Pramāṇa · Search · About · API

Indian Palaeography from about B.C. 350 to about A.D. 1300 c8: V. The Southern Alphabets

Indian Palaeography from about B.C. 350 to about A.D. 1300 c8

Language
English

Contents

V. The Southern Alphabets

§ 27. — Definition and varieties

c8.s1.1

[60] With BURNELL and FLEET, I understand by the term "southern alphabets" the scripts of plates VII. and VIII,¹ which, developed out of the characters of the Andhra period, have been generally used since about A. D. 350 in the territories south of the Vindhya, and most of which still survive in the modern alphabets of the Dravidian districts.

c8.s1.2

¹ Preparation of Plates VII. and VIII: —

c8.s1.3

Their most important common characteristics are: —

c8.s1.4

(1) The retention of the ancient forms, open at the top, of *gha*, *pa*, *pha*, *ṣa* and *sa*, of the old *ma*, and of the tripartite *ya* which is looped only occasionally, especially in the Grantha.

c8.s1.5

(2) The retention of the long stroke on the right of *la*, which however is mostly bent towards the left.

c8.s1.6

(3) The *ḍa* with the round back.

c8.s1.7

(4) The curves, originally open at the top, at the ends of the long verticals of *A*, *Ā*, *ka*, *ña* and *ra*, as well as of the subscript *ra* and of medial *u* and *ū*.

c8.s1.8

(5) The medial *ṛ* with a curled curve on the left, with occasional exceptions occurring in *kṛ*.

c8.s1.9

PLATE VII.

c8.s1.10

Cuttings from facsimiles.

c8.s1.11

Col. I; from FLEET, Gupta Inscriptions (CII. 3), No. 5, pl. 3 B; with *E* from No. 62, pl. 38, B. Cols. II, III; from F.GI (CII. 3), No. 18, pl. 11. Col. IV; from plate at IA. 7, 66. Col. V; from plate at IA. 5, 205; with *A*, *Ā*, *U*, *ghā*, *dhau*, *hā*, *kṣa*, *ttā*, from plate at IA. 6, 9, and *nta* from plate at IA. 7, 68. Col. VI; from F.GI (CII. 3), No. 38, pl. 24. Col. VII; from F.GI (CII. 3), No. 39, pl. 25. Col. VIII; from plate at EI. 2, 20, No. 1; with *I*, *ña*, *ba*, *ñca*, *brā*, *lya*, from No. 3, at p. 22. Col. IX; from plate at IA. 13, 78. Col. X; from F.GI (CII. 3), No. 55, pl. 34; with *U* and *AU* from No. 41, pl. 27, and *Ū* from Ajaṇṭā No. 3, B.ASRWI. 4, pl. 57. Col. XI; from F.GI (CII. 3), No. 56, pl. 35. Col. XII; from plate at IA. 7, 35. Col. XIII; from plate at IA. 7, 37; with *I*, *ṅśa*, *jye*, *ṇām*, *tsā*, from plate at IA. 6, 24. Col. XIV; from plate at IA. 10, 58; with *Ā*, *U*, and *ccha* from plates at IA. 7, 161, and *kḷ* from plate at IA. 6, 72, and *ḷa* from plate at IA. 8, 44. Col. XV; from plate at IA. 10, 104, FLEET's No. 94; with *Ī* (3, XV, *b*), *ṅge*, *śi*, and *ḷi* from FLEET's Nos. 99, 100, plate at IA. 10, 164, and *ḻa* from FLEET's No. 95, plate at IA. 10, 104. Col. XVI; from plates at IA. 8, 24 ff. Col. XVII; from plate at IA. 13, 137. Col. XVIII; from plates at IA. 8, 320. Col. XIX; from plate at IA. 13, 123. Col. XX; from plates at IA. 5, 50 ff. Col. XXI; from plates at IA. 5, 154 ff. Col. XXII; from HULTZSCH'S SII. 2, pl. 10. Col. XXIII; from HULTZSCH'S SII. 2, pl. 9. Col. XXIV; from HULTZSCH'S SII. 2, pl. 11.

c8.s1.12

PLATE VIII.

c8.s1.13

Cuttings from facsimiles.

c8.s1.14

Col. I; from plates at IA. 12, 158 ff. Col. II; from plate at IA. 11, 126, FLEET's No. 123. Col. III; from plates at IA. 12, 11. Col. IV; from plates at IA. 13, 186 ff. Col. V; from plates at IA. 7, 16. Col. VI; from plates at IA. 14, 50 ff. Col. VII; from plate at IA. 6, 138; with *A*, *U*, *cā*, and *ṭṭa* from plate at IA. 9, 75. Col. VIII; from plates at IA. 11, 12 ff. Col. IX; from plate at EI. 3, 62. Col. X; from plate at IA. 13, 275. Col. XI; from plate at IA. 18, 144. Col. XII; from plate at EI. 3, 18. Col. XIII; from HULTZSCH'S SII. 2, pl. 13. Col. XIV; from plate at EI. 3, 76. Col. XV; from plate at EI. 3, 14. Col. XVI; from HULTZSCH'S SII. 2, pl. 12. Cols. XVII, XVIII; from HULTZSCH'S SII. 2,. pl. 4. Cols. XIX, XX; from plate at EI. 3, 72, the lower part. Cols. XXI, XXII; from plate at EI. 3, 72, the upper part.

c8.s1.15

According to other peculiarities, the southern alphabets may be divided into the following varieties:¹—

c8.s1.16

¹ Compare B, ESIP. 14.

c8.s1.17

(1) The western variety, which, being strongly influenced by the northern alphabets, is the ruling script between about A. D. 400 and about A. D. 900 in Kāṭhiāvāḍ, Gujarāt, the western portion of the Marāṭhā districts, *i. e.* the Collectorates of Nāsik, Khāndesh and Sātārā, in the part of Haidarābād (Ajaṇṭā) contiguous to Khāndesh, and in the Koṅkaṇ, and which during the 5th century occasionally occurs also in Rājputāna and the Central Indian Agency, but altogether disappears in the 9th century in consequence of the inroads of the Nāgarī alphabet (see above, § 21).

c8.s1.18

(2) The Central-Indian script, which in its simplest form closely agrees with the western variety, but in its more developed form, the so-called "box-headed alphabet," shows greater differences, and which from the end of the 4th century is common in northern Haidarābād, the Central Provinces and parts of the Central-Indian Agency (Bundelkhaṇḍ), but appears also occasionally further south in the Bombay Presidency and even in Maisūr.

c8.s1.19

(3) The script of the Kanarese and Telugu districts of the Dekhan, — *i. e.* of the southern portion of the Bombay Presidency (the Southern Marāṭhā States, Sholāpur, Bījāpur, Beḷgaum, Dhārwār and Kārwār), of the southern territory of Haidarābād (roughly speaking south of Bīdar), of Maisūr, and of the north-east portion of the Madras Presidency (Vizagapatam, Godāvarī, Kistna, Karnūl, Bellary, Anantpur, Cuddapah, Nellore), — which appears first in the Kadamba inscriptions of the 5th and 6th centuries, and after a long development leads to the very similar and temporarily identical Kanarese and Telugu round-hand.

c8.s1.20

(4) The later Kaliṅga alphabet of the north-eastern coast of the Madras Presidency between Cicacole and the frontier of Orissa (Gañjām), which is strongly mixed with northern letters and in later times also with Grantha and Kanarese-Telugu characters, and which occurs in inscriptions of the 7th—12th centuries.

c8.s1.21

(5) The Grantha alphabet of the eastern coast of Madras, south of Pulikat (North and South Arcot, Salem, Trichinopoli, Madura and Tinnevelli), which first appears in the ancient Sanskrit inscriptions of the Pallava dynasties, and survives in the modern Grantha and its varieties, the Malayāḷam and the Tuḷu.

c8.s1.22

The Tamil alphabet of the same districts and of the western coast of Madras (Malabar) probably is derived from a northern script, imported in the 4th or 5th century, but greatly modified by the influence of the Grantha. A cursive variety of the Tamil alphabet is found in the Vaṭṭeḷuttu (the "round-hand," Burnell) or Cera-Pāṇḍya (Hultzsch),² which is known through inscriptions from the western coast and the extreme south of the -Peninsula, and according to Burnell [**61**] has fallen into disuse only in recent times.³ Though these two alphabets come from a different source, they have been included in this chapter, because they occur in the same districts as the other five.

c8.s1.23

² IA. 20, 286.

c8.s1.24

³ B, ESIP. 48.

