1. The Hebrew letters now in use, in which both the manuscripts of the O.T. are written and our editions of the Bible are printed, commonly called the square character (כְּתָב מְרֻבָּע‎), also the Assyrian character (כְּ׳ אַשּׁוּרִי‎), are not those originally employed.

Old Hebrew (or Old Canaanitish) writing, as it was used on public monuments in the beginning of the ninth and in the second half of the eighth century b.c., is to be seen in the inscription of Mêšaʿ, as well as in that of Siloam. The characters on the Maccabaean coins of the second century B.C., and also on ancient gems, still bear much resemblance to this. With the Old Hebrew writing the Phoenician is nearly identical. From the analogy of the history of other kinds of writing, it may be assumed that out of and along with this monumental character, a less antique and in some ways more convenient, rounded style was early developed, for use on softer materials, skins, bark, papyrus, and the like. This the Samaritans retained after their separation from the Jews, while the Jews gradually (between the sixth and the fourth century) exchanged it for an Aramaic character. From this gradually arose (from about the fourth to the middle of the third century) what is called the square character, which consequently bears great resemblance to the extant forms of Aramaic writing, such as the Egyptian-Aramaic, the Nabatean and especially the Palmyrene. Of Hebrew inscriptions in the older square character, that of ˓Arâq al-Emîr (15½ miles north-east of the mouth of the Jordan) probably belongs to 183 B.C.

The Jewish sarcophagus-inscriptions of the time of Christ, found in Jerusalem in 1905, almost without exception exhibit a pure square character. This altered little in the course of centuries, so that the age of a Hebrew MS. cannot easily be determined from the style of the writing. The oldest known biblical fragment is the Nash papyrus (found in 1902), containing the ten commandments and the beginning of Dt 6^{4 f.}, of the end of the first or beginning of the second century A.D. Of actual MSS. of the Bible the oldest is probably one of 820–850 A.D. described by Ginsburg, Introd., p. 469 ff., at the head of his sixty principal MSS.; next in age is the codex of Moses ben Asher at Cairo (897 A.D.). The date (916 A.D.) of the Codex prophetarum Babylon. Petropol. is quite certain.—In the synagogue-rolls a distinction is drawn between the Tam-character (said to be so called from Rabbi Tam, grandson of R. Yiṣḥāqî, in the twelfth century) with its straight strokes, square corners and ‘tittles’ (tāgîn), in German and Polish MSS., and the foreign character with rounded letters and tittles in Spanish MSS.

2. The Alphabet consists, like all Semitic alphabets, solely of consonants, twenty-two in number, some of which, however, have also a kind of vocalic power. The following Table shows their form, names, pronunciation, and numerical value –

FORM.	NAME.	PRONUNCIATION.	NUMERICAL VALUE.
א‎	˒Ālĕph	˒ spiritus lenis	1
ב‎	Bêth	b (bh)	2
ג‎	Gimĕl (Giml)	g (gh)	3
ד‎	Dālĕth	d (dh)	4
ה‎	Hē	h	5
ו‎	Wāw (Wāu)	w (u)	6
ז‎	Záyĭn	z, as in English (soft s)	7
ח‎	Ḥêth	ḥ, a strong guttural	8
ט‎	Ṭêth	ṭ, emphatic t	9
י‎	Yôd	y (i)	10
כ‍‎, final ך‎	Kaph	k (kh)	20
ל‎	Lāmĕd	l	30
מ‍‎, final ם‎	Mêm	m	40
נ‍‎, final ן‎	Nûn	n	50
ס‎	Sāmĕkh	s	60
ע‎	˓Áyĭn	ʿ a peculiar guttural (see below)	70
פ‎, final ף‎	Pê	p (f)	80
צ‍‎, final ץ‎	Ṣādê	ṣ, emphatic s	90
ק‎	Qôf	q, a strong k formed at the back of the palate	100
ר‎	Rêš	r	200
שׂ‎	Śîn	ś	300
שׁ‎	Šîn	š, pronounced sh	300
ת‎	Tāw (Tāu)	t (th)	400

