Indian Palaeography from about B.C. 350 to about A.D. 1300 c6.s4: § 4. — The origin of the Brāhma alphabet
Indian Palaeography from about B.C. 350 to about A.D. 1300 c6.s4
- Language
- English
Contents
- c6.s4.t1 A. — Borrowed signs — 4 parts
- c6.s4.t2 B. — Derivative consonants and initial vowels — 8 parts
- c6.s4.t3 C. — Medial vowels and absence of vowel in ligatures — 8 parts
§ 4. — The origin of the Brāhma alphabet
c6.s4.1
⁵ B.IS. III², 53—82.
c6.s4.2
[10] Among the numerous greatly differing proposals to explain the origin of the Brāhmī,⁶ there are five for which complete demonstrations have been attempted : — (1) A. CUNNINGHAM'S derivation from indigenous Indian hieroglyphics;⁷ (2) A. WEBER'S derivation from the most ancient Phoenician characters;⁸ (3) W. DEECKE'S derivation from the Assyrian cuneiform characters, through an ancient South-Semitic alphabet which is also the parent of the Sabaean or Himyaritic script;⁹ (4) I. TAYLOR'S derivation from a lost South-Arabian alphabet, the predecessor of the Sabaean;¹⁰ (5) J. HALÉVY'S derivation from a mixture of Aramaic, Kharosthī and Greek letters of the last quarter of the fourth century B. C.¹¹
c6.s4.3
⁶ R. N. Cust, Ling. and Or. Essays 2nd Ser., 27—52.
c6.s4.4
⁷ C.IA (CII. 1), 52 ff.
c6.s4.5
⁸ ZDMG. 10, 389 ff.; Ind. Skizzen 125 ff.
c6.s4.6
⁹ ZDMG. 31, 598 ff.
c6.s4.7
¹⁰ The Alphabet, 2, 314 ff.; restated with some modifications by F. MÜLLER, Mélanges Harlez 212 ff.
c6.s4.8
¹¹ JA. 1885, 268 ff.; Revue Sém. 1895, 223 ff.
c6.s4.9
CUNNINGHAM'S opinion, which was formerly shared by some eminent scholars, presupposes the use of Indian hieroglyphic pictures, of which hitherto no trace has been found. On the other hand, the legend of the Eran coin, which runs from the right to the left, and the letters seemingly turned round in the opposite direction which appear rarely in the Aśoka edicts and more frequently in the Bhattiprolu inscriptions, point to the correctness of the view taken as granted in all the other attempts at explanation, *viz.*, that Semitic signs are the prototypes of the Brāhma letters.
c6.s4.10
Among the remaining four proposals, J. HALÉVY'S *a priori* improbable theory may be at once eliminated, as it does not agree with the literary and paleographic evidence just discussed, which makes it more than probable that the Brāhmī was used several centuries before the beginning of the Maurya period, and had had a long history at the time to which the earliest Indian inscriptions belong. It is more difficult to make a choice between A. WEBER'S derivation from the oldest North-Semitic alphabet, and the view of W. DEECKE and I. TAYLOR, who derive the Brāhmī from an ancient South-Semitic script. Neither the one nor the other derivation can be declared to be *a priori* impossible; for, the results of modern researches make a high antiquity probable for also the Sabaean script, and point to the conclusion that this alphabet not only is older than the oldest Indian inscriptions, but that it existed at a period for which no evidence for the use of writing in India is available.¹ But according to these results, the question has to be put in a manner somewhat differing from that in which DEECKE and TAYLOR have put it. The point to be ascertained is no longer, whether the Brāhmī can be derived from an unknown predecessor of the Sabaean alphabet, but whether it can be derived directly from the actually known Sabaean characters.
c6.s4.11
¹ MORDTMANN and D. H. MÜLLER, *Sab. Denkmäler* (in DWA. Phil. Hist. Cl. 31), p. 108 f.
c6.s4.12
In all attempts at the derivation of alphabets, it is necessary to keep in mind three fundamental maxims, without which no satisfactory results can be obtained:—
c6.s4.13
(1) For the comparison of the characters to be derived, the oldest and fullest forms must be used, and the originals from which they are derived must belong to the types of one and the same period.
c6.s4.14
(2) The comparison may include only such irregular equations as can be supported by analogies from other cases where nations have borrowed foreign alphabets.