§ 28. — The western script and the script of Central India; Plates VII, and VIII

A. — The western script

c8.s2.t1.1

The western variety of the southern alphabets is found in the inscriptions of the Imperial Guptas and their vassals since the time of Candragupta II.,⁴ of the kings of Valabhī,⁵ of the Gurjaras of Broach,⁶ of some of the Calukyas of Bādāmi (Pulakeśin II. and Vijayabhaṭṭā- rikā), and of Nāsik and Gujarāt and their vassals,¹ of the Traikūṭakas,² of the Aśmakas (?) of Khāndesh,³ and of the Rāṣṭrakūṭas of Gujarāt,⁴ as well as in numerous votive inscriptions in the caves of Kaṇheri, Nāsik and Ajaṇṭā.⁵ Ordinarily, its characters no doubt were written with ink, just like those of the northern alphabets (see above, § 21). This is made highly probable by the use of wedges on the tops of the letters during the Gupta period (see plate VII, cols. I—III) and by the thick, frequently knob-like, heads of the signs of the Valabhī, Gurjara and Rāṣṭrakūṭa grants (plate VII, cols. IV—IX, and plate VIII, col. I), both of which ornaments can only be drawn with ink. Another argument is furnished by the fact that all the copper-plates from Gujarāt have been cut according to the ordinary size of the Bhūrja leaves (BURNELL), on which it is not possible to write with a *stilus*.

c8.s2.t1.2

⁴ Compare the facsimiles in Fleet's Gupta Inscriptions (CII. 3), Nos. 5, 14, and 62, plates 3 B, 8, 38 B, and Fleet's remarks.

c8.s2.t1.3

⁵ Compare the facsimiles in F. GI (CII. 3), Nos. 38, 39, plates 24, 25; IA. 1, 17; 5, 204 ff.; 6, 14 ff.; 7, 65 ff.; 8, 302; 9, 238; 14, 328; J. BBRAS. 11, 363; EI. 3, 320.

c8.s2.t1.4

⁶ Compare the facsimiles at J. RAS. 1865, 247; IA. 13, 78; [7, 62; 13, 116; 17, 200; disputed]; EI. 2, 19 ff.

c8.s2.t1.5

¹ Compare the facsimiles at EI. 3, 52; IA. 7, 164; 8, 45; 9, 124; J.BBRAS. 16, 1; Seventh Oriental Congress, Aryan Section, 238; IA. 19, 310.

c8.s2.t1.6

² Compare the facsimiles at B.ASRWI. No. 10, 58.

c8.s2.t1.7

³ Compare the facsimile at IA. 16, 98.

c8.s2.t1.8

⁴ Compare the facsimiles at IA. 12, 158; J.BBRAS. 16, 105; EI. 3, 56.

c8.s2.t1.9

⁵ Compare the facsimiles at B.ASRWI. 4, pl. 55, 9; pl. 58, 5 and 9; plates 59, 60; vol. 5, pl. 51, 6—9. Compare the facsimiles in F.GI. (CII. 3), No. 6, 17, 61, plates 4 A, 10, 38 A.

c8.s2.t1.10

The finds of nearly or quite contemporaneous inscriptions with northern characters in Rājputāna, the Central-Indian Agency,⁶ and Valabhī, as well as the Nāgarī signatures of the Gurjara princes,⁷ prove that northern scripts were being used simultaneously with this southern alphabet. And this circumstance is no doubt the cause of its showing traces of northern peculiarities in the following letters: — (1) in the *kha* with a large loop and a small hook (plate VII, 9, I—IX; VIII, 12, I), instead of which the true southern form appears only very rarely;⁸ (2) in the *ca*, rounded off on the right (plate VII, 13, I—IX; VIII, 16, I); (3) in the ancient *ta* without a loop (plate VII, 22, I—IX; VIII, 25, I); (4) in the narrow *dha* (plate VII, 25, I—IX; VIII, 28, I; compare plate IV, 25, I—III); (5) in the looped *na* (plate VII, 26, I—IX; VIII, 29, I), which agrees more exactly with the northern forms of plate IV, 26, than with the southern one of VII, 26, XIII (compare below, § 29, A); (6) in the Mātrās often placed above the line in medial *e* (plate VII, 26, V), *ai* (plate VII, 10, IV) and *ō* (plate VIII, 35, I), which latter, however, has a peculiar looped form in *lō* (plate VII, 34, III, IV); (7) in the medial *au*, consisting of three strokes above the line (VII, 25, V; 36, III); and compare plate IV, 7, IV); (8) in the subscript *ña*, which occasionally, as in plate VII, 42, VII, shows the northern cursive form. The inscriptions Nos. 17 and 62 of FLEET'S Gupta Inscriptions (CII. 3), plates 10, 38 B, which are not represented in plate VII, show, [62] besides, the northern *A* and *ka* without the curve at the foot. A *ka* of this description occurs also sometimes in the Valabhī inscriptions (plate VII, 8, V).

c8.s2.t1.11

⁷ Compare above, § 21 end.

c8.s2.t1.12

⁸ Compare, for instance, *likhitam*, facsimile at IA. 7, 72

c8.s2.t1.13

Irrespective of these northern peculiarities, which throughout remain almost unchanged, the characters of this script show three stages in their development, that of the 5th century (plate VII, cols. I—III), that of the 6th and 7th centuries (cols. IV—VI, VIII), and that of the 8th (col. IX) and 9th centuries (plate VIII, col. I) which last is very markedly cursive.

c8.s2.t1.14

Among the single letters the following deserve special remarks: —

c8.s2.t1.15

(1) The *I* (plate VII, 3, IV, ff.; VIII, 3, I), which here, as in most southern alphabets, consists of a curved line with a notch in the centre and of two dots below, and which appears to be a modification of a form like that in plate IV, 3, IX.

c8.s2.t1.16

(2) The *Ī* (plate VII, 3, I; VIII, 4, I), which, like that of the Bower MS. (plate VI, 4, I), has been developed by the transformation of two dots into a line, but in addition has the curved tail, characteristic of the southern alphabets.

c8.s2.t1.17

(3) The *E*, which usually consists of a triangle with the apex at the top, and is irregularly broadened on the left (plate VII, 6, I; and compare *AI* in VII, 6, VII), and which from the end of the 6th century frequently, especially in Gurjara inscriptions, is opened at the top (plate VII, 6, VI) and finally resembles a northern *la* (plate VIII, 8, I).