 

3. As the Table shows, five letters have a special form at the end of the word. They are called final letters, and were combined by the Jewish grammarians in the mnemonic word כַּמְנֶפֶץ‎ Kamnèphäṣ, or better, with A. Müller and Stade, כַּמְנַפֵּץ‎ i.e. as the breaker in pieces. Of these, ך‎, ן‎, ף‎, ץ‎ are distinguished from the common form by the shaft being drawn straight down, while in the usual form it is bent round towards the left. In the case of ם‎ the letter is completely closed.

4. Hebrew is read and written from right to left. Words must not be divided at the end of the lines; but, in order that no empty space may be left, in MSS. and printed texts, certain letters suitable for the purpose are dilated at the end or in the middle of the line. In our printed texts these literae dilatabiles are the five following: ﬡ ﬣ ﬥ ﬨ ﬦ‎ (mnemonic word אֲהַלְתֶּם‎ ˒^ahaltèm). In some MSS. other letters suitable for the purpose are also employed in this way, as ד‎, כ‍‎, ר‎.

Rem. 1. The forms of the letters originally represent the rude outlines of perceptible objects, the names of which, respectively, begin with the consonant represented (akrophony). Thus Yôd, in the earlier alphabets the rude picture of a hand, properly denotes hand (Heb. יָד‎), but as a letter simply the sound י‎ (y), with which this word begins; Áyĭn, originally a circle, properly an eye (עַ֫ין‎), stands for the consonant ע‎. In the Phoenician alphabet, especially, the resemblance of the forms to the objects denoted by the name is still for the most part recognizable (see the Table). In some letters (ג‎, ו‎, ז‎, ט‎, ש‎) the similarity is still preserved in the square character.

It is another question whether the present names are all original. They may be merely due to a later, and not always accurate, interpretation of the forms. Moreover, it is possible that in the period from about 1500 to 1000 B.C. the original forms underwent considerable change.

The usual explanation of the present names of the letters is: אָלֶף‎ ox, בֵּית‎ house, גִּמֶל‎ camel (according to Lidzbarski, see below, perhaps originally גַּרְזֶן‎ axe or pick-axe), דָּלֶת‎ (properly folding door; according to Lidzbarski, perhaps דַּד‎ the female breast), הֵא‎ air-hole (?), lattice-window (?), וָו‎ hook, nail, זַיִן‎ weapon (according to Nestle, comparing the Greek ζῆτα, rather זַיִת‎ olive-tree, חֵית‎ fence, barrier (but perhaps only differentiated from ה‎ by the left-hand stroke), טֵית‎ a winding (?), according to others a leather bottle or a snake (but perhaps only differentiated from ת‎ by a circle round it), יוֹד‎ hand, כַּף‎ bent hand, לָמֶד‎ ox-goad, מַיִם‎ water, נוּן‎ fish (‘perhaps originally נָחָשׁ‎ snake,’ as in Ethiopic), סָמֶךְ‎ prop (perhaps a modification of ז‎), עַיִן‎ eye, פֵּא‎ (also פֵּי‎ mouth), צָדֵי‎ fish-hook (?), קוֹף‎ eye of a needle, according to others back of the head (Lidzb., ‘perhaps קֶשֶׁת‎ bow’), רֵישׁ‎ head, שִׁין‎ tooth, תָּו‎ sign, cross.