c6.s4.15
(3) [11] In cases where the derivatives show considerable differences from the supposed prototypes, it is necessary to show that there are fixed principles, according to which the changes have been made.
c6.s4.16
If one wishes to keep to these principles in deriving the Brāhmī from Semitic signs, neither the Sabaean alphabet, nor its perhaps a little more archaic variety, the Lihyanian or Thammudaean,² will serve the purpose, in spite of a general resemblance in the *ductus* and of a special resemblance in two or three letters. The derivations proposed by DEECKE and TAYLOR do not fulfil the absolutely necessary conditions, and it will probably not be possible to obtain satisfactory results, even if all the impossible equations are given up, and the oldest Indian signs in every case are chosen for comparison. It would be necessary to assume that several Sabaean letters, such as *Aleph*, *Gimel*, *Zain*, *Teth*, *Phe*, *Qoph*, *Resh*, which show strong modifications of the North-Semitic forms, had been again made similar to their prototypes on being converted by the Hindus into *A*, *ga*, *ja*, *tha*, *pa*, *kha* and *ra*. In other cases, it would be impossible to show any connection between the Sabaean and the Indian signs. These difficulties disappear with the direct derivation of the Brāhmī from the oldest North-Semitic alphabet, which shows the same type from Phoenicia to Mesopotamia. The few inadmissible equations which WEBER’s earlier attempt contains, may be easily removed with the help of recently discovered forms, and it is not difficult to recognise the principles, according to which the Semitic signs have been converted into Indian ones.
c6.s4.17
² D. H. MÜLLER, *Denkmäler aus Arabien* (DWA. Phil. Hist. Cl. 37), p. 15 ff.
c6.s4.18
An examination of the old Indian alphabet in plate II. reveals the following peculiarities:—
c6.s4.19
(1) The letters are set up as straight as possible, and, with occasional exceptions in the case of *ṭa*, *ṭha* and *ba*, they are made equal in height.
c6.s4.20
(2) The majority consist of vertical lines with appendages attached mostly at the foot, occasionally at the foot and at the top, or rarely in the middle; but there is no case in which an appendage has been added to the top alone.
c6.s4.21
(3) At the top of the letters appear mostly the ends of verticals, less frequently short horizontal strokes, still more rarely curves on the tops of angles opening downwards, and, quite exceptionally, in *ma* and in one form of *jha*, two lines rising upwards. In no case does the top show several angles, placed side by side, with a vertical or slanting line hanging down, or a triangle or a circle with a pendant-line.
c6.s4.22
The causes of these characteristics of the Brāhmī are a certain pedantic formalism, found also in other Indian creations, a desire to frame signs suited for the formation of regular lines, and an aversion to top-heavy characters. The last peculiarity is probably due in part to the circumstance that since early times the Indians made their letters hang down from an imaginary or really drawn upper line,¹ and in part to the introduction of the vowel-signs, most of which are attached horizontally to the tops of the consonants. Signs with the ends of verticals at the top were, of course, best suited for such a script. Owing to these inclinations and aversions of the Hindus, the heavy tops of many Semitic letters had to be got rid of, by turning the signs topsy-turvy or laying them on their sides, by opening the angles, and so forth. Finally, the change in the direction of the writing necessitated a further change, inasmuch as the signs had to be turned from the right to the left, as in Greek.
c6.s4.23
¹ ^1 Compare Berūnī's India, 1, 172 (SACHAU).
c6.s4.24
[12] The details of the derivation, for which, with the exception of the evidently identical Nos. 1, 3—7, 9, 12, 16, 17, 19—22, only a greater or smaller degree of probability can be claimed, are shown in the subjoined comparative table, which has been drawn by Mr. S. PEPPER of Vienna. Cols. I, II, showing the oldest Phoenician characters and those from Mesa's stone, have been taken from PH. BERGER'S Histoire de l'Écriture dans l'Antiquité, pp. 185, 202. Col. III, comes from EUTING'S Tabula Scripturae Aramaicae of 1892. And cols. IV—VI, with the exception of the signs marked by asterisks as hypothetical, are taken from plate II. of this work. With respect to the single letters, I add the following explanatory remarks, brief abstracts of those in my Indian Studies, III^2, p. 58 ff.
