic dialect, which was spoken in the Paujab and in Kabulistan ± 1500 before Christ. Here arose the greater part of the rc-hymns, which in a later period, when the Iiido- Aryans had spread more to the East, were collected into the Rgvedasamhita. From this time dates the greater part of the Vedic literature. From an Indian dialect of the Veda-period sprang the samskrta bhasa (adorned , elaborated language) of Madhyadeca (the Mid- land , on the upper course of the Ganga and Yamuna) , which some centuries before Christ must have been a living language, be it not in quite the same form as in most of the literary Sanskrit works. To the time, when Sanskrit was yet living as a popular language, we may refer at least the bulk of the great epic poems Mahabharata and Ramayana. There are several reasons to assume that this spoken Sanskrit descends from an other Old-Indian dialect than that of the Vedic hymns (cf. § 43). The most indispensable books for Sanskrit students are: Whitney; A Sanskrit grammar 3 (Leipzig 1896), The roots, verb-forms and primary derivatives of the Sanskrit language (Leipzig 1885); Wackernagel , Altindische gram- matik I (Gottingen 1896); Bohtlingk & Roth, Sanskritwor- terbuch (Petersburg 1855 — 1875). See also the list of books in Lanman's Sanskrit reader (Boston 1888), XVII sqq. While Sanskrit continued as the language of the learned and educated, there arose numerous popular dialects or Pra- krit's {prakrta bhasa, ordinary language), many of which are known to us, partly from separate works, partly from Sanskrit plays (fifth, sixth and following centuries after Christ). At how early a time Sanskrit can no longer have been a popular language, appears from the Prakrit inscriptions of Acoka, king of Pataliputra, who reigned from 259 till 222 before INTRODUCTION. Christ. Particularly interesting is that Prakrit, in which the sacred books of the southern Buddhists are written and which is called Pali. Kern (Geschiedenis van het Buddhisme in Indie , Haarlem 1882—1884, 2, 339) says, that Pali is a dialect, the place of which is not yet fixed with entire certainty , but that at all events it was not the language of Acoka and the kingdom of Magadha. Oldenberg (Buddha, Berlin 1881, 76) thinks , that Pali was once the popular language of southern India. See about the ancient popular dialects of India Lassen , Institutiones linguae Pracriticae (Bonnae 1837); Jacobi, Ausge- wahlte erzahlungen in Maharastri (Leipzig 1886); Ernst Kuhn, Beitrage zur Pali-grammatik (Berlin 1875) and other works. From these Prakrits descend the modern Aryan vernaculars of India: Hindi (intermixed with numberless Persian and Arabian words it is called Hindustani or Urdu), Panjabi, Sindhl, Gujarat!, Marathi, Oriya, Bengali Ac. See Beames, A comparative grammar of the modern Aryan languages of India (London 1872—1879). § 4. Sounds and letters. The oldest Indian writing, that is known to us, is found in the inscriptions of Aqoka: here we distinguish two different systems of characters, one of which has been evidently derived from a Semitic alphabet and the other may be of the same origin. From the latter descend the younger Indian alphabets, to which belongs De- vanagari, properly the alphabet of Hindustan. The natives of India write Sanskrit in different modes , but by western scholars Devanagari is only used. Here I give a list of the Sanskrit letters, according to the most usual system of transliteration: each character has always the same phonetic value and all sounds of the lan- guage are represented by a separate character. 6 INTRODUCTION. Vowels. a, a, i, i, u, ii. e , o. T: f> It I- Diphthongs. ai, au. Mutes. Gutturals: k, kh, g, gh. Linguals: t, th, d, dh. Dentals: t, th, d, dh. Labials : p , ph , 6 , i h. Palatal affricates. c, ch, j, jh. Nasals. n , n , n, n, m and the anusvara and anunasika , \"hich in transliteration are both written m. Liquids, r, I. Semivowels. H> v - Sibilants. f, ?, «• Aspiration. h and the visarga h. INTRODUCTION. 7 The usual order of arrangement is slightly different from that giyen ahove. A few words must be said on the phonetic value of the Devanagari letters. The a, i and u are short vowels: their pronunciation is nearly the same as in German. The corresponding long vowels are a, i, u. The e and o are long vowels with a narrow pronunciation: they are contractions of ai and au. The r and I are sonant liquids; f and I are the correspon- ding long ones, but I does not occur in a single genuine Sanskrit word. The ai and au are diphthongs : their first component a is long, the second components are the semivowels i and u. The ~k and g are the ordinary European k- and ^-sounds.- Their corresponding aspirates are kh (k-\-h) and gh (g -\- h). The t and d are distinguished from the dentals by turning up the tip of the tongue to the palate. Their corresponding aspirates are th (t-\-h) and dh (d ■+- h). The t and d are ordinary dentals, i. e. they are formed by turning up the tip of the tongue to the upper teeth. Their corresponding aspirates are th (t -(- h) and dh (d-\-h). The p and b are ordinary labials and ph (p -\- h) and bh (b -|- h) are their corresponding aspirates. The c and j are palatal affricates : c consists of t -J- * and j of d-\-%. Engl, ch in chaff, chain, choose, churl and g, j in gem , gentleman , gin , jail , jam , jest , jockey are nearly the same sounds. Each of the nasals is akin to one of the series of conso- nants given above : the n is guttural (= engl. ng in long , sing), the n lingual, the n dental, the m labial and the n O INTRODUCTION. palatal. By m (anusvara and anunasika) is indicated the nasal pronunciation of a preceding vowel. The r must have been a lingual sound, because it has the power to lingualize neighbouring dentals. The I is a dental sound. The y is a semivowel , which is closely related to the vowel i; i.e. it is a consonant i like engl. y in year, yard, yoke. The v originally stood in the same relation to u and had the sound of engl. w in wall, word, wear . hut from an early time it was changed to a spirant (engl. v in veil, vain, vine), except when preceded by a consonant in the same syllable. The g is palatal like engl. sh in short, shield, shut. The s is lingual and distinguished from the * in the same way as t, d, n from t, d, n. The s is our ordinary surd *. By h and h are understood our usual h (spiritus asper) , but the visarga (h) is usually neglected by us in the pro- nunciation of Sanskrit words. PAET I. SYSTEM OF VOWELS. A. The representation of the Indogermanic vowels in Sanskrit. § 5. The Indogermanic vowels in general. There was a great variety of vowels in the mother-language, but their accurate pronunciation is not definable. We indicate these vowels by the following letters: a, a; e, e; o, oj 9j i, i) u, w. The e, e, 0,0 had a broad pronunciation. The 9 was per- haps something between a and engl. e in oldest, father. There were also sonant nasals and liquids, which are writ- ten m , %, r. I. We understand by 13, not only the dental n- sonans, but also the palatal and guttural sonant nasals. Some scholars deny the existence of these sounds in the mother- language : see Bechtel, Die hauptprobleme der Indogermanischen