________________
$5.]
INDIAN PALEOGRAPHY.
15
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 Girnar we have both forms of o, though there is only the 0 of col. VI, g. Similarly, the full initial U is recognisable in the combinations with consonants ending in verticals, as in lcu, pl. II, 9, V; du, 20, VII ; du, 25, V; bhu, 31, III, V (compare $ 16, D, 4); and in the dhu of Kälsi, No. 6, col. VI, 6: 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, 18, III, and dh , 26, 11, &c. Medial is identical with 0, 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 7 of the period used for the medial vowel. 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 Asoka edicts (loi, col. VI, B, e). The medial i has been developed out of the latter form by the addition of a stroke, indicating that the vowel is long (li, 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 B and a horizontal bar, medial ai is expressed by two parallel horizontal strokes (thai, No. 16, col. VI, A, c).
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 sainyuktākşara, "joined or ligature syllable," by which the phonologists and grammarians denote a syllable beginning with more consonants than one.
(2) - The system of the Drāvidi. The notation of the medial vowels in the inscriptions of Bhattiprolu differs from the usual one in so far as medial a is marked by the Brāhmi sign for , and medial å by a horizontal stroke from the end of which a vertical one hangs down; see ka, pl. II, 9, XIII; ka, 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.
65.-The time and the manner of the borrowing of the Semitio
alphabet. [17] According to the preceding discussion, the great majority of the Brāhma letters agree with the oldest types of the North-Semitic signs, which are found in the archaic Phoenician inscriptions and on the stone of Mesa, incised about B. C.890. But two characters, ha and ta, are derived from Mesopotamian forms of He and Taw, which belong to the middle of the eighth century B. C., and two, 8a-sa and ea, resemble A ramaic signs of the sixth century B. C. As the literary and epigraphic evidence leaves no doubt that the Hindus were not unlettered during the period B. C. 600-500, and as the other signs of the Aramaic alphabet of this period, each as Beth, Daleth, Ware, &c., are too far advanced to be considered as the prototypes of the corresponding Brähma letters, it becomes necessary to regard the seemingly modern forms of sa, ga and ka as the results of an Indian development, analogous to that of the corresponding Aramaic characters. This assumption, of course, remains tenable only as long as the two Aramaic letters are not shown to be more ancient by new epigraphic discoveries, which event, to judge from the results of the Sinjirli finds, does not seem to be impossible. But, for the present, they must be left out of consideration in fixing the terminus a quo
See below, $ 21, B, 3; pl. IV, 30, XII, XIV; pl. VII, 30, XII, XX, XXI. 2 B.18. 111%, 83-91,