A. — Borrowed signs
c6.s4.t1.1
No. 1, *A*, col. V, = *Aleph*, cols. I, II (WEBER doubtfully), [13] turned from right to left except on the Patnā seal (above, § 3, and pl. II, 1, I), with transposition of the vertical line to the end of the angle. — No. 2, *ba*, col. V, *a*, *b*, *c*, = *Beth*, cols. I, II (WEBER); the opening of the triangular top produced first a sign like that in col. IV, next the rhombus, col. V, *a*, and finally the square and the oblong, col. V, *b*, *c*. — No. 3, *ga*, col. V, = *Gimel*, cols. I, II. — No. 4, *dha*, col. V, *a*, *b*, = *Daleth*, cols. I, II (WEBER), set up straight with rounded back (compare the half-angular forms, pl. II, 26, IX, XIX, XXIII, and the triangular, pl. III, 24, VII—XIII), with or without the turn from right to left. — No. 5, *ha*, col. V, = *He* (WEBER doubtfully), the Siddāpura form, col. V, *a*, being probably derived from the *He* of col. III, *a* (Mina of Salmanassar, before B. C. 725), which was turned topsy-turvy and from right to left. The more similar *He* of the sixth century B. C. (col. III, *b*) cannot be the prototype, because it occurs in the period when the Brāhmī had been developed, and because then the Semitic *Aleph*, *Daleth*, *Cheth*, *Theth*, *Waw*, and *Qoph* had become cursive and had been changed so much that they could no longer have produced the Indian forms. — No. 6, *va*, col. V, *a*, *b*, = *Waw*, col. II (WEBER doubtfully), turned topsy-turvy and with the lower end shut. — No. 7, *ja*, col. V, = *Zain*, cols. I, II (WEBER); a displacement of the two bars produced the Drāviḍī letter, col. V, *a*; from this was derived, the letter being made with one stroke of the pen, the *ja* of the northern Brāhmī, col. V, *b*, with a loop, for which, owing to the use of ink, a dot was substituted in the *ja* of col. V, *c*. The usual Girnār form, col. V, *d*, was also derived from the Drāviḍa form, the letter being made with two strokes of the pen.
c6.s4.t1.2
No. 8, *gha*, col. V, *a*, *b*, = *Cheth*, cols. I, II (TAYLOR), the Semitic sign being laid on its side, col. IV (on account of its often sloping position), and the upper horizontal bar being changed into a vertical. — No. 9, *tha*, col. V, = *Theth*, col. I (WEBER), with the substitution of a dot for the cross in the centre, just as in the Assyrian letter, col. III. — No. 10, *ya*, col. V, = *Yod* (WEBER), the *Yod* of cols. I, II, being laid on its side, col. IV, the central stroke being lengthened, and the pendant on the right being turned upwards, hence first the *ya* of col. V, *a*, and later the cursive forms in col. V, *b*, *c*. — No. 11, *ka*, col. V, *a*, *b*, = *Kaph*, the upper side-bar of a form like that in col. II, having been converted into the top of the vertical, and the sign being then set up straight. — No. 12, *la*, col. V, = *Lamed*, cols. I, II (WEBER), preserved in its original position in the slightly differentiated *ḷ* of the Drāviḍī, col. VI (see below, B, 4, *c*), and in the Eraṇ form, col. IV, with the *Serif* on the top of the curve, turned from right to left in the usual form of the Aśoka edicts, col. V, *a*, and turned with a tail on the right, but without the *Serif*, in the Drāviḍī *l*, col. V, *b*. — No. 13, *ma*, col. V, = *Mem* (WEBER), derived from a form like that in col. II, with the change of the bent pendant into a loop, as in the hypothetical form in col. IV (analogous development in Euting, TSA. col. 58, *a*), and with superposition of the angle on the loop, col. V, *a* (analogous development in Euting, TSA. col. 59, *c*), whence the cursive form with semicircle at the top in col. V, *b*. — No. 14, *na*, col. V, = *Nun* (TAYLOR), the *Nun* in cols. I, II, being turned topsy-turvy as in col. IV, and the hook at the foot being converted into a straight stroke, for which development the *ña*, col. VI, *a*, formed out of the hypothetical sign by a regularisation of the hook and the addition of a differentiating bar at the top (see below B, 4, *d*), appears to be a witness.