lautlehre seit Schleicher (Gottingen 1892), 114 sqq. and J. Schmidt, Kritik der sonantentheorie (Weimar 1895), cf. Hirt, Idg. forschungen 7, 147 sqq. Combinations of vowels with semivowels , nasals and liquids were very common : 10 MANUAL OF SANSKRIT PHONETICS. at, at, ei, et, 01 , °&i au , au, eu , eu, , ou, ou ; am, am, em, em, om, dm, 9m; an, an , en , en, on, on. 9n; ar , ar , er, er, or, or, gr; al , al , el , el , ol , ol, 9l. Many scholars assume, that there were also long sonant nasals and liquids in the mother-language. See de Saussure, Memoire sur le systeme primitif des voyelles dans les langues Indo-Europeennes (Paris 1887), 262 sqq. and Brugmann, Grundriss der vergleichenden grammatik der Indogermanischen sprachen (Strassburg 1886 — 1893) , 1, 208 sq. 243 sqq. : against this theory Bechtel 203 sqq. § 6. The Indogermanic vowel-gradation. The Indoger- manic mother-language had four series of vowels, based on e, e, a and 6. De Saussure 134 sqq. has tried to prove, that the long base-vowels e, a, o are contractions of e with a short vowel and considers the e- , a- and o-series only as special cases of the e-series. Hiibschmann , Das Indogermanische vocalsystem (Strassburg 1885), Brugmann 1, 248 sqq., Bartholomae, Bezz. Beitr. 17, 91 sqq. and others assume an o-series and an o-series besides the four given above. Against this opinion Bechtel 256 sqq. The base-vowels e, e, a, o were weakened in all syllables, which had but a slight accent or no accent at all. In the latter case the radical vowel was lost without leaving any trace, but in syllables with a slight accent e was reduced to 9 (Bartholomae, Bezz. Beitr. 17, 109 sqq., cf. also skr. timird-s c dark : ' : tdmas darkness 5 ) , e to e or 9 , a to a or 9 , 6 to o or 9 (Bechtel 248). But a regular vowel-exchange is found also in accentuated SYSTEM OP TOWELS. 11 syllables : so we have an ablaut e : o , e: b , a: o , but the base- vowel 6 is not met with in a modified form. The causes of this ablaut are not yet clear. Kretschmer (Kuhn's Zeitschr. 31, 366 sqq.) has shown, that G. Meyer's way of explaining it (Kuhn's Zeitschr. 24, 227 sqq.) can not be the right one. The base-vowel e (o) often was lengthened to S (6) : Streit- berg (Idg. forschungen 3, 305 sqq.) has tried to explain this phenomenon. I see no reason to admit with Bartholomae , that there was also a dehnstufe of e, a, o. 1. e-senes. Weak: Strong : Lengthened : A. — ,9. e, 0. e, 0. 13. i (»), i. ei (ie) , oi do). Si (ie) , oi (id). C. u (u), a. eu (ue) , ou (mo). eu (uS) , ou (uo). D. m (m), 9m (m,9). em (me) , om (mo). em (mS) , dm (mo), E. n (n), sn (n»). en (fie) , on (no). en (tiS) , on (no). F. r (r) , 9r (rn). er (re) , or (ro). er (rS) , or (ro). Gk I (I), 9l (l 9 ). el (le) , ol (lo). el (IS), ol (Id). When e was lost before (or after) i, u, m, n, r, I these semivowels, nasals and liquids remained consonant before vowels, but became sonant before consonants. Concerning i, u, dm, on, dr , dl must be observed, that they occur not only as 9-degree of ei, eu, em, en, er , el, but also as weakest degree of eie, eue, erne, ene , ere, ele: see de Saussure 239 sqq. ; Brugmann 1, 104 sqq.; Bartholomae, Bezz. Beitr. 17, 109 sqq. I assume with Bartholomae, that * and u are contractions of 9 -\- i and 9 + u and stand on a level with 9m, gn, 9r , si. 12 MANUAL OF SANSKRIT PHONETICS. II. e-series. Weak: Strong: A. — ,9, e. e, o. B. i (i), i, ei (ie). ei (ie) , oi {id). G. u (u), u, eu (we). eu (ue), ou (uo). The relation of i, u to i, u in ei- (ie-) and eu- (ue-) roots is the same as in ei- (ie-) and eu- (ue-) roots: we find i, u before vowels and i, u before consonants. About the ei- and , Stint, lat. odor, oleo. Strong: gr. oha%x, tiiailvic. Cf. Brngmann 1 , 250 sqq., who differs in many points from the exposition given above. It can not be denied, that it is very difficult to distinguish the e-series from the a-series, cf. skr. caru- , gladsome, plea- sant: lat. cams, goth. Mrs; obg. stem, stam: gr. "gthihi (dor. " &c. idg. *aw3-, *«»e-: skr. dniti, he breathes, anild-s , wind,., gr. avepoc , lat. animus, anima , goth. praet. wzo», expired. SYSTEM OF TOWELS. 15 idg. *anti: ved. dnti, opposite, before, gr. xvrl , lat. ante, goth. and. idg. *apo: skr. ap«, away, forth, off, gr. xiri , goth. af. idg. *arg-: skr. drjuna-s, white, ved. rajatd-s, silvery, skr. rajatd-m, silver (perhaps idg. *rag- by the side of *arg-; Avestian has SrZzata-, silver), gr. xpyupog , lat. argentum. idg. *ghans- : skr. hamsd-s , goose , swan , gr. %yv , lat. awer (^hanser) , ohg. yarcs. idg. **«^- (weak form of *«'«#-) : skr. ydjam.i (for *yajdmi , idg. *iagb) , I worship, gr. xtyftxi (*iag-io-mai), cf. skr. yajhd-s, worship , sacrifice , gr. xy vis (idg. *iagnd-s) , ved. #«/<& , worship , gr. Sty os (idg. *iagos), ved. yajya- , venerable, gr. ay/c? (idg. *iagio-s , *ia§iio-s). idg. *kakwd-, *kahld- : skr. /caMi! (<#) , kakudman (nt) , top , summit , lat. cacumen (*cacildmen). idg. *kark-\ skr. karkd-s , karkata-s , crab, gr. Kxpnivos , lat. cancer (*carcer or * car cent) idg. *mad-: ved. madami (*maddmi), I am drunk, gr. [^xSxai , lat. madeo (see Zubaty, Archiv f. Slav. phil. 13, 418 sqq.). idg. *skando : skr. skdndami (*skanddmi) , I spring , lat. scando. Heterosyllabic a* , o?. idg. *nebhos : skr. ndbhas , cloud , sky , gr. vfCpos , cf. gr. veQeXvi, lat. nebula, ohg. nebul. idg. *penqe : skr. pdnca, five, gr. irevTe, lat. qwinque, goth. ^?»«/'. idg. ^ej 1 - : skr. pdcami , I cook , gr. 7rierf- : skr. pdtami , I fly ; I fall , gr. ireTo^xi , wiirra , lat. ^)e&?. idg. *£« : skr. ca , and , gr. tc , lat. §■««. idg. *dorkeip : skr. dargdyami , I let see , I show, goth. gatarhja. idg. *domo-s ,. *domu-s: ved. ddma-s , house, gr. "hoyLoq , lat. idg. *gegona: skr. jajana, I produced, gr. yiyoi/«. idg. * < ghono-s: skr. gJiand-s , slayer (ved.), compact, com- pacted mass, cloud, gr. (povo?, cf. skr. hdnmi, gr. df/W. 20 MANUAL OF SANSKRIT PHONETICS. idg *koksa: skr. kaksa (kdksa-s) , region of the girth , girdle , cincture, circular wall, enclosed court, lat. coxa, mhg. hahse. idg. *qoteros : skr. katard-s , who (from two) , gr. irirspoi: , ion. xiTepog, cf. goth. Awapar. idg. *moni-s : skr. mani-s , jewel (though Vedic, this word must be a pracritism; genuine-skr. would be *mani-s), os. meni , cf. lat. monile. idg. *p6fcs: skr. pdti-s, lord, husband, gr. irfoi? , goth. -/ia^tf , cf. the feminine idg. *potni (gen. *potniies or *potniias) : skr. pdtni, lady, mistress, wife, gr. irorvia. idg. *pr6M; skr. prdti , against, gr. tt/kjt/ (irp&q). idg. *rStho-s , *rotha : skr. rdtha-s , chariot , lat. rofa , ohg. «w2. Of. also gr. piio? , which allows however an other explication. idg. *ulqo-s, *M,qo-s : skr : vfka-s, wolf, gr. \vxog , goth. wulfs. idg. *e-bhereto : skr. dbharata , gr. icpipsro , 3 pers. sing, imperf. med. from the root *M^- , to bear. Heterosyllabic oi , ou : idg. *duoid-s: skr. dvayd-s , double, gr. So/o'«. idg. *6ui-s : skr. dvi-s , sheep , gr. oig , olg , lat. ovis , ohg. awi. Tautosyllabic oi , ou = skr. e , o (contracted from ai , au) : idg. *leldiqa: skr. rireca, I left, gr. Xekotirix., goth. feifc. idg. *u,6ida: skr. »«fo, I know, gr. oJSa , goth. «wtf. idg. HMrois: skr. Ihdres, thou mayst bear , gr. cpepo/? . goth. idg. HhebMudha: skr. bubSdha, I perceived, cf. goth. ioajb. idg. *sun6us : skr. .swreos , goth. sunaus , gen. of skr. sunu-s , son, goth. sunus. I can not agree with Brugmann (Kuhn's Zeitschx. 