c8.s2.t1.18

(4) The *ḍa*, which in its oldest form (plate VII, 19, II), as mostly in the southern alphabets, is undistinguishable from *da*, but from the 6th century develops a little tail (plate VII, 19, IV—IX), or, in some inscriptions of the 8th and 9th centuries, a loop at the end (plate VII, 43, VII; plate VIII, 22, I).

c8.s2.t1.19

(5) The *tha* with a ringlet on the base-line (plate VII, 23, III, IV, VI) instead of the cross-bar (plate VII, 23, I, II), developed out of the ancient dot, or since the end of the 6th century with the southern notch in the base (plate VII, 23, VII—IX; plate VIII, 26, I).¹

c8.s2.t1.20

¹ Transitional forms occur in the Calukya inscriptions.

c8.s2.t1.21

(6) The *la* with the diminutive main portion of the original sign and the enormous tail (plate VII, 34, VI, VIII), which latter since the 7th century frequently becomes the sole representative of the letter (plate VII, 34, VII, IX).

c8.s2.t1.22

(7) The *śa*, which shows regularly in the Gurjara inscriptions (plate VIII, 39, I) and the Nāsik Calukya inscription,² and occasionally in the Valabhī inscriptions,³ a cursive combination of the cross-bar with the vertical on the right, which occurs also in the north.⁴

c8.s2.t1.23

² Compare facsimile at *IA.* 9, 124.

c8.s2.t1.24

³ Compare *IA.* 6, 10, and facsimile at 14, 328.

c8.s2.t1.25

⁴ Compare facsimile at *J.ASB.* 64, 1, plate 9, No. 2.

c8.s2.t1.26

(8) The *sa*, which occasionally shows (plate VII, 38, V) a cursive combination of the left limb with the *Serif* occurring also in southern scripts (plate VIII, 41, XI).

c8.s2.t1.27

(9) A number of cursive forms in ligatures, thus: — (*a*) The prefixed *ña* which often loses the hook on the right and looks like *ṇa* (compare also plate V, 19, V, VII. (*b*) The prefixed *na*, which especially before *ta*, *tha*, *dha* and *na* (see the *nta* of *anumantavyaḥ*, plate VII, 42, V) consists of a horizontal or bent stroke and looks like *ta*.¹⁶ (*c*) The subscript *ka*, which occasionally, as in *ṣka* (plate VII, 46, VIII), is looped on the left (compare *IA.* 11, 305). (*d*) The subscript *ca* of *ñca* (plate VII, 41, VIII, IX), which since the 6th century remains open on the right and bears the hook of *ña* on its base. (*e*) The subscript *ṇa*, which already since early times is merely indicated by a loop (see *rṇṇa*, plate VII, 41, IV). (*f*) The subscript *tha*, which, as in other southern alphabets (compare, *e. g.*, plate VII, 45, XX), is changed to a double curve open on the right (plate VII, 45, IV; plate VIII, 49, I).

B. — The script of Central India

c8.s2.t2.1

The Central-Indian script is found fully developed in the inscriptions of Samudragupta at Eraṇ and of Candragupta II. at Udayagiri,⁶ on the copper-plates of the kings of Śarabhapura,⁷ of the Vākāṭakas,⁸ and of Tīvara king of Kosala,⁹ and in two early Kadamba inscriptions.¹⁰ In all these documents, the heads of the letters bear small squares, which are either hollow (plate VII, col. XI) or filled in (plate VII, col. X). These squares, to which on account of their resemblances to small boxes the script owes the name "box-headed," are, like the wedges, artificial developments of the *Serifs*. The solid, filled in, squares probably have been invented by writers who [63] used ink, and the hollow ones by persons writing with a *stilus*, who feared to tear their palm-leaves. Both varieties of "box-heads" occur occasionally or constantly in other districts and in connection with other alphabets (see, *e. g.*, the Valabhī inscription of plate VII, col. V, the archaic Kadamba inscription of plate VII, col. XII, the Pallava inscription of plate VII, col. XX), and even in Nos. 21 and 21, A, of the Campā inscriptions from Further India.¹ But the very peculiar appearance of the Central-Indian inscriptions of this class is due to the more or less rigorous modification of the letters by the contraction of their breadth and the conversion of all curves into angular strokes. This is best visible in the grants, figured in EI. 3, 260, and in Fleet's Gupta Inscriptions (CII. 3), Nos. 40, 41, 56, 81, plates 26, 27, 35, 45, among which No. 56 is represented in col. XI. of our plate VII, while col. X. offers the less carefully modified characters of F.GI (CII. 3), No. 55, plate 34. Both these inscriptions were issued in the same year from the *Dharmādhikaraṇa* of the Vākāṭaka king Pravarasena II.

c8.s2.t2.2

⁵ See also my remarks in *IA.* 7, 110, and below, § 28, B.

c8.s2.t2.3

⁶ *F.GI* (*CII.* 3), Nos. 2, 3, pl. 2, A, B.

c8.s2.t2.4

⁷ *Op. cit.*, Nos. 40, 41, plates 26, 27.

c8.s2.t2.5

⁸ *Op. cit.*, Nos. 53—55, plates 33, A, to 35; *IA.* 12, 239; *B.ASRWI.* 4, pl. 56, No. 4; pl. 57, No. 3; *EI.* 3, 260; the earliest of them belong in BHAGVANLAL INDRAJI'S and my opinion to the 5th, according to FLEET to the 7th, century.

c8.s2.t2.6

⁹ *F.GI* (*CII.* 3), No. 81, pl. 45; according to FLEET from the 8th or 9th century; according to KIELHORN, *EI.* 4, 258, undoubtedly from the 8th.

c8.s2.t2.7

¹⁰ See FLEET, *IA.* 21, 93: of the same type is, according to an impression presented to me by L. RICE, the Tāḷgund (Sthānakundūra) Praśasti of Kubja from the reign of Śāntivarman, *Ep. Carn.* 7, Sk. 176 (*EI.* 8).

c8.s2.t2.8

¹ Bergaigne-Barth, Inscriptions Sanskrit du Campā et du Cambodge, 2, 23; the Campā inscriptions show the northern *ka* and *ra* without curves at the end.

c8.s2.t2.9

Traces of the influence of the northern alphabets are visible in this script just as in the western variety, and particularly in the letters *ta*, *dha*, *na*, and in the Mātrās of medial *e*, *ai* and *o*, which in F.GI (CII. 3), No. 81, plate 45 (not in our plate), shew the peculiar tailed northern form of the 7th and 8th centuries. But in the ligatures (see, for instance, *nta*, plate VII, 43, X), we meet repeatedly with the looped *ta* and with the *na* without the loop, and even an independent looped *ta* appears exceptionally² in the word *snātānām* (No. 55, line 7; No. 56, line 6). Medial *au* has the tripartite western and northern form in F.GI (CII. 3), Nos. 2, 3, 40, 81, plates 2, A, B, 26, 45, but the southern bipartite form (see *dau*, plate VII, 24, XI) in the Vākāṭaka inscriptions. The *kha*, which has a big hook and small loop, and the oblong *ca* with the vertical on the right, likewise agree with the southern forms. But F.GI (CII. 3), No. 2, line 17, offers once, in *śulkā*, the northern *ka* without the curve at the foot.

c8.s2.t2.10

² Fleet and Kielhorn assume that the writers by mistake put *na* for *ta* and *vice versâ*.

c8.s2.t2.11

The other letters of this script frequently show greater or smaller variations. Our plate offers a few in the case of *A*, *ja*, *tha*, *ba* and *la*. More have been pointed out by Fleet and Kielhorn in their editions of the inscriptions in F.GI (CII. 3) and in EI. 3. I may add to Fleet's remarks, that his Nos. 40, 41, and 81 have the angular form of *ma* of the later Kanarese-Telugu alphabet (see below, § 29, B, 6).