c6.s4.t1.3
No. 15, *sa*, *ṣa*, cols. V, IV, = *Samekh* (WEBER doubtfully); a *Samekh* like that of col. I, *b*, being made cursive by the Hindus, as shown in col. IV, and turned topsy-turvy, [14] whereby the Drāviḍa *ṣ*, col. V, was obtained, which originally served both for *s* and *ṣ*. Later, this sign was divided into the signs for the etymologically connected *sa* and *ṣa*. By transferring the cross-bar to the outside of the curve, arose the *sa* of the southern Brāhmī in col. VI, *a*, and (turned round) that in col. VI, *b*, while the removal of the bar to the inside of the curve produced the *ṣa* of the same script, col. VI, *c*. The Drāviḍī adopted the new *sa* for its *s*, and retained the old sign for *ṣ*. The northern Brāhmī developed out of the southern *sa* that with the curve, col. VI, *d*, and out of this a new *ṣa*, col. VI, *e*. An immediate derivation of the Drāviḍa *ṣ* from the *Samekh* of the sixth century B. C. in col. III. is not possible, for the reasons stated under No. 5, and because the characteristic ancient cross-bar is wanting in it. — No. 16, *E*, col. V, = *Ain*, cols. I, II (Weber), the Indian sign being changed slightly or not at all in the ancient forms of Kālsī, col. IV. and col. V, *b*, as well as in that of Sāñci and Hāthigumphā, col. V, *a*, but later made triangular, col. V, *c*, *d*, *e*, in order to avoid a confusion with *tha* and *dha*. — No. 17, *pa*, col. V, = *Phe*, cols. I, II (Weber), turned topsy-turvy; in its original position in the Eraṇ form, col. IV; turned sideways in col. V.
c6.s4.t1.4
No. 18, *ca*, col. V, = *Tsade*, cols. I, II, turned topsy-turvy, the second hook on the right being bent at the same time towards the vertical as in the hypothetical form of col. IV, whence arose, with the turn sideways, the angular or round *ca* of the Brāhmī in col. V, *a*, *b*, and the tailed one of the Drāviḍī, col. V, *c*. — No. 19, *kha*, col. V, = *Qoph*, cols. I, II, turned topsy-turvy with the addition of a curve at the top, col. V, *a*, in order to distinguish the letter from *va*. Owing to the use of ink, the circle at the foot was converted into a dot, col. V, *b*. — No. 20, *ra*, col. V, = *Resh*, cols. I, II (Weber), the triangular head of the letter being opened and the vertical attached to the base of the former triangle, whence arose the forms in col. V, *a*, *b*, and later the ornamental ones, col. V, *c*, *d*, in which the angles were repeated. — No. 21, *śa*, col. V, = *Shin*, cols. I, II (Weber), the two angles, standing side by side, being placed the one inside the other, and the sign being then turned topsy-turvy, col. V, *a*, *b*, *c*. The more closely resembling Aramaic *Shin* of the sixth century B. C., col. III, cannot be the prototype of *śa*, for the same reasons as those stated above under No. 5, and is merely an analogous transformation, which the Aramaens, Phoenicians and Ethiopians have made independently at various periods. The older form with two angles has been preserved in the western sign for 100 = *śu* (see my Indian Studies, III², 71, 117). — No. 22, *ta*, col. V, = *Taw*, cols. I, II (Weber); from a form like that of Sinjirli, col. III, *b*, or the Assyrian of the time of Salmanassar, col. III, *a*, was derived the *ta* of col. V, *a*, *b*, and hence the regularised form of col. V, *c*.
B. — Derivative consonants and initial vowels
c6.s4.t2.1
The derivative signs, invented by the Hindus themselves, have been formed by means of the following contrivances:—
c6.s4.t2.2
(1) One of the elements of a phonetically cognate letter is transposed: (*a*) in *sa* and *sa*, where the cross-bar of the oldest sign has been displaced (see above, A, No. 15); (*b*) in *da*, which has been derived from *dha* (Weber) by dividing the vertical stroke, and by attaching the two pieces to the upper and lower ends of the curve, whence first the *da* of the Drāviḍī and of the Patnā seal, No. 4, col. VI, *a*, was derived, and, with the turn to the left, the ordinary form of the Brāhmī, No. 4, col. VI, *b*, and further the angular *da*, No. 4, col. VI, *f*.