24, 1 sqq., Morphol. unters. 3, 91 sqq.), Osthoff (Morphol. unters. 1,207 sqq. note, cf. 4, 303 note), Btreitberg (Idg. forschungen 3, SYSTEM OF VOWELS 21 364 sqq.) and others, who suppose, that the idg. o in Aryan always became a in open syllables. J. Schmidt (Kuhn's Zeitschr. 25, 1 sqq.) and Meillet (Mem. de la Soc. de Ling. 9, 142 sqq.) have proved, that the idg. o in open syllables generally is represented by a. The a of skr. jdnu, knee, must not be compared with the o of gr. yovv , but with the a of gr. •yavia, and skr. ddru, wood, is not identical with gr. So'/ju, but differs from it in the radical vowel (o : o).' So the a of skr. pddam , datdram does not correspond to the o of gr. icSbot,, SaTop*, but to the o of gr. dor. irin;, lat. datorem. About the vocalisation of the causatives (skr. svapdyami = lat. sopio) and iteratives (ved. patdyami, cf. gr. vorko^xi) see not only Meillet, but also Delbruck, Idg. forschungen 4, 13iJ sq. De Saussure 96 sq. and Bartholomae (Bezz. Beitr. 17, 93, 103) have made probable, that there was originally a diffe- rence between the o in the e-series and the o , which is weakened from o : in Armenian o : e is represented by o , but o : 6 has become a. In as early a period as the Aryan or Indo-Iranian these two o's had fallen together. § 12. Idg. o. The idg. 6 became a and fell together with the idg. a (see § 8) and e (see § 10). This transition took place in the same time as the change of o to a, i. e. in the Aryan period (see § 11). idg. *ok4-s: skr. aqu-s, swift, gr. dmvg , cf. lat. comp. deior. idg. *omd-s: skr. amd-s, raw, gr. d>/*6i. idg. *dbno-m: skr. danarm, gift, lat. donum, cf. idg.*didoti, skr. dddati (for *didati), gives, gr. liluxri. idg. *ulqod\ skr. vfkad, abl. of vrka-s , wolf, cf. the Latin ablatives ending in -3d, -6. idg. Hheretod : ved. bhdratad, must bear, gr. (pipera, cf. l&t.ferto. 22 MANUAL OT SANSKRIT PHONETICS. m idg. *8ueso{r): skr. svdsa (r), sister, lat. soror, goth. swistar, cf. gr. &>0 • duyxrtip , avexpiSg. idg. *bhero : skr. bhdrami (*bMra) , gr. : skr. hirdmi, I pour out, I strew. idg. *fors: ved. tirami, cf. idg. *fero, skr. tar ami, I cross, and gr. tW/3w , Tepsrpov. idg. *'fo««: ved. tirds, through, cf. lat. trans. Idg. dr , gl before consonants, not in the neighbourhood of labials, skr. ir: idg. *fars-\ ved. girsdn-, skr. girsd-m, head, cf. gr. xopavi , xoppyj and skr. giras (idg. * faros) , head, gv. xapx, xxpvi (*fara). idg. *ddlgho-s: skr. dirgkd-s, long , cf. gr. SoA^o's , si/Sf *«#>}? , lat. indulged. idg. *^s^-« (perhaps *gdr-s): skr. yw-, voice. Further a great number of passive participles as girnd-: girdmi , hirnd- : hirdmi &c. It is a matter of course , that skr. jirnd-, worn out, old, must not be identified with lat. granum (idg. *grano-). The '«?• of ved. jurdti , jilryati , decays , grows frail, and other verbs, where no labial is found, is not yet explained: probably these words are loans from a dialect, where pr became ur {ur) in every condition or they have taken their u from other verbs, in which u was developed accor- ding to the rule (sphurdti). On the other hand we should expect ur instead of ir in ved. irmd-s , arm, which with lat. armus, goth. arms goes back on idg. *grmd-s (m is a labial!). 26 MANUAL OF SANSKRIT PHONETICS. Idg. 9r , 9l before Towels , in the neighbourhood of labials , skr. ur: idg. *u9ren-: ved. wraria-s, ram, lamb, cf. gr. xpvo; , gen. of *a.fijv. idg. *p9lu-s : skr. puru- , much , gr. tto^u? (*?t«Ai;$) , cf. goth. film. idg. *sph9rt: skr. sphurdmi, I make a quick motion, I dart, I twitch, cf. gr. airxipa &c. idg. * < spry?, goth. kaurus and also lat. gravis. Idg. 9r, el before consonants, in the neighbourhood of labials, skr. ur : idg. *ugrdhu6-s: skr. urdAvd-s, high, tending upwards, gr. dor. @op9o- (*rxpho-), cf. gr. 3pSog , lat. arduus (without u in the anlaut, idg. *9rdhu6-s). idg. ^uplmi-s: skr. urmi-s, wave, &gs.wielm,wylm,ohg.walm. idg. *u9bid : skr. M , wool , cf. goth. «w2£a (but lat. lana = gr. Aafci/ij). idg. *pdruo-s\ skr. purva-s , first, cf. oslav.^rftpfi (i. e.prwU) and with w? goth. fruma. idg. *pdlMo-s\ skr. purnd-s , full, cf. goth. fulls. Lat. plenus has idg. e and corresponds to ved. prana-. idg. *ps^-5 : skr. jo&r, stronghold, fortified town, cf. gr. to'A/?. idg. "mdldhen- : skr. murdhd (n) , head, cf. ags. tnolda. idg. *hhdr§o-s\ skr. bhurja-s birch, cf. ohg. pirihha. idg. *ddrua: skr. dirva, millet-grass, dutch tanoe, wheat. § 14. Idg. i. Idg. « = skr. ». idg. *imes, *imen : skr. imas, we go , gr. Iptsv. idg. *bhibheimi: skr. bib/temi, I am afraid, ohg. bibem. idg. *<&««: skr. <&»/, in heaven, gr. Ail. idg. *«*-<£ : skr.. i-d-dm, this, lat. e'^, goth. it-a. SYSTEM OF VOWBLS. 27 idg. *uidme, *uidmen: skr. vidmd, we know, gr. ISpev, faftev, goth. witum. idg. *6ui-s : skr. dvis , sheep , gr. o'ig, o/V, lat. ovis , goth. awi-. idg. *suadistho-s : skr. svadistkars , sweetest , gr. tfhtrroc c_ ' (*as5a, dog, gr. xuwv. idg. *Uut6-s : skr. grutd-s , heard , heard of, famed , gr. x\vr6q, lat. in-dutus (cf. with w ags. AWd). 28 MANUAL Or SANSKRIT PHONETICS. idg. *rudAi>r6-s: skr. rudhird-s , red, cf. gr. ipuipie (idg. *erudhr6-s), lat. ruber. The root *rendh- , *ereudh- is a secon- dary formation from *ereu , cf. skr. arund-s and arums , ruddy. idg. *suadurs : skr. svadu-s , sweet , gr. v$vc. idg. *nu : skr. nu , now , gr. vu , goth. mw. I do not accept the theory, according to which idg. ru became r, when there was an u in the following syllable: qrrwti , hears , goes back on *kl-ne^urti (root *hehir, cf. de Saus- sure 244, PederBen, Idg. forschungen 2, 307). § 17. Idg. u. Idg. u = skr. H: idg. *bhutis (root *bhu- , *bheug-, *bheue-) : skr. bhuti-s , being, prosperity, cf. gr. Qvatq. idg. *dhumo-s : skr. dhumd-s , smoke , vapor, gr. SUftdt; , lat. fumus. idg. *kuro-s: skr. gura-s, mighty, bold, hero, gr. -Kupoi in <»-ieD/>e$ , not valid , cf. to xupog , xupiog. idg. *»» : ved. rahs- , mouse , gr. ytXis , lat. »&ws , ohg. mus. idg. *»w : ved. nu , now , gr. vuv. § 18. Idg. m. The idg. m became a during the Aryan period ; so it fell together with n (see § 19), a (see § 7), e (see §9) and o (see §11): idg. * ( gmskhdti: skr. gdcchati , goes, cf. gr. imperat. (Zxtrxe. The root is *^em-. idg. * < gmt6-s: skr. gatd-s , gone, gr. /3«toV, lat. -ventus and idg. * < gmti^s: skr. gdti-s , going, issue, refuge, gr. ficitrit; , goth. -qumpi- (root * l gem-). idg. *krp,t6-m: skr. qatd-m, hundred, gr. s-kktov , lat. centum, goth. /toe?. That we have to do with ?«, appears from lith. szimtas. idg. *dehm: skr. j? , f/3ij , dor. ? (3xv, 1/35? , I/3«, are not derived from the root * c gem-, but from its sy- nonyme * c §a- (see Persson, Studien zur lehre der wurzeler- weiterung und wurzelvariation , Upsala 1891, 70). The a of skr. dgam, gr. dor. ?