(a) In Assyrian ‘one’ is represented by a vertical wedge (Y), and the other units by combinations of such wedges, but ‘ten’ by a sign which is quite similar to the sign for u (<<, cf. in Delitzsch’s Assyr, Gramm. p. 18 with p. 40). The other numbers are indicated by combinations of this sign for ‘ten’ with the vertical and the horizontal wedge. These Assyrian figures might be called purely linear, were it not that the number ‘sixty’ is expressed by ‘| Susu, or soss’; cf. further, Bezold, Oriental Diplomacy (London, 1893), p. 120f., and, above all, Th. Dangin, Recherches sur VOrigine de Vécriture cunéiforme (Paris, 1898), pp. 82 ff., where the figures employed in the oldest cuneiform inscriptions are collected with great completeness.
(8) In the hieroglyphic texts of the Egyptians ‘one’ is indicated by a vertical line, and the num- bers from ‘ two’ to ‘nine’ by vertical strokes placed side by side (e.g. Ill 111). ‘In dates the units are indicated also by horizontal strokes (—, =, etc.).’ But the sign for ‘ten’ is ἢ, ‘hundred’ is repre- sented by C, ete. (cf. Erman, Aegypt, Gramm. 1894, § 140). Essentially identical is the Phanictan system of figures: | to III III ΠῚ; ‘ten’ is indicated by “τ or by a similar obliquely drawn and curved line which evidently arose from O, the earlier form of y, with which the word roy ‘ten’ begins. Then follows a special sign for ‘twenty’ and for ‘hundred’ (ef. Schréder, Die Phin. Sprache, Pp 186 ff, and CIS i. 30, 40, 43, 50, 94, ete.). Only the sign O for ‘ten’ has been found up till now in the Zinjirli inscriptions, namely ‘of = 30,’ and #900°=70’ (Sachau, Ausgrabungen in Sendschirli, 3893, p. 71). Upon the same principle the signs for numbers are chosen in Minwo-Sabeun, where **one” is expressed by a vertical stroke’ (Priitorius, ZDMG xxvi. p. 750), but ‘five’ by J, the initial letter of ry (i) 94, if the Minwo-Sabwan letters
VOL. I11,—36
NUMBER
are transcribed in Ethiopic. The number ‘ten’ is indicated by the sign O, an older form of (7 (y), with which the word for ‘ten’ begins which answers to the Ethiopic ΠΟΘΈΝ. (For the other figures see Priitorius, /.c., and Hommel, Siidarab. Chrestomathie, 1893, p. 8.). Only slightly modified is the system of figures which one finds employed in the Palmyrene inscriptions, namely | to III; ‘five’ = a sign which appears to the present writer to be a simplification of the above 5. Arabian © ; ‘ten’=a sign which may have arisen from O (y), etc. (cf. Merx, Gramm. Syr. p. 17). This second
rinciple upon which numbers are indicated may
called the lineo-acrostic.
(y) In India an older system of figures was dis- placed by that which is adopted in the Sanskrit texts: Y, 3,3, ete. (cf. e.g. Stenzler, Hlementar- buch der Sanskrit-Sprache, §7). This way of in- dicating numbers is the pure acrostic. For the sign % represents the vowel Ὁ, with which the word &q_ (eka, ‘one’) begins, ete. These figures
are employed also by the Arabs (ef. |, Γ΄, I~, ete.), who themselves call this method of indicating numbers ar-rakmu-lhindijju (Caspari-Miiller, Arab. Gramm.° § 33), while Europeans are accustomed to call it the Arabic method.
(5) The fourth leading method of shortening the expression of numbers is the alphabetic. The following traces of it have been noted by the
resent writer: the Greek inscriptions of older
ate show the following figures, I, 11, III, Ill, Γ (S. Reinach, l.c. p. 217, recalls the Π of ITENTE), Tl, ete., A (cf. AEKA), ete. Similar signs are found in inscriptions from Epidauros belonging to the 4th cent. B.c. According to B. Keil (in Hermes, xxv. p. 319), as the present writer’s colleague, G. Kérte, has pointed out to him, the latest specimens of this system are found in CJ Attic. ii. 2, No 985 (written c. 90 B.c.). But somewhat earlier than B.c. 50 the alphabetic hopes of figures appears to have been introduced, according to B.
eil (in above-cited art. p. 320), and it is found, e.g., in CI Attic. iii. 644 (the time of Augustus or Claudius), ete. ‘In the oldest system of this class, the letters possess the following values: A=], B=2, T=3, A=4, E=6, T=6, H=7, 9=8, I=9, K=10, ete.’ (Reinach, 1.6. p. 220). It is clear from all this that Gow (‘The Greek Numeral Alphabet,’ in Journal of Philology, 1884, p. 278) has rightly rejected the hypothesis of a Pheenician origin for this Greek method of indicating numbers. The alphabetic method adopted for Greek figures was copied in Coptic-Arabic and in Ethiopic writ- ings (Pritorius, deth. Gramm. § 14). Further, in many Syriac manuscripts (cf. the Codices Musei Britannici enumerated by Land in his Anecdota Syriaca, p. 94) one finds signs for numbers which have a genetic connexion with the above-mentioned figures of the Palmyrene inscriptions (cf. further, on the notation of the Syrians, Gottheil, ZD./c, 1889, p. 121 ff.). But these figures, which occur pretty frequently in the Codices of 5th-7th cent., afterwards fell into disuse (Merx, Gramm. Syr. p. 16), and the alphabetic method of indicating numbers was adopted (¢.g. Jiid=10; Kaph =20, etc.) ; cf. further, Néldeke, Syr. Gramm. pe 279. This alphabetic method was, and is still, largely employed by the Arads (Caspari-Miiller’, § 33). It was also partially adopted by the Nabateans, in whose inscriptions one finds ‘a mixed system’ of figures (Sachau, ZDMG, 1884, Ρ. 541: ‘ten=Jod, and gence a wet and the same method is not unexampled even in New Persian (cf. Salemann-Shukowski, Neupers. Gramm. p. 4f.).