With regard to the origin of this alphabet, it may be taken as proved that it is not earlier (or very little earlier) than the fifteenth century B.C., since otherwise the el-Amarna tablets would not have been written exclusively in cuneiform. It seems equally certain on various grounds, that it originated on Canaanitish soil. It is, however, still an open question whether the inventors of it borrowed

(a) From the Egyptian system—not, as was formerly supposed, by direct adoption of hieroglyphic signs (an explanation of twelve or thirteen characters was revived by J. Halévy in Rev. Sémit. 1901, p. 356 ff., 1902, p. 331 ff., and in the Verhandlungen des xiii. … Orient.-Kongr. zu Hamb., Leiden, 1904, p. 199 ff.; but cf. Lidzbarski, Ephemeris, i. 261 ff.), or of hieratic characters derived from them (so E. de Rougé), but by the adoption of the acrophonic principle (see e) by which e.g. the hand, in Egyptian tot, represents the letter t, the lion = laboi, the letter l. This view still seems the most probable.

(b) From the Babylonian (cuneiform) system. Wuttke’s and W. Deecke’s derivation of the old-Semitic alphabet from new-Assyrian cuneiform is impossible for chronological reasons. More recently Peters and Hommel have sought to derive it from the old-Babylonian, and Ball from the archaic Assyrian cuneiform. A vigorous discussion has been aroused by the theory of Frdr. Delitzsch (in Die Entstehung des ält. Schriftsystems od. der Urspr. der Keilschriftzeichen dargel., Lpz. 1897; and with the same title ‘Ein Nachwort’, Lpz. 1898, preceded by a very clear outline of the theory) that the old-Semitic alphabet arose in Canaan under the influence both of the Egyptian system (whence the acrophonic principle) and of the old-Babylonian, whence the principle of the graphic representation of objects and ideas by means of simple, and mostly rectilinear, signs. He holds that the choice of the objects was probably (in about fifteen cases) influenced by the Babylonian system. The correspondence of names had all the more effect since, according to Zimmern (ZDMG. 1896, p. 667 ff.), out of twelve names which are certainly identical, eight appear in the same order in the Babylonian arrangement of signs. But it must first be shown that the present names of the ‘Phoenician’ letters really denote the original picture. The identity of the objects may perhaps be due simply to the choice of the commonest things (animals, implements, limbs) in both systems.

The derivation of the Semitic alphabet from the signs of the Zodiac and their names, first attempted by Seyffarth in 1834, has been revived by Winckler, who refers twelve fundamental sounds to the Babylonian Zodiac. Hommel connects the original alphabet with the moon and its phases, and certain constellations. This theory is by no means convincing.

(c) From the hieroglyphic system of writing discovered in 1894 by A. J. Evans in inscriptions in Crete (esp. at Cnossus) and elsewhere. According to Kluge (1897) and others, this represents the ‘Mycenaean script’ used about 3000–1000 B.C., and according to Fries (‘Die neuesten Forschungen über d. Urspr. des phöniz. Alph.’ in ZDPV. xxii. 118 ff.) really supplies the original forms of the Phoenician alphabet as brought to Palestine by the Philistines about 1100 B.C., but ‘the Phoenician-Canaanite-Hebrews gave to the Mycenaean signs names derived from the earlier cuneiform signs’. The hypothesis of Fries is thus connected With that of Delitzsch. But although the derivation of the Phoenician forms from ‘Mycenaean’ types appears in some cases very plausible, in ethers there are grave difficulties, and moreover the date, 1100 B.C., assigned for the introduction of the alphabet is clearly too late.

(d) From a system, derived from Asia Minor, closely related to the Cypriote syllabary (Praetorius, Der Urspr. des kanaan. Alphabets, Berlin, 1906). On this theory the Canaanites transformed the syllabic into an apparently alphabetic writing. In reality, however, they merely retained a single sign for the various syllables, so that e.g. ק‎ is not really q, but qa, qe, qi, &c. Of the five Cypriote vowels also they retained only the star (in Cypriote = a) simplified into an ˒ālef (see alphabetical table) to express the vowels at the beginning of syllables, and i and u as Yod and Waw. Praetorius claims to explain about half the twenty-two Canaanite letters in this way, but there are various objections to his ingenious hypothesis.