/3«i/ is an idg. a (not m) and the original paradigm was *e c gam, *e l gas, *6 < gjat &c. In a similar way are to be explained said- : sandm i , void- : vandmi , ghata- (and ghati-) : hdnmi and some other cases. But jatd- , born , seems to have taken its j instead of jh by the influence of jdnati , jandyati , jajdna &c. The regular form *jhata- (idg. *gnat6-) corresponds to lat. -gnatus , gall, -gnatus. The j of the praesens jdyale has the same origin as that of jatd- {fay ate: tandti, May ale: hhdnati have been formed after the type of jayate). The substitution of _;' for jh in jatd- , jay ate is Indo-Iranian , cf. pers. zad, birth, zadan, to bear, to be born. In skr. jandmi , I know (= pers. danam) we may SYSTEM OF VOWELS. 31 explain the / instead of jh by dissimilation (*jnanami), cf. gr. yiyvaxTKu , lat. -gnosco. If ved. atirs, a waterbird, is to be compared with gr. vtjtrinx, the following explanation seems possible : there was an -i- (.-ei-. -oi-) stem *ndti- 1 'nte'i- in the mother-language, which in Aryan became regularly *ndti-, *alm-. Ved. aU- is a con- tamination of *ndti- and *atdi-. Lat. anas, ohg. anut, lith. aniis, oslav. qty are not clear (idg. *anat-, *ant-1). But perhaps it is preferable to separate ati- from vyaax (*natia) and to identify it with gr. wt/? (-/So?) , as has been proposed by Speyer (Museum 2, 435). Unexplained are Ted. dta- , dta , frame of a door : lat. anta ; ved. ydtar- , the wife of the husband's brother : gr. s'tvxTipsc : lat. janitriees &c. Von Bradke (Idg. forschungen 4, 87 sqq.) compares skr. jard- , lover , paramour , with jdmaiar- , son in law , gr. •yxpfipoi; , which combination from a semantic point of view would not seem very probable. Less credible yet is his etymology of skr. dard-, wife (: gr. 'Sxfixp, 85 sqq.), for Johansson (Idg. forschungen 3, 229 sqq.) has shown its iden- tity with gr. SwAo? ( : SoDAo?). § 21. Idg. r, 1. The idg. r and I are represented by r. Skr. klptd-, put in order, arranged, the only word that con- tains I, is a pracritism for *krptd- (the etymology of kalp- is to be found in my Etym. worterbuch der gotischen sprache , Amsterdam 1896, s. v.'Aalba). Idg. % ~ B ^ r - 2' : idg. *{kpo-s: skr. fksa-s, bear, gr. upxTos, lat. ursus. idg. *rneumi, *-rnumai: ved. rndmi, I rise, I go, I attain, gr. Spvufti (see Schmidt, Kuhn's Zeitschr. 32, 376 sqq.). XpWfiXl. idg. *rsen-: skr. rsa-bhd-s, bull, gr. &p, lat. ago, icel. inf. aha. ved. djra-s, field, gr. xypbc, lat. ager, goth. akrs. skr. skdndami, I spring, lat. scando. skr. a»«00, yocat. of arafe, mother, cf. gr. vocatives as vufi0», ^etrirOTX. Skr. a = idg. e: skr. dsti, is, gr. Io-t;, lat. est, goth. is<. skr. dbharam, I bore, gr. % wurzelerweiterung &c, Upsala 1891, 59—67, cf. P. B. Beitr. 17, 437 sqq. and my Btym. wb. der got. sprache s. v. alhs, bairhts, biraubon, gras, gredus, hruk, halbo, stilari). This Indogermanic change of r and I does not properly belong to the province of Indian phonetics and so we may leave it unnoticed. Likewise such cases as skr. bhundkti: goth. brukjan, skr. bhandkti: goth. brihan are to be explained by phonetic rules of the mother- language. § 42. Idg. r. In general the idg. r remained unchanged. Instances are: idg. *regio: skr. rdjyami , I am (get) red, affected with a strong feeling, gr. ?e&. idg. *re-s : skr. rd-s , wealth , lat. res. idg. *relc-s, stem *reg- : skr. rat (instead of *rak from *raks, which was to be expected), lat. rex. idg. *reud-: skr. rddimi, I weep, lat. rudo, ags. reotan. idg. *bhero : skr. bhdrami , I bear , gr. lat. tfopo. idg. *speJc-s: ved. sjoai (instead of *spak, stem spag-), lat. -.sjj&p in auspex , haruspex , cf. skr. pdqyami (Dhatup. also spdgami), lat, specio , ohg. spehon. How far the Aryan jj<£ (skr. ^, iran. _/) goes hack on idg. ph, is still unsettled (cf. Hoffmann, Bezz. Beitr. 18, 154 sqq.). Instances of skr. ph, iran. f: skr. kapha- , phlegm , avest. Jcafa-. skr. gaphd-, hoof, avest. 4 became &?A in Aryan, from which skr. (£/< (see § 63). idg. *mizdho-: ved. mldhd-m, prize, contest, gr. [i.ivt&q , cf. goth. mizdo. Idg. rfA» remained unchanged in Indian : idg. *bhudhn6-s : ved. ludhnd-s, bottom , cf. gr. irvbuviv, ?rui/5«£, lat. fundus and ohg. bodam. When followed by an aspirate, dh lost its aspiration: idg. *dhidhemi: skr. dadhami, I put, gr. tUviim. Idg. dh -\- I is, represented by ddh: see § 49. In many Sanskrit words and forms, when dh might be expected, we find however h. Instances: idg. *-dhi, ved. -dhi, -hi, skr. -hi, suff. 2 pers. sing, im- perat. act., gr. -fa. SYSTEM oj consonants: guttural explosives. 63 idg. *-medhai, *-medA9: ved. & skr. -make, -mahi, Buff. 1 pers. plur. med., avest. -maide, gr. -nefa. Of. von Bradke (Zeitschr. der D. Morgenl. Ges. 40, 657 sqq.), "who explains -such cases by Pracritic influence. Guttural explosives. § 52. General remarks. A few years ago only two series of gutturals were assumed, viz. a palatal and a velar series. Bezzenberger (Bezz. Beitr. 16, 234 sqq.) has proved, that there was a third series — different from the palatals and velars — , which may be called the middle gutturals. In the western Indogermanic languages, viz. in Greek, Italic, Ger- manic and Celtic, the velars were labialized and the middle gutturals have fallen together with the palatals. In Aryan, Armenian, Phrygian-Thracian , Albanian and Balto-Slavonic the velars were not labialized, but have fallen together with the middle gutturals: here the palatals were changed to spirants. The velars and middle gutturals being represented in Aryan by one series, it is preferable not to separate them in a book on Sanskrit phonetics. § 53. The palatalization-rule. The palatalization-rule con- cerns the velars and middle gutturals in general : therefore it is to be treated in a separate paragraph, the more so as the discovery of this rule has had a decisive influence on the modern vowel-theories. It has been the strongest, if not the only argument for adopting , that the multiplicity of vowels in the European languages and in Armenian is more original than the Aryan simplicity. This rule, which was discovered by several scholars at the same time (see Schmidt, Kuhn's Zeitschr. 25, 1 — 179), may be formulated as follows: 64 MANUAL Ot SANSKEIT PHONETICS. During the Aryan period , before the change of e to a, the gutturals (idg. middle gutturals and velars), were palatalized and changed to palatal affricates, when followed by idg. e, i, i, and were preserved as gutturals in any other position (cf. however Meillet , Mem. de la Soc. de Ling. 9, 376 sqq.). Instances of skr. e, j, h (from jh) , going back on gutturals before e, i, i: idg. *qe: skr. ea, and, gr. re, lat. que. idg. *qeric-s, skr. caru-s, kettle, pot, icel. Averr. idg. *qetubres: skr. catvdras, four, goth. fidwor , cf. gr. Tso?. idg. *-qiti-s: skr. dpa-citi-s, reward, homage, punishment, gr. xiro-Tiiri/;, rhig. idg. *qid: skr. cw£, -cunque, gr. ri , lat. grn^. idg. *hiki-s: skr. p^ci-s, light, clear, pure, cf. guJcrd-s, clear, bright. idg. *hieuetai: skr. cydvate, moves, falls, cf. gr. aiva. Instances of gutturals, which were not followed by e, i, i and therefore remained unchanged: idg. *JcakM-: skr. kaMt (d), kakudman (nt), top, summit, lat. cacumen (from 'cacwdmen). idg. *karh- : skr. harhds , harhatc^s, crab , gr. Kctpxtvo? , lat. ccmcer (from *carcer or *carcen). SYSTEM OF CONSONANTS: GUTTURAL EXPLOSIVES. 