§ 29. — The Kanarese and Telugu alphabet; Plates VII. and VIII

A. — The archaic variety

c8.s3.t1.1

[64] The archaic variety of this script is found: — (*a*) In the west, in the inscriptions of the Kadambas of Vaijayantī or Banavāsi (plate VII, cols. XII, XIII), and of the early Calukyas of Vātāpi or Bādāmi, *e.g.* of Kīrtivarman I. and Maṅgaleśa (plate VII, col. XIV), Pulakeśin II., and Vikramāditya I. (sometimes). (*b*) In the east, on the Sālaṅkāyana plates, and on those of the first two Calukyas of Veṅgī, Viṣṇuvardhana I. and Jayasiṁha I. (plate VII, col. XVII).³ The date of the Sālaṅkāyana plates, which used to be assigned to the 4th century,⁴ is uncertain.⁵ The Kadamba grants probably belong partly to the 5th and partly to the 6th centuries; for, Kākusthavarman, who issued the oldest known record, was the contemporary of one of the Imperial Guptas, probably of Samudragupta,⁶ and his descendants all ruled before the overthrow of the Kadamba kingdom by Kīrtivarman I., between A. D. 566-67 and 596-97. The archaic Calukya inscriptions fall between A. D. 578 and about 660.⁷

c8.s3.t1.2

³ Compare facsimiles of Sālaṅkāyana inscriptions at B.ESIP. plate 24; IA. 5, 176; EI. 4, 144; of Kadamba inscriptions at IA. 6, 23 ff.; 7, 33 ff.; J.BBRAS. 12, 300; of Western Calukya inscriptions at IA. 6, 72, 75; 8, 44, 237; 9, 100; 10, 58; 19, 58; and of Eastern Calukya inscriptions at B.ESIP. pl. 27.

c8.s3.t1.3

⁴ B.ESIP. 16, pl. 1.

c8.s3.t1.4

⁵ Fleet, IA. 20, 94.

c8.s3.t1.5

⁶ Academy, 1895, 229.

c8.s3.t1.6

⁷ See Fleet's dates of the Calukyas, EI. 3, table at p. 2; IA. 20, 95 ff.

c8.s3.t1.7

During this period, the characters of the western and eastern documents do not differ much. The alphabet of the Sālaṅkāyana plates⁸ agrees very closely with that of plate VII, col. XIII; and in the first half of the 7th century the letters of the Calukya inscriptions from Vātāpi and from Veṅgi show an almost perfect resemblance.¹ But the more considerable differences between cols. XII. and XIII, which both are derived from grants of the Kadamba Mṛgeśavarman issued within a period of only five years, have to be explained by the assumption that the letters of col. XIII, with which nearly all the other Kadamba inscriptions agree, imitate writing with ink, and those of col. XII. writing with the *stilus*. This explanation is suggested by the thinness of the signs of col. XII, and by the much greater thickness of those in col. XIII, and by the wedges and solid squares at their heads (compare above, § 28, B).

c8.s3.t1.8

⁸ B.ESIP. pl. 1.

c8.s3.t1.9

¹ Compare also the facsimile at IA. 6, 72, and B.ESI P. pl. 27.

c8.s3.t1.10

The letters of the older documents of this period remain very similar to those of the Andhra inscriptions of plate III, the so-called "cave-characters." In the Śālaṅkāyana grant, and in those of the Kadambas Kākusthavarman, Śāntivarman, Mṛgeśavarman and Ravivarman, we find only few, and by no means constant, traces of the development of the later characteristic round forms. Thus, col. XII. no doubt offers rather far advanced signs for *A* and *ra*, but at the same time a more archaic *A*, and the facsimile frequently shows even an angular *ra* with a not very long upward stroke. In the grants of the last Kadamba king Harivarman and in those of the Calukyas between A. D. 578 and 660, the *A*, *Ā*, *ka* and *ra*, characteristic of the next stage of development, occur not rarely, but never constantly. Thus col. XIV, derived from the Bādāmi inscription of Kīrtivarman I. and Maṅgaleśa, has the *ka* closed on the left. But this form is the only one used there, and it never appears on Maṅgaleśa's copper-plate, nor on the Haidarābād plates of his successor Pulakeśin II.² Further, this *ka*, as well as the closed *ra* of 33, col. XV, occur on the Nerūr plates of Pulakeśin II.³ Finally, the Aihoḷe stone inscription, of the time of Pulakeśin II.,⁴ has exclusively the older *ka* and *ra*, but occasionally the later *A* of col. XV. This vacillation indicates that between A. D. 578 and 660, and perhaps even earlier, the round-hand forms of the middle Kanarese alphabet existed, but that they either had not completely displaced the older ones, or that they were not yet considered as really suitable for inscriptions, though the clerks occasionally introduced them by mistake into the official documents (compare above, § 3, page 8).

c8.s3.t1.11

² IA. 6, 72.

c8.s3.t1.12

³ IA. 8, 44.

c8.s3.t1.13

⁴ See the plates at IA. 8, 241; EI. 6, 6.

c8.s3.t1.14

Among the other signs, the following may be noted especially:—

c8.s3.t1.15

(1) The *ṇa* (plate VII, 21, XII—XIV, XVII) which is never looped, but looks as if it were cursively developed from a looped form similar to that of col. I, ff.

c8.s3.t1.16

(2) The *ta*, which keeps the old form of the western inscriptions without a loop in 22, XIII, but shows in cols. XII, XIV, XVII, a cursive development from the looped *ta* of cols. XX—XXIII, which likewise is not rare in Kadamba and Calukya inscriptions of this period.

c8.s3.t1.17

(3) The tailed *ḍa* (24, XIV, XVII) agreeing exactly with the western form [65] of *ḍa* (19, IV—IX).

c8.s3.t1.18

(4) The *na*, which sometimes has the looped form (26, XIII), and more frequently that without the loop (26, XII, XIV—XVII); the latter being, however, apparently derived from the looped one.

c8.s3.t1.19

(5) The very exceptionally looped *ya* (in *yō*, 45, XIV), which thus is identical with the much older northern form.

c8.s3.t1.20

(6) The medial vowels: — (*a*) *ū* in *pū* (27, XIII), a cursive substitute for the *ū* of *yū* (32, VI), *cū* (13, IV), &c.; (*b*) the subscript *ṛ* of *kṛ* (8, XII, XVII; 41, XIV), somewhat resembling a northern *ṛ* (which latter actually occurs once on the seal figured in IA. 6, 24, in *Mṛgeśa*), but probably independently derived from a not uncommon *ṛ* in the shape of an unconnected semicircle before *ka*; (*c*) the exceedingly rare *ḷ* of *kḷ* (42, XIV), which, differing from the northern subscript *ḷ* (plate VI, 35, XVII), but agreeing with the northern initial sign of the Cambridge MS., consists merely of a cursive *la*; (*d*) the Mātrā of *e* (in *ṇe*, 21, XII), of *ai* (in *vai*, 13, XII; and *vai*, 35, XIII), and of *o* and *au* (in *thau*, 23, XII), which, except in connection with *le* (see *le*, 34, XII, and *lo*, 34, XIII, XVII), frequently stands at the foot of the consonant; (*e*) the *au* (in *pau*, 27, XII, XIV), the right-hand portion of which invariably and in all southern alphabets consists of a hook, formed by a cursive combination of the second Mātrā with the *ā*-stroke (compare *yau*, plate III, 31, VI).