2. As to the order of the letters, we possess early evidence in the alphabetic poems: ψ 9 (א‎–כ‍‎, cf. ψ 10¹ ל‎, and vv. ^12–17 ק‎–ת‎; cf. Gray in the Expositor, 1906, p. 233 ff., and Rosenthal, ZAW. 1896, p. 40, who shows that ψ 9^3,15,17 כ‍‎, ל‎, נ‍‎, exactly fit in between ח‎, ט‎, י‎, and that ψ 10^1,3,5 therefore has the reverse order ל‎, ך‎, י‎); also ψψ 25 and 34 (both without a separate ו‎-verse and with פ‎ repeated at the end^[15]); 37, 111, 112, 119 (in which every eight verses begin with the same letter, each strophe, as discovered by D. H. Müller of Vienna, containing the eight leading words of ψ 19^{8 ff.}, tôrā, ˓edûth, &c.); La 1–4 (in 2–4 פ‎ before ע‎, in chap. 3 every three verses with the same initial, see Löhr, ZAW. 1904, p. 1 ff., in chap. 3 at any rate as many verses as letters in the alphabet); Pr 24^1,3,5, 24^10–31 (in the LXX with פ‎ before ע‎^[16]); also in Na 1^2–10 Pastor Frohnmeyer of Württemberg (ob. 1880) detected traces of an alphabetic arrangement, but the attempt of Gunkel, Bickell, Arnold (ZAW. 1901, p. 225 ff.), Happel (Der Ps. Nah, Würzb. 1900) to discover further traces, has not been successful. [Cf. Gray in Expositor, 1898, p. 207 ff.; Driver, in the Century Bible, Nahum, p. 26.]—Bickell, Ztschr f. Kath. Theol., 1882, p. 319 ff., had already deduced from the versions the alphabetical character of Ecclus 51^13-30, with the omission of the ו‎-verse and with פ‎ at the end. His conjectures have been brilliantly confirmed by the discovery of the Hebrew original, although the order from ג‎ to ל‎ is partly disturbed or obscured. If ו‎ before צ‎ is deleted, ten letters are in their right positions, and seven can be restored to their places with certainty.

The sequence of the three softest labial, palatal, and dental sounds ב‎, ג‎, ד‎, and of the three liquids ל‎, מ‍‎, נ‍‎, indicates an attempt at classification. At the same time other considerations also appear to have had influence. Thus it is certainly not accidental, that two letters, representing a hand (Yôd, Kaph), as also two (if Qôph = back of the head) which represent the head, and in general several forms denoting objects naturally connected (Mêm and Nûn, ˓Ayĭn and Pê), stand next to one another.

The order, names, and numerical values of the letters have passed over from the Phoenicians to the Greeks, in whose alphabet the letters Α to Υ are borrowed from the Old Semitic. So also the Old Italic alphabets as well as the Roman, and consequently all alphabets derived either from this or from the Greek, are directly or indirectly dependent on the Phoenician.

3. a. In default of special arithmetical figures, the consonants were used also as numerical signs. The earliest traces of this usage are, however, first found on the Maccabean coins. These numerical letters were afterwards commonly employed, e.g. for marking the numbers of chapters and verses in the editions of the Bible. The units are denoted by א‎–ט‎, the tens by י‎–צ‍‎, 100–400 by ק‎–ת‎, the numbers from 500–900 by ת‎ (=400), with the addition of the remaining hundreds, e.g. תק‎ 500. In compound numbers the greater precedes (on the right), thus יא‎ 11, קכא‎ 121. But 15 is expressed by טו‎ 9+6, not יה‎ (which is a form of the divine name, being the first two consonants of יהוה‎). For a similar reason טז‎ is also mostly written for 16, instead of יו‎, which in compound proper names, like יוֹאֵל‎, also represents the name of God, יהוה‎.