65 idg. *skandi: skr. skdndami , I spring, lat. soando. idg. *kalo-s: skr. kala-s, dark, black, cf. gr. xt^^ig (*x«a/j), spot, j«jA«s (*x«A«?), „v6(pi^ KvvSpoq kx) x el f* e P lv *1 jftip»" ', lat. caligo , mist, darkness. idg. *karw-s: ved. karw-s, praiser, singer, cf. gr. wpuZ, dor. xxpu!;. idg. *qasetai: skr. kdsate, coughs, cf. ags. hwosta, dutch ^oe*^. That the radical vowel was a (not o), Is seen from lith. Jcoseti , to cough. idg. *e i gat: skr. o^, went, gr. 2/3>?, dor. 2/35. idg. *qotero-s: skr. katard-s, who (from two), gr. iroTepog, ion. xSrspog , goth. Awapar. idg. *k6kso-s, *kd/csa: skr. kdksa-s, region of the girth, girdle &c, underwood, kdksa, girdle, circular wall, lat. coxa, mhg. hahse. idg. *jug6-m : skr. yugd-m , yoke , age , gr. £uyoV, lat. jugum, goth. ./w£. idg. * phonos: skr. ghand-s, slayer (ved.), compact w , forest , oslav. gora, mountain. idg. * < gmt6-s : skr. gatd-s, gone , gr. @xto$ , lat. -ventns. idg. *isqr-t: skr. ya>£r£ (*yakrt, cf. avest. yakare), liver, gr. jjJTras^, lat. jecur. idg. *kreu9s: red. krawis, raw flesh, gr. xpsx?, cf. lat. cruor and ags. ^raz , icel. /jrar. idg. *kukr6-s, *Tcukl6-s : skr. gukrd-s (very common in Vedic) , 5 66 MANUAL OF SANSKRIT PHONETICS. pukld-s, clear, bright, white, cf. comparat. qdcvyan (ms) and guci-s, light, clear, pure. idg. *ugro-s: skr. ugrd-s, mighty, terrible, cf. comparat. djiyan (ms), akin to lat. augeo , goth. aukan &c. idg. *stighnutai\ skr. stighnute , ascends (Dhatup. askand ane) , cf. gr. trtsixa, goth. steiga. idg. *penqtw : skr. pankti-s, row of five, row, cf. gr. 7riftirT0s. idg. *siseqti: ved. sisakli, follows, accompanies, cf. sdcate, gr. eireTou , lat. sequitur. § 54. Idg. q, k. In Aryan the idg. £ and k have fallen together. Before palatal vowels and the semivowel i they were palatalized, in all other cases they are represented by k. Instances of idg. q, k, skr. k: idg. *qdsetai: skr. kdsate , coughs, cf. ags. hwosta. idg. *qotero-s : skr. katard-s, who (from two) , gr. irorepoq , ion. xorepoi;, goth. hwapar. idg. *iiqr-t : skr. ydkrt (*ydkrt) , gr. iJ?r#/> , lat. jecur. idg. *penqti-s: skr. panktt-s, row of five, row, cf. gr. vs/jlvto?. idg. *kakud-: skr. faM (^), kakudman (nt), top, summit, lat. cacumen. idg. *kark-: skr. karkd-s, karkata-s, crab, gr. xxpxivog , lat. idg. *k6kso-s, *k6ksa; skr. kdksa-s, region of the girth &c, kdksa, girdle, circular wall, lat. coaw, mhg. hahse. idg. *kreuss: ved. kravis, raw flesh, gr. xpixc, cf. lat. «-«o?- &c. Instances of idg. y , £ , skr. e : idg. *qerus: skr. carw-s , kettle, pot, icel. hverr. idg. *qeiu6res : skr. catvdras, four, goth. Jidwor, cf. gr. Thsotpsc, lat. quatuor. idg. *penqe: skr. pdhca, five, gr. a-fvre, lat. quinque > goth. j£w/. SYSTEM OP CONSONANTS: GUTTURAL EXPLOSIVES. 67 idg. *gid: skr. cid, -cunque, gr. ri , lat. idg. *leuketai: skr. rdcate, shines, pleases, cf. gr. Xsuko? , lat. fe?i(, goth. liuhap. idg. *Jcuki-s: skr. ?««-«, light, clear, pure, cf. cukrd-s, bright, light, clear. idg. *Meuetai: skr. cydvate, moves, falls, cf. gr. asiiai. § 55. Idg. qh, kh. Certain examples of idg. qh, kh are very rare : idg. "sqhalb: skr. skhdlami, I stumble, cf. gr. tr$#Mw (*sqhlio)1 idg. *konMo-s: skr. gankhd-s, shell, gr. Koy%oc , cf. lat. Instances of skr. M, iran. #: skr. Mara- , ass , avest. x ara -- skr. Hw, well, avest. #«»-, %ao. skr. sdJcha, friend, avest. ha%a. In skr. Jcumbhd-, pot, the initial # was originally aspirated, cf. avest. %umba-. The loss of the h was caused by dissimilation. Concerning the combination cch (ch) it is not yet certain, whether it should be explained from idg. shh before palatal vowels or from idg. shh. In the former case gdcchami , icchdmi, prcchdmi would have cch by analogy, because idg. * c gmshhd, *isshh6, *prhshhd would have given *gashhami, *ishhami , *jorslchami ; the cch of gdcchati, icchdti, prcchdti Tegvl&Tlj -woxlIA. represent the idg. shh before idg. e : idg. * < gm,shhiti , Hsshheti , *prhshheti. In the latter case gdcchami, icchdmi, prcchdmi also would be organic forms. Cf. on this question Zubaty, Kuhn's Zeitschr. 31, 9 sqq. and Bartholomae, Studien zur idg. sprach- geschichte 2, 3 sqq. Instances: skr. chinddmi , I cut off, cf. avest. sid- (s from sh), gr. asks , cf. avest. peresaiti (s from «£) , lat. posed, ohg. forscon. § 56. Idg. ^, g. In Aryan the idg. j and y have fallen together. Before palatal vowels and the semivowel * they were palatalized, in all other cases they are represented by g. Instances of idg. g, g, skr. g: idg. * c g6u-s: skr. gdu-s , cow, gr. /Sou? &c. idg. * ! tpru-s: skr. guru-s, heavy, cf. gr. (3xpti<; , goth. kaurus (and lat. gravis). idg. *egat: skr. ayatf, went, gr. c(3>i. idg. * c gmt6-s: skr. gatd-s, gone, gr. @xtoc , lat. -ventus. idg. *jugo-m: skr. yugd-m, yoke, age, gr. %vydv, lat. jugum, goth. y&£. idg. *ugrd-s: skr. ugrd-s, mighty, terrible, cf. lat. auged , goth. aukan. Instances of idg. j, ^, skr. /: idg. * c geretai : ved. jdrate, crackles , invokes, cf. obg. quirit. idg. *auges-: skr. o)'a*-, strength, power, cf. ugrd-s, mighty, terrible, lat, auged, goth. aukan. % 57. Idg. jh, gh. In Aryan the idg. jA and gh have fallen together. Before palatal vowels and the semivowel * they became jh, in all other cases they are represented by gh. In Indian jh became h. Instances of idg. gh , gh , skr. gh : idg. * phonos: skr. g hands , slayer (ved.), compact &c, gr. : skr. mrjami, I rub, wipe, strip, cf. gr. xn.h'Kya, lat. mulgeo, ohg. milcAu &c. Lith. melzu, oslav. mVazq, prove, that we have $. The idg. combination ^S(^) regularly became in Aryan £6(£), from which skr. db(A): cf. § 63. Example: idg. *uigbMs: skr. vidbMs, instrum. pi. of vig-, nom. »«£, settlement, community, clan, tribe, people, cf. gr. olxog, lat. vicus &c. § 61. Idg. gh. During the Aryan period the idg. gh became %A. This aspirated spirant is in general represented by skr. A. The intermediate jA is preserved in skr. ujjAitd-s, left , from *ud-jMtars : jdhami , I leave (ujjAami has been formed by analogy). Instances of idg. gh , skr. A : idg. *gAeimen-: ved. Mman- , winter, gr. %tiiL», xeifii&v, cf. skr. Aimd-m, snow, gr. %ic!iv, lat. Aiems. The gA appears from lith. zema, oslav. zima &c. idg. *§Aans6-s: skr. Aamsds, goose, swan, cf. gr. %viv, lat. cmser, ohg. gans. Lith. %qsis proves, that we have gA ; surprising is the g of oslav. gqsz (perhaps an ancient loan from Germanic). SYSTEM OF CONSONANTS: SPIRANTS. 73 idg. *lei§hmi: skr. lehmi, I lick, cf. gr. >>ei%u, lat. lingo, goth. -laigon. The gh appears from lith. leMu, oslav. li%q&c. idg. *bhagh7> fy 17 ^ idg. *ueqsio: skr. vaksyami, I shall say: w«c»m, I say, cf. vdk, voice, lat. vox &c. idg. *dheksip: skr. dhaJcsyami, I shall burn: ddkami, I burn, lith. <#e^w , cf. goth. dags. idg. *uehi: skr. »«/£s«, thou art willing, cf. vdgmi, I am willing, gr. kxcbv. It is to be noticed, that skr. s was turned back into *, when immediately followed by r or r: idg. *tisres, *iisrbhis &c. : skr. tisrds, fern, three, instrum. tisfbhis , cf. avest. tisaro , irish feom. Brugmann 1, 412 and Bartholomae (Studien zur idg. sprach- geschichte 2, 5) assume, that the idg. sq, sh before palatal Towels became go in Aryan, cf. ved. paged, behind, later, west, avest. pasca (originally an instrumental, ending in idg. e). Against this opinion Meillet, Mem. de la Soc. de Ling. 9, 375 sq. It would seem, that the idg. sqh, shh before palatal vowels are represented in Indian by eeh {ch) , but Bartholomae (Stu- dien zur idg. spracbgeschichte 2, 3 sqq.) has tried to prove, that eeh {ch) goes back on a combination of h. Of. § 55, where examples of cch are given. The idg. ss after «-vowels became ts in Indian. In most cases however ** had been simplified before the end of the Indogermanic period, cf. skr. dsi, thou art, avest. ahi, gr. el, idg. *esi from *es-si. Instances of idg. ss, skr. ts: idg. *vessio; skr. vatsydmi, I shall put on, gr. £, lat. ego, goth. ik. idg. *dhuydte{f)\ skr. <#«^a (r), daughter, gr. Svyxryp, cf. goth. dauhtar. idg. *guy-\ skr. guhami, I hide, cf. 