B. — The middle variety

c8.s3.t2.1

This second variety is found from about A. D. 650 to about A. D. 950: — (*a*) In the west, in the inscriptions of the Calukyas of Vātāpi or Bādāmi, of their successors the Rāṣṭrakūṭas of Mānyakheṭa (in cases when they did not use the Nāgarī, see above, page 51), of the Gaṅgas of Maisūr, and of some smaller dynasties; (*b*) in the east, on the copper-plates of the Calukyas of Veṅgī and of their vassals. During this period, some marked differences are observable in the *ductus* between the several classes of documents. The copper-plates of the Western Calukyas (plate VII, col. XVI)¹ mostly show carelessly drawn cursive signs sloping towards the right, and their stone inscriptions (plate VII, col. XV) upright, carefully made, letters, which especially in the ligatures are abnormally large. With the characters of the latter agree those of the inscriptions of the Rāṣṭrakūṭas (plate VIII, cols II, III),² with the exception of the sign-manual on the Baroda copper-plate of Dhruva II.³ In this royal signature and in the inscriptions of the Calukyas of Veṅgī (plate VIII, cols. IV, V), the letters are broader and shorter, and in this respect resemble very closely the Old-Kanarese.⁴

c8.s3.t2.2

¹ Compare the facsimiles at IA. 6, 86, 88; 7, 300; J.BBRAS. 16, 223 ff.

c8.s3.t2.3

² Compare the facsimiles at IA. 10, 61 ff., 104, 166, 170; 11, 126; 20, 70; Ep. Carn. 3, 80, 87, 92 (for the last of these see also EI. 6, 54).

c8.s3.t2.4

³ See the facsimile at IA. 14, 200.

c8.s3.t2.5

⁴ Compare the facsimiles at IA. 12, 92; 13, 214, 248; EI. 3, 104.

c8.s3.t2.6

In addition to the above-mentioned rounded forms of *A*, *Ā*, *ka* and *ra*, which become constant during this period, the following letters deserve special remarks: —

c8.s3.t2.7

(1) The very rare *Ṛ* (plate VII, 5, XVI; compare also the earlier letter in the facsimile at IA. 6, 23, end), which seems to be a modification of the northern form of plate VI, 7, I, II.

c8.s3.t2.8

(2) The strongly cursive *kha* (plate VIII, 12, III—V), which is identical with the Old-Kanarese letter, and which according to FLEET⁵ never occurs before about A. D. 800, but actually appears in the cognate Pallava inscriptions (plate VII, 9, XXIII; compare below, § 31, B, 4) already since the 7th century.

c8.s3.t2.9

⁵ EI. 3, 162 f.

c8.s3.t2.10

(3) The *ca*, which from the 9th century begins to open in *ñca* (plate VII, 41, XIX; plate VIII, 19, III, IV).

c8.s3.t2.11

(4) The *da* (plate VIII, 27, II, IV, V) the tail of which begins to turn upwards since the 9th century.

c8.s3.t2.12

(5) The *ba*, opened above (plate VIII, 32, V), which according to FLEET⁶ first occurs about A. D. 850.

c8.s3.t2.13

⁶ EI. 3, 163.

c8.s3.t2.14

(6) The *ma* (plate VII, 31, XVII; VIII, 34, II—V), the upper part of which is drawn towards the right and placed nearly on the same level as the lower one, and which thus becomes the precursor of the Old-Kanarese *ma*.

c8.s3.t2.15

(7) The abnormal cursive *la* (plate VII, 34, XVI), which elsewhere appears only as the second part of ligatures (as in *ślo*, plate VII, 44, XVIII).

c8.s3.t2.16

(8) The Mātrās, which occasionally stand below the consonant (as in *dhe*, plate VIII, 28, V).

c8.s3.t2.17

(9) The vertical Virāma, above final *m* (plate VII, 41, XVIII; plate VIII, 46, V) and final *n* (plate VIII, 45, V).

c8.s3.t2.18

(10) The Dravidian *ṟa* (plate VII, 45, XV, XVIII; 46, XVI; plate VIII, 47, II, III) [66] and *ḻa* (plate VII, 46, XV, XVIII; plate VIII, 49, II, V), which first appear in the 7th century. The first of them, *ṟa*, may possibly represent two round *ra*, and *ḻa* may be a modification of a *la* like that in plate VII, 40, XIV, XVI. The occurrence of these signs proves that the Kanarese language had a literature already in the 7th century.

C. — The Old-Kanarese alphabet

c8.s3.t3.1

The third and last variety of the Kanarese-Telugu alphabet, which BURNELL calls “the transitional” and FLEET more appropriately “Old-Kanarese,” does not differ much from the modern Kanarese and Telugu scripts. In the east, it first appears in the Veṅgi inscriptions of the 11th century; in the west, a little earlier, in a Gaṅga inscription of A. D. 978 and in a not much later Calukya inscription.¹ Some of its characteristics, like the opening of the loop of *ma* and of the head of *va*, appear however in the sign-manual of Dhruva II. on the Baroda plates, mentioned above under B. The specimens of this script² in plate VIII, among which cols. VI. and VII. date from the 11th century, col. VIII. from the 12th, and col. IX (according to HULTZSCH, Telugu) from the 14th, show the gradual progress very distinctly.

c8.s3.t3.2

¹ BURGESS and FLEET, Pāli, Saṁskṛit, and Old-Canarese inscriptions, Nos. 271, 214; see also, for the Gaṅga record, IA. 6, 102.

c8.s3.t3.3

² Compare also the facsimiles at IA. 9, 74; 14, 56; EI. 3, 26, 88, 194, 228; Ep. Carn. 3, 116, 121; B.ASWI. No. 10, 100; and J.RAS. 1891, 135 (the original of PRINSEP’s, Kistna alphabet, which is archaic and retrogracd *A*, *ka*, *ra*, *la*).

c8.s3.t3.4

One of the most characteristic marks of the Old-Kanarese consists in the angles over all Mātṛkās which do not bear superscripted vowel-signs. These angles, which in col. VI. resemble those of the modern Telugu and in cols. VII, VIII, those of the modern Kanarese, probably are cursive representatives of wedges, and have been invented because the latter did not suit the writing with the *stilus*. Since the 6th century, they occur more or less frequently in single inscriptions from other districts, such as Guhasena’s grant of A. D. 559-60 (plate VII, col. IV) and Ravikīrti’s Aihoḷe Praśasti,³ sometimes together with wedges. But it is only in this alphabet that they become a constant distinctive feature.

c8.s3.t3.5

³ IA 8, 241; EI. 6, 6.

c8.s3.t3.6

The most important among the changes in the several signs are:—

c8.s3.t3.7

(1) The opening of the heads of *E* (plate VIII, 8, VI, VIII), of *ca* (16, VI—IX), of *bha* (33, VI—IX, which in col. IX. becomes identical with *ba* by the connection of the two base-strokes), and of *va* (38, VII—IX), as well as of the loop of *ma* (34, VI, VIII) and of the right limb of *cha* (17, VI—IX; compare also col. V).