c6.s4.t2.3
(2) A borrowed or derivative letter is mutilated in order to obtain one with a similar phonetic value: (*a*) from *da*, No. 4, col. VI, *a*, comes [15] by the removal of the lower end the half round *ḍa* of Kālsī and the later southern inscriptions, col. VI, *c*; similarly, from the angular *da*, col. VI, *g*, the ordinary angular *ḍa*, col. VI, *h*, of the Aśoka edicts (Weber); (*b*) from *tha*, No. 9, col. V, comes *ṭha*, col. VI, *a*, by the removal of the central dot; and from the latter again *ṭa*, col. VI, *b*, is derived by bisection, the round *ṭha* being considered as the product of an unaspirated letter and a curve of aspiration, which appears (see below, 5) in various other letters (Weber); (*c*) from the triangular *E*, No. 16, col. V, *c*, *d*, *e*, comes the *I* with three dots, col. VI, B, *a*, *b*, *c*, which just indicate the outlines of the older sign (Prinsep), the derivation being suggested by the fact that grammatically *e* is the guṇa-vowel of *i*, for which therefore a lighter form of *e* appeared suitable; (*d*) through a bisection of the lower portion of *va*, No. 6, col. V, *b*, and a straightening of the remaining pendant, is derived *U*, col. VI, *a* (see my Indian Studies, III², 74), the derivation being suggested by the fact that *u* commonly represents *va* in weak grammatical forms (*samprasāraṇa*); (*e*) if the later small circle (pl. IV, 38, VI) is the original form of the Anusvāra, No. 13, col. VI, *a*, *b*, and the dot a cursive substitute, the sign may be explained as a mutilated small *ma*, which has lost the angle at the top, and has been thus treated like the small vowelless consonants appearing in the inscriptions of the first centuries A. D. (see, *e. g.*, pl. III, 41, VIII); compare also the derivation of the Kharoṣṭhī Anusvāra from *ma* (see below, § 9, B, 4).
c6.s4.t2.4
(3) Short horizontal strokes, which originally, before the change in the direction of the writing, stood on the left, are used to derive the long vowels *Ā*, No. 1, col. VI, and *Ū*, No. 6, col. VI, *d*, from short *A* and *U*. On account of the peculiar shape of *I*, a dot is used instead for the formation of *Ī*, No. 16, col. VI, B, *g*.
c6.s4.t2.5
(4) Short horizontal strokes, originally added on the right, denote a change in the quality of the sounds: (*a*) in *O*, No. 6, col. VI, *f*, *g*, derived from *U*, col. VI, *a* (with the bar in the original and the later position), because grammatically *o* is the *guṇa*-vowel of *u*; (*b*) in *AI*, No. 16, col. VI, A, *b*, derived from *E*, because grammatically *ai* is the *vṛddhi*-vowel of *e*; (*c*) in the *ḷ* of the Drāviḍī, No. 12, col. VI, from the original form of *la* (*Lamed*), cols. I, II, in which case the bar still stands on the right, because the letter has not been turned; (*d*) in *ña*, No. 14, col. VI, *a*, from the original inverted *Nun*, col. IV; compare above under A, No. 14; (*e*) in *ṅa* (see my Indian Studies, III², pp. 31, 76; also page 35, below, § 16, C, 12) from *na*, No. 14, col. V, with a displacement of the lower horizontal stroke towards the right, the letter being kept in its original position; (*f*) in *ṇa*, No. 14, col. VI, *b*, from *na*, the bar protruding at both sides of the vertical in order to avoid the identity with *nā*, *ne* and *O*.
c6.s4.t2.6
(5) The aspiration is expressed by a curve in the *gh* of the Drāviḍī, No. 3, col. VI, formed out of *g*, and in the ordinary Brāhmī *dha*, No. 4, col. VI, *d*, from *ḍa*, col. VI, *c*, in *pha*, No. 17, col. VI, from *pa*, col. V, and in *cha*, No. 18, col. VI, *a*; in the last sign the curve has been attached to both ends of the vertical, and this proceeding led to the development of the cursive *cha* of col. VI, *b*. More rarely a hook is substituted for the curve, and then the original sign is mutilated; thus *bha*, No. 2, col. VI, is derived from *ba* by omitting the base-stroke, and *jha*, No. 7, col. VI, from the Drāviḍa *j*, col. V, *a*, by dropping both bars at the ends of the vertical. Both the hook and the curve are cursive substitutes for *ha*, which in the Tibetan alphabet¹ is used again in order to form *gha*, *bha*, &c.
c6.s4.t2.7
¹ *AR.* 2, plate at p. 400.
c6.s4.t2.8
(6) [**16**] The *ḷa* of the Brāhmī, No. 4, col. VI, *e*, has been derived, by the addition of a small semicircle, for which we have an open angle in Sāñci (pl. II, 41, XVIII), from the half round *ḍa* of col. VI, *c*, the derivation being very probably suggested by the phonetic affinity of *ḍa* to *ḷa*, which two letters are frequently exchanged in Vedic and classical Sanskrit and in the Prākrit dialects.