'Xiyuy/a , "flyyyo? , ryyif?. idg. # »ey-: skr. mahan (nt), great, cf. ved. mdhi, nom. ace. neutr. great , adv. very, subst. greatness, gr. iteyx (idg. *meya), § 67. Idg. p , S. In general skr. ks corresponds to gr. £, but in some cases we find gr. nr, %9 or (p$ instead of £. We must assume, that the skr. s, which is represented by gr. t, S, goes back on dental sounds, which were different from s and from the ordinary idg. dental explosives. Perhaps these dental sounds were spirants (p , d). See Brugmann 1, 409 sq. and Kretschmer, Kuhn's Zeitschr. 31, 429 sqq. Instances of skr. ks , gr. xr, idg. hp, kp : ved. kse'mi, I dwell, skr. ksetra-m, field, ksiti-s , dwelling, abode , earth , ved. pi. tribes, gr. rhod. xrolvx, dwelling-place , gr. xrl^a, KTiirit; , xy,(piKTloveq , ivtcTifievcii;. Avest. saeiti proves, that the initial guttural was k. ved. ksdyami, I possess, gr. xtxo/axi, Krij/^x. Avest. %sa- yeiti proves, that the original anlaut was k. skr. ksandmi , I hurt, gr. XTeiva. Opers. -%sata- = skr. ksatd-. skr. fksctrs, bear, gr. xpxrog, lat. ursus. Avest. arMa-(t). skr. tdksa (n), carpenter, gr. tcktuv, cf. lat. texere. Avest. tas-. Instance of skr. ks , gr. 06 , idg. qph : skr. ksindmi (ved.) , later ksinomi, ksdyami, I destroy, ksiti-s , destruction, gr. (pSlvu, (pShig. Avest. %$aya- proves, that 80 MANUAL OF 8ANSKRIT PHONETICS. skr. h , gr. cf. gr. #«#«/, lat. humus, avest. zem-, lith. %eme , oslav. zemlja, which prove, that skr. ks, gr. %b here represents an idg. combination of voiced consonants. skr. ksdrdmi, I flow, melt away, perish, gr. Cpdelpa. Avest. yzaraiti proves, that the initial group of consonants was originally voiced. B. The relation of the Indian consonants to THE iNDOGERMANIC. Semivowels. § 68.. Skr. y. In skr. y have fallen together the idg. i and j. Skr. y = idg. i: skr. yds, who, gr. a?. skr. ydkrt , liver , originally *ydkrt , gr. yirxp , lat. jecur. skr. trdyas , three , gr. rpeTg , lat. tres, goth. preis. skr. mddhyarS, middle , gr. iisgos , lat. medius, goth. midjis. skr. j^d, bowstring, cf. gr. /3/o'$. Skr. y — idg. ^ : skr. yugd-m , yoke , gr. tpyov, lat. jugum , goth. _/«£. skr. ydsami (ydsyami), I seethe, I boil, gr. £««, cf. ohg. skr. ydvas, corn , barley, cf. gr. £««/ , lith. _/««><«*. skr. tydjami, I leave, cf. gr. , lat. is-tum, goth. pan-a. §71. Skr. n. Skr. » = idg. w: skr. %d«-s , ship , gr. c«u« , cf. lat. «««»'*, icel. » °hg- ga>ns and lat. cmser. skr. andmca, perf. of agnSmi, I attain, dmga-s, portion, part, cf. gr. vivsyKov, lat. nanciseor. ved. drhhas, distress, lat. angus- in angustus, cf. gr. «y#w, lat. ango and goth. aggwus. Liquids. § 76. Skr. r. Skr. r regularly corresponds to idg. r, but in many cases it goes back on idg. I. Skr. r = idg. r: skr. rdjyami, I am (get) red, affected with a strong feeling, gr. #?». skr. rd-s, riches, wealth, lat. ras. skr. rat (*rak from *rafa) , king, lat. ra». skr. bhdrami, I bear, gr. , lat. &m#<5, goth. -laigon. About the relation of the Vedic dialect to Sanskrit (the Brahmana's and the epic poems included) see § 43. § 77. Skr. 1. Skr. I regularly corresponds to idg. I: skr. Idmbate , hangs down, cf. lat. labor. skr. Ukmi, I lick, cf. gr. Ae/#« &c. skr. limpdmi, I smear, cf. gr. >>iiro<; , hnt«.pi<; &c. skr. Mbhyami, I desire, cf. lat. lubet , goth. livfs. skr. dlpa-s, small, cf. lith. alpnas, weak. But we often find skr. I = idg. r (cf. § 42): skr. luncami, I pull out, I pull off &c, cf. gr. 3puif»« and ohg. feim. skr. phalgu-s, weak, worthless, useless, cf. gr. (peXyuvsr xirvvereT, Xviptt (Hesych.). skr. gaphd-s, hoof, cf. ohg. huof. § 80. Skr. b. Skr. b = idg. b: skr. bdla-m , strength , power , cf. oslav. bolij and lat. debilis. skr. buh-hara-s, roar, bellow, cf. gr. /3uktijs, fivxtxvy , lat. skr. Idmbate, hangs down, cf. lat. labitur. ved. upa-bdd-s, stamping, trampling, cf. gr. £V/j35«<. In the neighbourhood of a following aspirate b often repre- sents idg. bh: skr. bSdhami, I perceive, goth. biuda, cf. gr. irevSoficti, irvv&xvofixi. ved. budhnd-s, bottom, cf. gr. nuSfitjv, ?ru!/5fl:|f, lat. fundus and ohg. bodam. In younger texts we often meet with dialectic words, in which & is written instead of v. § 81. Skr. bh. Skr. bh = idg. bh : skr. bhdrami, I bear, gr. (pfjoa, l&t. fero , goth. Wra. skr. bhdvami, I am, cf. gr. , axivha.'hyt.&i;, lat. seindo. Before bh the skr. <# sometimes goes back on idg. z: ved. madbMs , instr. pi. of mds, moon, month. Skr. ed in some cases represents ar. azd (see § 63) : ved. sedydt , 3 pers. sing, optat. perf. of sad-, to sit , avest. hazdyap. In the neighbourhood of a following aspirate d often goes back on idg. dh: skr. dddhami, I put, gr. t16w*i. SYSTEM OF CONSONANTS: LINGUAL EXPLOSIVES. 87 § 85. Skr. dh. Skr. dh = idg. dh: skr. dhdrsami , I dare , cf. gr. dpxvvs , Sxppsa , goth. ga-dars. skr. dJivmd-s, smoke , vapor , gr. SvfiS? , lat. fumus. skr. dhdyami, I suck, I drink, goth. daddja, cf. gr. 9%4, <&&&, ^^ the skr. . Skr. &s = guttural + p (d~): see § 67. § 91. Skr. kh. Skr. kh = idg. M (q&): skr. cankhd-s, shell, gr. x.oy%oq. skr. skhdlami, I stumhle, cf. gr. o-fpaAAa? § 92. Skr. g. Skr. y corresponds to idg. j, ^. Skr. g = idg. j: skr. gdu-s, bull, cow, gr. /3ou?, lat. *vos (60s), ohg. cAwo. skr. guru-s, heavy, cf. gr. (3xpus, goth. tow-as (also lat. gravis). skr. d^atf, went, gr. ?/3»j. Skr. ^ = idg. g : skr. y«?«-?» , yoke , age , gr. £uyaV, lat. jugnm , goth. jul: skr. aprrf-*, mighty, terrible, cf. lat. auged, goth. aw/fo». In the neighbourhood of a following aspirate we often find g from ^h or ^A: skr. gfdhyami, I am eager, cf. goth. gredus. system or consonants: palatal aitkicates. 91 § 93. Skr. gh. Skr. gh goes back on idg. j/j or gh. Skr. gh = idg. gh: skr. ghand-s, slayer (ved.), compact &c, gr. Qovos , cf. ohg. ■ea, ags. g4d. Skr. ^A = idg. gh: skr. dirghd-s, long, cf. gr. ScA/^o'?. Palatal affricates. § 94. Skr. c. Skr. c in general represents idg. q, h before palatal Towels and i. Skr. c = idg. q : skr. ear4-s , pot, kettle, icel. Aew. skr. catvaras, four, goth. fidwor, cf. gr. rhvapet;, lat. quatuor. skr. joiwca , five , gr. srsi/Tf , lat. quinque, goth. < $»^. Skr. c = idg. Jc: skr. cydvate , moves, falls, cf. gr. . skr. rocate, shines, pleases, cf. gr. tevxii; , lat. to, feetf, goth. Uuhap. In cases as ucca- , high , the first c is an assimilated t (here originally af). § 95. Skr. ch. On the origin of skr. ch see§ 55. Instances: skr. chinddmi, I cut off, cf. avest. sid-, gr. strength, power, cf. ugrd-s, mighty, terrible, lat. augeo and goth. aukan. 