c8.s3.t3.8

(2) The cursive looped forms of *A*, *Ā* (1, 2, VII—IX), and of *I*, *Ī* (3, 4, VI—IX; compare their precursors in 3, II, and 4, III, V), and of *śa* (39, VII—IX), the central cross-bar of which is connected with the curved end of the right side.

c8.s3.t3.9

(3) The conversion of the long-drawn loops of *ka* (11, VI—IX) and of *ra* (36, VI—IX) into much smaller circles.

c8.s3.t3.10

(4) The cursive rounding off of the angles of *ṇa* (24, VI—IX), *na* (29, VI—IX), and *sa* (41, VI—IX).

c8.s3.t3.11

(5) The development of new loops or ringlets to the right of the top of *Ṛ* (7, IX), *ṅa* (15, VIII, IX) and *ja* (18, VI—IX; compare col. V).

c8.s3.t3.12

(6) The exclusive employment of the medial *u* turning upwards on the right (see, for instance, *pu*, 30, IX), which in earlier times is restricted to *gu*, *tu*, *bhu* and *śu*, but later appears also in *su* (plate VIII, 41, II, III).

c8.s3.t3.13

(7) Finally, the appearance of the Anusvāra on the line (see *raṁ*, 36, VIII), which cannot be a survival from ancient times, but must be an innovation intended to make the lines more equal (compare page 59 above, § 26, A, 5).¹

c8.s3.t3.14

¹ Compare for this paragraph B ESIP. 15 ff.

§ 30. — The later Kaliṅga script; Plates VII. and VIII

c8.s4.1

[67] This script has been found hitherto only on the copper-plates of the Gaṅga kings of Kaliṅganagara, the modern Kaliṅgapattanam in Gañjām, which in olden times was the residence of the Ceta king Khāravela and his successors (see page 39 f. above). The dates of these documents run from the year 87 of the Gāṅgeya era. Though its exact beginning has not yet been determined, FLEET has shown that the oldest Gaṅga grants probably belong to the 7th century.²

c8.s4.2

² IA. 13, 274; 15, 133.

c8.s4.3

The signs of these documents resemble, up to the Gāṅgeya year 183, partly the letters of the Central-Indian script (above, § 28, B) and partly those of the western variety, which exhibits the medial *au*, of the Ajaṇṭā inscriptions (above, § 28, A), and they show only a few peculiar forms. A specimen of the Kaliṅga script of the latter kind has been given in plate VII, col. XIX, from the Chicacole grant of the Gāṅgeya year 148, in which only the Grantha-like *A* (2, XIX), and the *ga* (10, XIX) and *śa* (36, XIX) with curves on the left, differ greatly from the corresponding Valabhī letters. The alphabet of the Acyutapuram plates³ of the Gāṅgeya year 87, which exhibits angular forms with solid box-heads, closely resembles the Central-Indian writing; but its *na* is identical with that of the modern Nāgarī. The Chicacole plates⁴ of the Gāṅgeya year 128 show in general the same type; but they offer the ordinary looped *na* of the north and west, and the looped *ta* of the archaic Grantha (22, XX, ff.). Finally, the Chicacole plates⁵ of the Gāṅgeya year 183 come close to the script of plate VII, col. X; but their *na* is again that of the late Nāgarī, and their medial *ā* mostly stands above the line, as in various northern and also Grantha documents of the 7th and 8th centuries.

c8.s4.4

³ EI. 3, 128.

c8.s4.5

⁴ IA. 13, 120; compare 15, 131 f.

c8.s4.6

⁵ EI. 3, 132.

c8.s4.7

In the grants of the 3rd and 4th centuries of the Gāṅgeya era, and in a late undated inscription, the mixture of the characters is much greater, and the same letter is often expressed by greatly differing signs. In plate VIII, col. X, from the Chicacole plates of the Gāṅgeya year 51, that is 251,⁶ and in col. XI, from the Vizagapatam plates of the year 254, and in col. XII, from the Alamanda plates of the year 304, we find a northern *A*, *Ā* (1, 2, X—XII), *I* (3, XI), *U* (5, X), *ka* (44, XI, XII), *kha* (12, XI), *ṅga* (15, X), *ṅka* (15, XII), *ja* (18, XII), *ña* (in *jñā*, 19, X), *ḍa* (22, XII), *ṇa* (24, XI, XII), *dha* (28, 45, XI), *na* (48, X), and *pra* (47, XII). The other letters are of southern origin, and belong partly to the middle Kanarese, partly to the middle Grantha, or are peculiar developments. The restricted space available in plate VIII. has made it impossible to enter all the variants for each letter. But the three different forms of *ja* (18, 46, and 47, X) show how very great the variations are.

c8.s4.8

⁶ The words *śata-dvaya* probably have been left out by mistake after *saṁvatsara*.

c8.s4.9

Still stronger are the mixture and variations in the Chicacole plates of the Gāṅgeya year 351,⁷ and in the undated grant of Vajrahasta from the 11th century (KIELHORN),⁸ neither of which is represented in our plate. In the first-named document each letter has, according to FLEET, at least two, but sometimes three or four forms. The majority of the signs belong to the southern Nāgarī. But Old-Kanarese and late Grantha signs likewise occur. In Vajrahasta's grant there are, according to KIELHORN'S calculation, 320 Nāgarī letters and 410 southern ones of different types, and each letter again has at least two and sometimes [68] four or more forms. KIELHORN points out that the writer has shown a certain art in the grouping of the variants; and he is no doubt right in hinting that the mixture is due to the vanity of royal scribes, who wished to show that they were acquainted with a number of alphabets. For the same reason. the writer of the Chicacole plates of the Gāṅgeya year 183 has used three different systems of numeral notation in expressing the date (see below, § 34). The kingdom of the Gaṅgas of Kaliṅga lay between the districts in which the Nāgarī and the Kanarese-Telugu scripts were used, and it was not far from the territory of the Grantha. Its population was probably mixed, and used all these scripts,¹ as well as, in earlier times, those employed in the older western and Central-Indian inscriptions. The professional clerks and writers of course had to master all the alphabets.

c8.s4.10

⁷ IA. 14, 10 f.; HULTZSCH's undoubtedly correct reading of the date has been adopted by FLEET in his Dynasties of the Kanarese Districts, Bombay Gazetteer, vol. 1, part 2, p. 297, note 8, the printed sheets of which I owe to the author's courtesy. FLEET declares this inscription, as well as those represented in plate VIII, cols. X, XII, to be suspicious, — in my opinion, without sufficient reasons.

c8.s4.11

⁸ EI. 3, 220.

c8.s4.12

¹ The use of northern characters is proved by the Bugaḍa plates, EI. 3, 41; compare also B.ESIP. 53, and plate 22 *b*.