The thousands are sometimes denoted by the units with two dots placed above, e.g. אׄׄ‎ 1000.

b. The reckoning of the years in Jewish writings (generally ליצירה‎ after the creation) follows either the full chronology (לִפְרָט גָּדוֹל‎ or לפ׳ ג׳‎), with the addition of the thousands, or the abridged chronology (לפ׳ קָטוֹן‎), in which they are omitted. In the dates of the first thousand years after Christ, the Christian era is obtained by the addition of 240, in the second thousand years by the addition of 1240 (i.e. if the date falls between Jan. 1 and the Jewish new year; otherwise add 1239), the thousands of the Creation era being omitted.

4. Abbreviations of words are not found in the text of the O.T., but they occur on coins, and their use is extremely frequent amongst the later Jews. A point, or later an oblique stroke, serves as the sign of abridgement in old MSS. and editions, e.g. ישׂ׳‎ for יִשְׂרָאֵל‎, פ׳‎ for פְּלֹנִי‎ aliquis, ד׳‎ for דָּבָר‎ aliquid, וגו׳‎ for וְגוֹמַר‎ et complens, i.e. and so on. Also in the middle of what is apparently a word, such strokes indicate that it is an abbreviation or a vox memorialis (cf. e.g. תא״ם‎). Two such strokes are employed to mark the different classes of weak verbs.—Note also יְיָ‎ or יָי‎ (also ה׳‎) for יְהוָֹה‎.

5. Peculiarities in the tradition of the O.T. text, which are already mentioned in the Talmud, are—

(1) The 15 puncta extraordinaria, about which the tradition (from Siphri on Nu 9¹⁰ onwards) differs considerably, even as to their number; on particular consonants, Gn 16⁵, 18⁹, 19^33.35, Nu 9¹⁰; or on whole words, Gn 33⁴, Gn 37¹², Nu 3³⁹, 21³⁰, 29¹⁵, Dt 29²⁸, 2 S 19²⁰, Is 44⁹, Ez 41²⁰, 18⁹, ψ 27¹³, —all no doubt critical marks; cf. Strack, Prolegomena Critica, p. 88 ff.; L. Blau, Masoretische Untersuchungen, Strassburg, 1891, p. 6 ff., and Einleitung in die hl. Schrift, Budapest, 1894; Königsberger, Jüd. Lit.-Blatt, 1891, nos. 29–31, and Aus Masorah u. Talmudkritik, Berlin, 1892, p. 6 ff.; Mayer-Lambert, REJ. 30 (1895), no. 59; and especially Ginsburg, Introd., p. 318 ff.; also on the ten points found in the Pentateuch, see Butin (Baltimore, 1906), who considers that they are as old as the Christian era and probably mark a letter, &c., to be deleted.

(2) The literae majusculae (e.g. ב‎ Gn 1¹, ו‎ Lv 11⁴² as the middle consonant of the Pentateuch, י‎ Nu 14¹⁷), and minusculae (e.g. ה‎ Gn 2⁴).

(3) The literae suspensae (Ginsburg, Introd., p. 334 ff.) נ‍‎ Ju 18³⁰ (which points to the reading משֶׁה‎ for מְנַשֶּׁה‎, ע‎ ψ 80¹⁴ (the middle of the Psalms) and Jb 38¹³, 15.

(4) The ‘mutilated’ Wāw in שלום‎ Nu 25¹², and ק‎ Ex 32²⁵ (בקמיהם‎), and Nu 7² (הפקודים‎).

(5) Mêm clausum in לםרבה‎ Is 9⁶, and Mêm apertum in המ‍ פרוצים‎ Neh 2¹³.

(6) Nûn inversum before Nu 10³⁵, and after ver. 36, as also before ψ 107^23–28 and ⁴⁰; according to Ginsburg, Introd., p. 341 ff., a sort of bracket to indicate that the verses are out of place; cf. Krauss, ZAW. 1902, p. 57 ff., who regards the inverted Nûns as an imitation of the Greek obelus.