92 MANUAL OF SANSKRIT PHONETICS. Skr. j = idg. g : ved. jdnas, race , family, gr. ykvos , lat. genus, cf. goth. Jeuni. skr. jtinu, knee, cf. gr. y&w, lat. genu, goth. Jcniu. skr. jusdmi, I take pleasure, I relish, cf. gr. e*'As. skr. ddgami, I bite, cf. gr. S«xi/». ved. «{»a («), stone, gr. «x/*ai/. skr. daddrga, I have seen, gr. Se'Sopx*. Skr. g = idg. *: SYSTEM OF CONSONANTS: SPIKANTS. 93 skr. gvdgura-s, father in law, gr. kxupig , cf. lat. socer and goth. swaihra. skr. gagas, hare, cf. ohg. haso. § 99. Skr. s. Skr. s lingualized from idg. s: skr. tisthami, I stand, gr. JVt^^;. skr. ustd-s, turned, lat. ustus. skr. vfJcesu, loc. pi. of vfJca-s, wolf, cf. oslav. vlttcecM. ved. jostar-, loving, cf. gr. yewrrvipiov. skr. dveksi, 2 pers. of dvesmi, I hate. skr. dhrsnSmi, dhdrsami, I dare, cf. gr. Sappiu and goth. In the combination sf the skr. s often represents idg. I: skr. asfd (ved.), awftw, eight, gr. o>cn$, lat. octfo, goth. ahtau. skr. disti-s, indication &c, cf. ohg. inziht, lat. diclio, gr. How the s of saf, six, is to be explained, is yet uncer- iain: probably an initial consonant has been lost, cf. avest. %svas. Skr. § = idg. p, ct: see § 67. § 100. Skr. s. Skr. * = idg. *: skr. saptd, seven, gr. !tt«, lat. septem, cf. goth. siiun. ved. sdna-s, old, gr. fein. swi , cf. lat. senece and goth. sineigs. skr. sdhas, strength, might, violence, goth. sigis. skr. svdsa, (r) , sister, lat. soror, goth. swistar. skr. srdvami, I flow, I stream, gr. piu, cf. ohg. strowm. skr. »&i!e , clothes one's self, puts on , cf. gr. svvupii , hdfc , lat. vestis, goth. wasjan. § 101. Skr. h. Skr. ^ may go back on idg. gh, gh be- fore palatal vowels and i, further on idg. gh and in dialectic words and forms on idg. bh and dh. Skr. h = idg. gh: 94 MANUAL Or SAN6KKIT PHONETICS. skr. Miami, I slay, cf. gr. 6eiva and skr. hatya, slaughter, ohg. gundea, ags. g4ct. ved. haras, heat, gr. Sipog. Skr. h = idg. gh: ved. druhas, harming spirits, cf. druhyami, I harm, ohg. triogan. Skr. h = idg. y^: ved. Aeman-, winter, gr. #«<>««, %eiyt.&v, cf. gr. ^(«v, lat. hiems &c. skr. hamsdrs, goose, swan &c, cf. gr. %yv, lat. anser, ohg. ^a»*» skr. bahu-s , arm, gr. jrifau?, ohg. d»o^. skr. Uhmi , I lick, cf. gr. *«'#», lat. £i»^o, goth. -laigon. Skr. ^ = idg. 5^: skr. Mrami , I bear, I take , cf. bMrarni , I bear, gr. Cpspa &c. skr. grhndmi, I seize, cf. ved. grbhnami. Skr. ^ = idg. dh : Skr. -^i , suffix of the 2 pers. sing, imperat. act., cf. -dhi ,. gr. -Si. skr. -make, -maM, suffixes of the 1 pers. pi. med., gr. -ftiSx. In some words the skr. h seems to correspond to gr. y, goth. h: see § 66. PART Til. SANDHI-SYSTEM. § 102. General remarks. The rules, which govern the changes of the initial and final sounds of words in the sen- tence and in composition, are called aemdAi-Tvlea. Of course already in the mother-language the final sounds of preceding- words and the initial sounds of following words had a mutual influence on each other, but it is impossible to reconstruct a complete tableau of this Indogermanic sandhi. Yet there are reasons to suppose , that already in the Indogermanic period the explosives and spirants were voiced or voiceless according to their position before voiced or voiceless explosives and spirants. The double forms gr. icpori = skr. prdti and gr. irpot; = skr. prdti/ may prove, that in the mother-language i before sonants became a semivowel, which suggests the supposition, that this was also the case with u. The simul- taneous existence of the Indian dual-endings -au = goth. -au (in ahtau) and -a = gr. -co allows the conclusion, that yet in a time of dialectic continuity the idg. diphthong -au before certain consonants or groups of consonants was simplified to -6. Gr. xxftav and skr. dgma, gr. noniip and skr. pita stand in a similar relation to each other. By the side of roots 96 MANUAL OF SANSKRIT PHONETICS. beginning with * we often find doublets without that initial consonant (skr. sthdgami: lat. tego, goth. stawtan'. skr. tuddmi) and hence we may conclude, that before the end of the Indo- germanic period the initial * was lost, when preceded by a word ending in *. About these and other phenomena, which are explained by Indogermanic sandhi, see Brugmann 1, 490 sqq. The Indian sandhi has gradually developed itself from that of the mother-language, because of which only a historical treatment of the sandhi-rules may claim a scientific value, but the uncertainty, which till now prevails in the history of the initial and final sounds, makes it preferable to arrange them systematically. § 103. Final and initial vowels. It is a rule in compo- sition as well as in the sentence , that -a -|- a- are contracted to a , -i -\- i- to i , -u -\- u- to ii. An a-vowel (-«) combines with an i-vowel (?-) to e, with an ««-vowel («-) to o, with r- to or, with e- and ai- to ai, with o- and au- to cm, but -?, -u and -f before a dissimilar vowel or diphthong are changed each to its corresponding consonant (j), v, r). The original, but contracted diphthongs -e and -o remain unchanged before a-, which however disappears; before all other vowels they become a (we seldom find -ay and -av, which of course are the regular representatives of the antevocalic ar. -ai and -au: see § 7). The diphthongs -ai and -au before all vowels may be retained unchanged (if so, they are written -ay, -av), but it is also permitted to change them to -a. About some liberty in Sanskrit as to the treatment of final and initial vowels see Whitney § 127 a, % 129 c, § 133 a, § 134 c and cf. Kern, Taal en letteren 6, 332. The peculia- rities of the Vedic vowelsandhi are noticed in Whitney's grammar, to which it will suffice to refer. SANDHI-SYSTEM. 97 Contraction of similar vowels: na -)- asti = nasti , not is. tatra -j- ayatah = tatrayatah , there arrived. gatva -\- abravit = gatvabravit , having gone he (she) said. raja -\- adadat = rajadadat, the king took, -f- iha = astiha , is here, -j- igvarah = adhigvarah, over-lord. devi -f- iyam = deviyam , this goddess. stri -|- iksate = striksate , the woman sees. su -f- uJctam = suktam, hymn. bahu- -\- uru- == bahwru-, arms and thighs. Combination of a, with dissimilar vowels : tistha + iha = tistheha, remain here. ratha- -\- isd = rathesd , shaft of a chariot. bharyd -f- «>a = bharyeva , as a wife. ha, -\- idrgi = kedrgi , who (fern.) is such a one. /Was- -(- ucchistam = hutocchistam , remains of a sacrifice. wvfco- + udara- = vrkodara-, who has a wolfs belly (epithet of Bhima). sahasa -f" utthaya = sahasoUhaya , hastily arisen. rambhd- -\- urw- = rambhoru- , who has thighs as a banana. fopfrt -f" rsayah = saptarsayah , the seven sages , the seven stars of the Great Bear. maha- -\- rsabhah = maharsabhah , a great (large) bull. «c« = lomaqa uvaca, Lornac^ said. 106 MANUAL OF SANSKRIT PHONETICS. Change of -as to -a before voiced consonants and before vowels : vihagas -\- day ante = vihaga day ante, the birds fly. naras -j- vadanti = tiara vadanti, the men say. balas -j- annam prechanti = bala annam prechanti, the children ask food. In particular we must notice the treatment of the pronouns sa, that, and esa, this. Though originally