§ 31. — The Grantha alphabet; Plates VII. and VIII

A. — The archaic variety

c8.s5.t1.1

For the history of the Sanskrit alphabets in the Tamil districts during the period after A. D. 350, we have only the Sanskrit inscriptions of the Pallavas, Colas and Pāṇḍyas from the eastern coast, among which only those of the first-named dynasty can lay claim to a higher antiquity. Corresponding inscriptions from the western coast are hitherto wanting. For this reason, and because only a small number of the eastern documents have been published with good facsimiles, it is as yet impossible to give a complete view of the gradual development of the letters.

c8.s5.t1.2

The most archaic forms of the Sanskrit scripts of the Tamil districts, which usually are classed as "Grantha," are found on the copper-plates of the Pallava kings of Palakkada and (? or) Daśanapura² (plate VII, cols. XX, XXI) from the 5th or the 6th century (?), with which the ancient inscriptions, Nos. 1 to 16, of the Dharmarājaratha (plate VII, col. XXII)³ closely agree. These inscriptions, together with a few others,⁴ exhibit what may be called the archaic Grantha, the latest example of which occurs in the Bādāmi inscription, incised, according to FLEET'S newest researches,⁵ by the Pallava Narasiṁha I., during his expedition against the Calukya Pulakeśin II. (A. D. 609 and about 642) in the second quarter of the 7th century; and it seems to have gone out soon after, as the Kūram plates of Narasiṁha's son Parameśvara I. show letters of a much more advanced type. It is met with also in the stone inscription from Jambu in Java; see IA. 4, 356.

c8.s5.t1.3

² IA. 5, 50, 154; compare B.ESIP. 36, note 2.

c8.s5.t1.4

³ I owe the facsimi es of this inscription and of those used for pl. VII, col. XXIV, and pl. VIII, col. XIII, to HULTZSCH'S kindness; see now his SII. 2, part 3.

c8.s5.t1.5

⁴ IA. 9, 100, No. 82, 102, No. 85; 13, 48; EI. 1, 397.

c8.s5.t1.6

⁵ Dynasties of the Kanarese Districts, Bombay Gazetteer, vol. 1, part 2, p. 328.

c8.s5.t1.7

The characters of the archaic Grantha in general agree with those of the archaic Kanarese-Telugu (see above, § 29, A), but shew a few peculiarities which remain constant in the later varieties: thus: —

c8.s5.t1.8

(1) The *tha*, the central dot of which is converted into a loop, attached to the right side (plate VII, 23, XXI); compare the *tha* of col. XX, where the straight stroke of the Kanarese-Telugu script appears.

c8.s5.t1.9

(2) The *śa* with the cross-bar converted into a curve or loop and attached to the right side (plate VII, 36, XX—XXII, 45, XXII); compare also the cursive *śa* of the western script, mentioned above, § 28, A, 7.

c8.s5.t1.10

(3) The *ṣa* with the cross-bar treated similarly (plate VII, 37, XX); compare the *ṣa* of col. XXI, which shows the older form.

c8.s5.t1.11

The characters of plate VII, cols. XX, XXI, show no closer connection with those of the Prākrit inscriptions of the Pallavas, discussed above in § 20, D.

B. — The middle variety

c8.s5.t2.1

The earliest inscription of the much more advanced forms of the second variety or the middle Grantha, is found on the Kūram copper-plates (plate VII, col. XXIV) of the reign of Parameśvara I., the adversary of the Western Calukya Vikramāditya I. (A. D. 655—680).¹ @H [69] Compared with this document, which appears to offer a real clerk's script, the monumental inscription of the Kailāsanātha temple (plate VII, col. XXIII), built according to Fleet² by Narasiṁha II., the son of Parameśvara I., is retrograde, and shows more archaic forms for several paleographically important letters. On the other hand, the Kaśākūḍi copper-plates (plate VIII, col. XIII), incised in the time of Nandivarman who succeeded Mahendra III., the second son of Narasiṁha II., and warred with the Western Calukya Vikramāditya II. (A. D. 733—749),³ agree more closely with the Kūram plates, and offer, besides some archaic forms, also much more advanced ones.

c8.s5.t2.2

¹ Hultzsch, *SII.* 1, 144 ff.; Fleet, *op. cit.* (preceding note), 322 f.

c8.s5.t2.3

² Fleet, *op. cit.*, 329 f.

c8.s5.t2.4

³ Fleet, *op. cit.*, 323 ff.

c8.s5.t2.5

The most important innovations, either constantly or occasionally observable in this second variety of the Grantha, are: —

c8.s5.t2.6

(1) The development of a second vertical in *A*, *Ā*, *ka* and *ra* (plate VII, 1, 2, 8, 33, XXIII, XXIV; plate VIII, 1, 2, 11, 36, XIII), as well as in medial *u* and *ū* (plate VII, 31, 38, XXIV; plate VIII, 34, 40, XIII), out of the ancient hook; compare the transitional forms in the facsimiles at IA. 9, 100, 102.

c8.s5.t2.7

(2) The connection of one of the dots of *I* with the upper curved line (plate VII, 3, XXIII, XXIV; plate VIII, 3, XIII, *a*, *b*).

c8.s5.t2.8

(3) The opening of the top of *E* (plate VII, 5, XXIV), which however shows closed up forms in col. XXIII, and in plate VIII, 8, XIII.

c8.s5.t2.9

(4) The development of a loop to the left of the foot of *kha*, and the opening up of the right side of the letter (plate VII, 9, XXIII), as in the Kanarese-Telugu script (see above, § 29, B, 2).

c8.s5.t2.10

(5) The upward turn of the Serif at the left-hand lines of *ga* and *śa* (plate VII, 10, 36, XXIV; plate VIII, 13, 39, XIII; not in plate VII, col. XXIII).

c8.s5.t2.11

(6) The opening up of the loops of *cha* (plate VIII, 17, XIII), and perhaps also in the indistinct *cha* of the Kūram plates, i, line 5.

c8.s5.t2.12

(7) The transposition of the vertical of *ja* to the right end of the top-bar, and the conversion of the central bar into a loop connected with the lowest bar (plate VII, 15, XXIV; plate VIII, 18, XIII; not in plate VII, col. XXIII).

c8.s5.t2.13

(8) The incipient opening up of the tops of *dha* and *tha* (plate VII, 23, 25, XXIII, XXIV; plate VIII, 26, 28, XIII).

c8.s5.t2.14

(9) The opening up of the top of *ba*, and the transposition of the original top-line to the left of the left-hand vertical (plate VII, 29, XXIV; plate VIII, 32, XIII; not in plate VII, col. XXIII).

c8.s5.t2.15

(10) The adoption of the later northern *bha* (see above, § 24, A, 24), or the development of an exactly similar sign (plate VII, 30, XXIV; plate VIII, 33, XIII; not in plate VII, col. XXIII).

c8.s5.t2.16

(11) The combination of the left-hand vertical of *sa* with the left end of the old side-limb, and of the right end of the side-limb with the base-stroke (plate VII, 38, XXIV; a transitional form in col. XXII, and a different cursive form in plate VIII, 41, XIII).

c8.s5.t2.17

(12) The frequent separation of medial *ā*, *e*, *ai*, *o*, *au*, from the Mātṛkā (constant in plate VIII, col. XIII), as well as the use of the *ā* standing above the line, as in the northern alphabet of this period and in the Central-Indian script (compare plate VII, 17, 19, 21, 31—33, XXIII; 8, 24, XXIV).

c8.s5.t2.18

(13) The expression of the Virāma (as in the Kanarese-Telugu script) by a vertical stroke above, or in the Kaśākūḍi plate also to the right of, the final consonant (plate VII, 41, XXIII; plate VIII, 47, XIII; and compare the facsimiles).

c8.s5.t2.19

(14) The transposition of the Anusvāra to the right of the Mātṛkā (plate VII, 38, XXIV) below the level of the top-line, as in the Kanarese-Telugu script.

c8.s5.t2.20

(15) The occasional development of small angles, open above, at the tops of the verticals, for the left part of which a dot usually appears in plate VIII, col. XIII.

c8.s5.t2.21

The fully-developed and very constant characteristics of the alphabet of the Kūram plates make it probable that they have not arisen within the period of twenty to thirty years, which lies between the issue of the Kūram grant and the incision of the much more archaic Bādāmi inscription of Narasiṁha I. (see above, under *A*). Very likely the Kūram alphabet had a longer history.