C. — Medial vowels and absence of vowel in ligatures
c6.s4.t3.1
(1) — The system of the Brāhmī
c6.s4.t3.2
In accordance with the expressions of the Sanskrit phonologists and grammarians, who take into account the spoken language alone² and who call the *k*-sound *ka-kāra*, the *g*-sound *ga-kāra*, &c., the medial *a* is inherent in all consonants, and consequently medial *ā* is expressed by the stroke which distinguishes *A* from *Ā*.
c6.s4.t3.3
² M.M.*HASL.* 505 ff.
c6.s4.t3.4
The other medial vowels are either the full initial vowel-signs or cursive derivatives from them, which are placed mostly at the top or rarely at the foot of the consonants. The identity of the medial *o* with the initial *O* is distinctly recognisable in all letters with verticals at the top, as in *ko*, No. 6, col. VI, *h*, *i*, where, on the removal of the dagger-shaped *k* below the second cross-bar, the signs in col. VI, *f*, *g*, reappear; compare also *go* in *mago*, Girnār edict I, line 11, where an initial *O* has been placed above *g*. In the Jaugada edicts, where only the *O* of col. VI, *f*, occurs, the medial *o* has invariably the same form. But in Girnār we have both forms of *o*, though there is only the *O* of col. VI, *g*. Similarly, the full initial *U* is recognisable in the combinations with consonants ending in verticals, as in *ku*, pl. II, 9, V; *ḍu*, 20, VII; *du*, 25, V; *bhu*, 31, III, V (compare § 16, D, 4); and in the *dhu* of Kālsī, No. 6, col. VI, *b*: more usually *u* is represented cursively, either by the horizontal stroke of *U*, as in *dhu*, No. 6, col. VI, *c*, or by its vertical as in *cu*, pl. II, 13, III, and *dhu*, 26, II, &c. Medial *ū* is identical with *Ū*, if combined with consonants ending in verticals; elsewhere it is cursively expressed by two lines, commonly placed horizontally, as in *dhū*, No. 6, col. VI, *e*: but in the later inscriptions we occasionally find the *Ū* of the period used for the medial vowel.¹ Medial *i* was probably at first expressed by the three dots of the initial *I* (*ki*, No. 16, col. VI, B, *d*), which afterwards were joined cursively by lines and converted into the angle used in most of the Aśoka edicts (*ki*, col. VI, B, *e*). The medial *ī* has been developed out of the latter form by the addition of a stroke, indicating that the vowel is long (*kī*, col. VI, B, *f*; see above, under B, 3). In order to express medial *e*, the triangle of the initial *E* has been reduced cursively first to an angle, open on the left, as in *ge*, pl. II, 11, III, and more commonly to a straight line (*ke*, No. 16, col. VI, A, *a*). In accordance with the form of the initial *AI*, which consists of *E* and a horizontal bar, medial *ai* is expressed by two parallel horizontal strokes (*thai*, No. 16, col. VI, A, *c*).
c6.s4.t3.5
¹ See below, § 24, B, 3; pl. IV, 30, XII, XIV; pl. VII, 30, XII, XX, XXI.
c6.s4.t3.6
The absence of a vowel is indicated by interlacing the sign for the consonants immediately following each other, and in such ligatures the second sign is often mutilated; see below, § 16, E, 2. This proceeding appears to be a practical illustration of the term *saṁyuktākṣara*, "a joined or ligature syllable," by which the phonologists and grammarians denote a syllable beginning with more consonants than one.
c6.s4.t3.7
(2) — The system of the Drāviḍī
c6.s4.t3.8
The notation of the medial vowels in the inscriptions of Bhaṭṭiprolu differs from the usual one in so far as medial *a* is marked by the Brāhmī sign for *ā*, and medial *ā* by a horizontal stroke from the end of which a vertical one hangs down; see *ka*, pl. II, 9, XIII; *kā*, 9, XIV. Hence the consonants have no inherent *a*. The device is no doubt of later origin, and has been invented in order to avoid the necessity for ligatures.