they do not end in -s, yet their pausa-form is sah, esah and before a- they appear as so, eso, after the -o of which forms the initial a- disappears. Examples : sa -f- dadarga = sa dadarga, he saw. sa -\- puman = sa puman, that man. sa -\- avadat = so'vadat , he said. rwurkhas -f- sa = mwrhhah sah, he is a fool. The exclamation bhos before vowels and before voiced con- sonants appears as. bho (instead of *bhor): the cause of this irregularity is the origin of the word, which is a contracted form of bhavas, voc. of bhavant-, your honor (pron. pers. used in reverent allocution). § 110. Treatment of -r. The sandhi-rules concerning -r are but partly the organic result of phonetic processes : much in the sandhi of -r is due to analogy. I shall simply give an account of the rules without trying to trace them to their origin. Before c- (ch-) the -r is turned into -g; before t- (tk-) into -s; before t- {th-) into -*. Before k (kA), p (j>A), before sibi- lants and in pausa the -r becomes visarga. Before an initial r- the final -r disappears after lengthening a preceding short vowel. In any other case -r remains unchanged. SANDHI-BTSTBM. 107 Substitution of sibilants for r before c- (ch-) , t-, t- : punar -\- cacara = punag cacara, again he wandered. nisedur -\- chaydyam = nisedug chayayam, they sat down in the shadow. pitur + tankas = pitus tankah, the father's pickaxe. ilciir -\- te = item te, they said. Change of -r to visarga before k (kh) , p {ph) , sibilants and in pausa (only in pausa the change seems to have been or- ganic, see Brugmann 1, 494): pwitar + karoti = punah karoti , again he (she) does. dadur -\- phalani = daduh phalani , they have given fruits. babhuvur -\- samaje = baihuvuh samaje , they were at the assembly. vanam -J- vicerur = vanam viceruh, they wandered through the forest. Change of -ur, -ar to -u, -a before r- : ucur -f- ramam = ucu ramam, they said to Rama. punar -\- raksati = puna raksati, again he protects. The unchanged -r before voiced consonants and before vowels : pitar -j- dehi = pitar dehi, father (voc), give. punar -\- jagama = punar jagarna , again he went. punar -\- aha = punar aha, again he said. dadur -f- annam = dadur annam, they gave food. § 111. Changes of initial consonants. In § 108 we saw, that g- with a preceding dental combines to cch. An initial ch- becomes cch-, when preceded by a short vowel, the pre- position (prefix) a or the prohibitive adverb ma. Finally we mention the rule, that an initial h- is changed to the aspirate of a preceding media, whether this media is original or a softened tenuis (see § 106). Change of ch- to cch-: 108 MANUAL OF SANSKRIT PHONETICS. ahara -)- chattram = ahara cchattram, bring the parasol. a -\- chaditas = acchaditah, covered. Change of h- to the aspirate of a preceding media : jprthah -\- haranti = prthag gharanti , they take separately. tad + hastas = taddhastah, his hand. PART IV. ACCENT. § 112. The accent of the Indogermanic period. Verner's law (Kuhn's Zeitschr. 23, 97 sqq.) has been an eyident proof of the fact, that the Indian stress, as it is handed down to us in some Vedic hooks and by ancient Indian grammarians, generally fell on the same syllables as in the Indogermanic mother-language. In Indian, Balto-Slavonic and Ur-Germanic the so-called free accent prevailed , i. e. that neither by the number of syllables nor by the quantity of any syllable the stress was bound to a certain place. Therefore we must assume , that the accent of the mother-language was also a free one. Concerning the accent of the sentence we may learn from the comparison of the Indogermanic languages, that already in a period of dialectic continuity some conjunctions, the inter- rogative pronouns, when used as indefinites, the personal pronouns, when not used emphatically or antithetically, in many cases also the vocatives and the verbum finitum in an independent clause were deprived of their stress by the pre- ceding word. Long before the separation of the Indogermanic dialects 110 MANUAL OP SANSKRIT PHONETICS. all vowels of so-called toneless syllables were weakened : Stra- chan (Bezz. Beitr. 14, 173 sqq.) and Kretschmer (Kubn's Zeitschr. 31, 325 sqq.) have shown, that this rule is not to be restricted to pretonic Towels. Many scholars therefore as- sume , that the Indogermanic accent of an early period must have been a strong expiratorical one (but cf. Finck, Ueber das verhaltnis des baltisch-slavischen nominalaccents zum ur- indogermanischen , Marburg 1895, 29 sqq.) and that it became less expiratorical or even musical towards the end of the proethnic period. Now there is no longer any doubt, but that the mother- language had two accents of different quality, one of which is represented by the Sanskrit udatta and the Greek o^elx, while the other has left its traces in the dissyllabic value of some long vowels in the Vedic dialect, the Greek irtpifiraitevii and the Lithuanian scMeifton. See Bezzenberger , Bezz. Beitr. 7, 66 sqq.; Hanssen, Kuhn's Zeitschr. 27, 612 sqq.; Hirt, Idg. forschungen 1, 1 sqq. 195 sqq. To Hirt we owe an excellent monography on all questions concerning the Indogermanic accentuation (Hirt, Der indoger- manische akzent, Strassburg 1895). § 113. Accentual agreement between Sanskrit and Greek. The liberty of accent , which prevailed in the mother-language, was restricted in Greek by many special rules (see Hirt, Der idg. akzent 24 sqq.). Yet there are many words — dissyl- labic and trisyllabic — , which have the same accent in Indian and in Greek. A few examples will suffice: skr. padds, gr. iroSoi; , gen. of skr. pat, foot, gr. dor. sr««. skr. trdyas, three , gr. rpelq , but loc. skr. trim, cf. gr. rpivt. skr. pdhca , five, gr. irevTs , goth. fimf. ACCENT. Ill skr. ddga, ten, gr. $ixx, goth. skr. saptd, seven, gr. Ittt«. skr. astd(u), eight, gr. iwr&. skr. fksa-s, bear, gr. apxrog. skr. jdmbha-s, set of teeth , gr. yofiCpog. ved. ddma-s, house, gr. Mfiog. skr. dgva-s, horse , gr. "wot; , goth. aihwa-. skr. iirdhvd-s, high , cf. gr. o'/j^s. Ted. midhd-{m), prize, contest, gr. fuvM-tg). skr. dkwmd-s, smoke, gr. &v(j,6c. skr. yajhd-s, sacrifice, gn dyvSg. skr. grutd-s, heard, heard of, gr. Ktorog. skr. jnatd-s, known , gr. yvarog. skr. gwr4-s, heavy, gr. (Zxpvg. skr. svadu-s, sweet, gr. $$ug. skr. mddku, honey, gr. peSv. skr. ndbhas, sky, gr. vsCpcg. ved. jdnas, race , gr. yevog. skr. mdnas, mind , gr. (tevog. skr. dndhas, herb, gr. &v6og. skr. grdvas, fame, gr.- xAe'e?. skr. sddas, seat, gr. ?5o?. ved. bhdrma (n), bearing, gr. (pippx. ved. &w« («), pouring, oblation, gr. x e ^/* a ' ved. vdsma (»), cover, gr. lesb. reppx. skr. _p»tfd , father , gr. ttxtiIip , goth. fadar. skr. <&?wa, brother in law, gr. S«^. skr. bhrdta, brother, gr. (ppxrup, goth. bropar. skr. svdsa, sister, cf. gr. ?o/)«$. In trisyllabic words and forms: skr. uttara-s, further, left &c, gr. v/rrepog. 112 MANUAL SANSKRIT OT PHONETICS. skr. duhitar, gr. ivyxrep, voc. of skr. duhitd. daughter, gr. iuyxTvip. skr. svddiyas neutr., sweeter, gr. ij'S/ov. ved. astdpat, with eight feet or parts, gr. tttrdirovs. skr. pitdras, fathers, gr. Trxrepes. skr. janitd, father, gr. ycverijp. ved. isird-s, strong, quick &c, gr. )sp6g. skr. hahuld-s, numerous