C. — The transitional Grantha

c8.s5.t3.1

The series of the published datable Pallava inscriptions of the 8th century ends for the present with the Kaśākūḍi plates; and facsimiles of documents of the next following centuries [70] are not accessible to me. I am, therefore, unable to exactly fix the time when the third or transitional variety of the Grantha, BURNELL'S Coḷa or middle Grantha, came into use, which is found in the inscriptions from the reign of the Bāṇa king Vikramāditya¹ about A. D. 1150 (plate VIII, col. XIV) and of Sundara-Pāṇḍya,² A. D. 1250 (plate VIII, col. XV), as well as in other documents.³ It would however appear, both from the Grantha signs occurring in the Gaṅga inscriptions (plate VIII, cols. XI, XII) and from BURNELL'S Coḷa-Grantha alphabet of A. D. 1080,⁴ that the new developments originated partly towards the end of the 8th century and partly in the 9th and 10th, about the same time when the Old-Kanarese script (above, § 29, C) was formed.

c8.s5.t3.2

¹ *EI.* 3, 75.

c8.s5.t3.3

² *EI.* 3, 8.

c8.s5.t3.4

³ Compare facsimiles at IA. 6, 142; 8, 274; 9, 45 (*EI.* 3, 79 f.); *EI.* 3, 228; Ep. Carn. 3, 166; *SII.* 2, pl. 2; the last inscription and the last but two are older than the 11th century.

c8.s5.t3.5

⁴ *B.ESIP.* plate 13.

c8.s5.t3.6

The most important changes, which the transitional Grantha shows, are as follows:—

c8.s5.t3.7

(1) The suppression of the last remaining dot of *I* (plate VIII, 3, XIV, XV; compare 3, XIII, *a*).

c8.s5.t3.8

(2) The formation of a still more cursive *E* (8, XIV) out of the Kūram letter (plate VII, 6, XXIV).

c8.s5.t3.9

(3) The formation of a still more cursive *kha* (plate VIII, 12, XIV, XV), closely resembling the later Kanarese-Telugu sign (plate VIII, 12, III, ff.), out of the letter of plate VII, 9, XXIII.

c8.s5.t3.10

(4) The development of a single or double curve to the left of *gha* (plate VIII, 14, XIV, XV).

c8.s5.t3.11

(5) The opening up of the top of *ca*, and the conversion of its left side into an acute angle (plate VIII, 16, XIV, XV).

c8.s5.t3.12

(6) The addition of a curve to the right end of *ḍa* (plate VIII, 22, XIV, XV).

c8.s5.t3.13

(7) The development of an additional loop in *ṇa* (plate VIII, 24, XIV, XV), in accordance with the practice of the Tamil alphabet (see below, § 32, A).

c8.s5.t3.14

(8) The complete opening up of the tops of *tha* and *dha* (plate VIII, 26, 28, XIV, XV).

c8.s5.t3.15

(9) The development of a curve at the left side of *pa* (plate VIII, 30, XIV, XV).

c8.s5.t3.16

(10) The closing up of the top of *ma* (plate VIII, 34, XIV, XV), found already in the Gaṅga inscription of about A. D. 775 (plate VIII, 46, XI).

c8.s5.t3.17

(11) The suppression of the circle or loop on the right side of *ya* (plate VIII, 35, XIV, XV), whereby the letter obtains a very archaic appearance.

c8.s5.t3.18

(12) The opening up of the top of *va*, and the addition of a curve to its left side (plate VIII, 38, XVI, XV).

c8.s5.t3.19

(13) The complete separation of medial *ā*, *e*, *ai*, *o* from the Mātṛkās, and the formation of a separate sign for the second half of *au*, consisting of two small curves with a vertical on the right.

c8.s5.t3.20

It is worthy of note that the later alphabet of col. XV. has some more archaic signs than the earlier one of col. XIV. The reason no doubt is that the latter imitates the hand of the clerks of the royal office, while the former shows the monumental forms, suited for a public building. All the Grantha inscriptions imitate characters written with a *stilus*.

§ 32. — The Tamil and Vaṭṭeḷuttu alphabets; Plate VIII

A. — The Tamil

c8.s6.t1.1

The Tamil, as well as its southern and western cursive variety, the Vaṭṭeḷuttu or "roundhand," differs from the Sanskrit alphabet by the absence not only of the ligatures, but also of the signs for the aspirates, for the *mediae* (expressed by the corresponding *tenues*), for the sibilants (among which the palatal one is expressed by *ca*), for the spirant *ha*, for the Anusvāra and for the Visarga, as well as by the development of new letters for final *n*, and for *ṟa*, *ḻa* and *ḷa*, which latter three characters do not resemble those for the corresponding sounds in the Kanarese-Telugu script. The great simplicity of the alphabet fully agrees with the theories of the Tamil grammarians, and is explained by the peculiar phonetics of the Tamil language. Like all the older Dravidian dialects, the Tamil possesses no aspirates and no spirant. Further, it has no *ja*, and only one sibilant, which, according to CALDWELL, lies between *śa*, *ṣa* and *ca*, and which, if doubled, becomes a distinct *cca*. **[71]** The use of separate signs for the *tenues* and *mediae* was unnecessary on account of their mutual convertibility. The Tamil uses in the beginning of words only *tenues*, and in the middle only double *tenues* or single *mediae*. Hence, all words and affixes beginning with gutturals, linguals, dentals and labials, have double forms.¹ A knowledge of these simple rules makes mistakes, regarding the real phonetic value of *ka*, *ṭa*, *ta* and *pa*, impossible. The use of ligatures probably has been discarded because the Tamil allows even in loan-words no other combinations of consonants but repetitions of the same sound, and because it seemed more convenient to use in these cases the Virāma.²

c8.s6.t1.2

¹ CALDWELL, Comparative Grammar of the Dravidian Languages, 21—27.

c8.s6.t1.3

² Differently BURNELL, ESIP. 44, 47 ff., who considers the Vaṭṭeḷuttu as independent of the Brāhmī, but likewise of Semitic origin, and declares the Tamil alphabet to be the result of a Brahmanical adaptation of the Grantha letters to the phonetical system of the Vaṭṭeḷuttu. This view has already been characterised "as hardly in accordance with the facts" by CALDWELL, op. cit., 9.

c8.s6.t1.4

The occurrence of signs for the Dravidian liquids, which, though the sounds correspond with those of the older Kanarese and Telugu, differ from the characters of the Kanarese-Telugu script, indicates that the Tamil alphabet is independent of the latter and has been derived from a different source. HULTZSCH's important discovery of the Kūram plates,¹ with a large section in the Tamil script and language of the 7th century, confirms this inference. The Tamil alphabet of these plates agrees only in part with their Grantha, and many of its letters offer characteristics of the northern alphabets.

c8.s6.t1.5

¹ SII. 1, 147; compare 2, plate 12; the characters of the Vallam Cave inscription, op. cit., 2, plate 10, fully agree,

First 183 of 220 passages. Read on

Linked texts (16)