probable that the Hebrew which incorporated it was also written in Aramaic characters. It may well be that Esther and Ecclesiastes were originally written in Aramaic characters, as well as many of the Apocrypha. There can be little doubt that the Psalter,' Proverbs, Job, and Lamentations were originally written with the ancient letters. It is also probable in the case of Ezra, Nehemiah, Chronicles, and Ruth. It is doubtful with the other writings.

During this period of the formation of the official Canon, and of the substitution of the Aramaic characters for the Hebrew, there were certain changes in the text which have left. their permanent traces.

(a) Emendations were made chiefly for religious reasons.

The substitution of the word Lord, ", for the divine name Yahweh,, was certainly prior to the earliest layer of the Septuagint Version; for kúptos is constantly substituted for it. There are traces of such substitution in the Hebrew text itself.

The substitution of Bosheth,, shame, for Baal,, the god of the Canaanites, and also for Baal in proper names compounded with Baal, was made before the Septuagint translation of the Prophets, but was not thoroughly carried out in all the texts.* The change in proper names is usual in Samuel, where the Chronicler preserves the original form. This seems to indicate that this change was made by the scribes chiefly in the time before the final admission of Chronicles into the Canon. The

1 Perles (Analekten, 1895, pp. 50 seq.) gives examples of errors in the Psalter and Job, which can only come from the ancient Hebrew letters.

2 Baumgartner (Etude critique sur l'Etat du Texte du Livre des Proverbes, Leipzig, 1890) makes it plain that, while the larger proportion of the errors of transliteration in the text of Proverbs is due to mistakes in the distinguishing of similar letters of the Egyptian Aramaic alphabet, and a smaller number to mistakes in the older Aramaic alphabet, there is still a limited number that can be explained only by the ancient Hebrew alphabet.

8 Ginsburg (Introduction, p. 293) gives Ezra 6+ as an example of a mistake of Aleph for Tav in the old Hebrew alphabet. But Baumgartner (1.c., s. 279) thinks that such mistakes might be as well explained from the ancient Aramaic alphabet also.

4 Cf. ǹ Báaλ, Jer. 223, 79, 1113. 17, 195; Hos. 210, 131; Rom. 114; which implies the reading of aloxúvn for ßáaλ. See Dillmann, Baal mit d. weibl. Artikel. in the Monatsberichte d. Königl. Acad. d. Wiss. zu Berlin, 1881.

5 However, in 2 Sam. 1121 the Septuagint, Syriac, and Vulgate versions all read, and in 2 Sam. 238 Lucian's text of the Septuagint preserves 'Ico ßáaλ.

same is true of the reading of Shame, Bosheth, , for King, Melekh,, when applied to the god of the Ammonites.1

(b) The earlier scribes also acted as editors. They divided first the Law and then the Psalter into five books. These divisions are not logical divisions. The natural divisions in both cases would be into three books. The divisions are mechanical, and they were doubtless made for liturgical reasons. Another ancient division for both the Law and the Psalter, into seven books, is mentioned in the Talmud.2 These divisions all may have reference to the use of the Law and the Psalter at the feasts of the Jews.

(e) The scribes also divided the sacred books into sections. These sections do not correspond altogether with the later sections of the Talmudic and Massoretic periods, but they were doubtless arranged for public reading in the synagogues. Two such sections are mentioned in the New Testament.3

(d) No verses are known so far as prose writings are concerned; but the ancient poems in the historical books, and the poetical books of Psalms, Lamentations, and the Wisdom Literature, were certainly written in distich, tristich, tetrastich, and the like. It is probable that the greater portion of the poetry in other books was written in this way also. This enabled Josephus and even Jerome to speak of trimeters, tetrameters, and hexameters. But this method of writing poetry was subsequently lost, except for the ancient poems in the Pentateuch, because of the Massoretic system of accentuation for cantilation in the synagogue.1

II.

THE TEXT OF THE CANON OF THE SOPHERIM

There is no evidence of any attempt to establish an official Hebrew text until after the destruction of Jerusalem by the

1 Lev. 1821 (Sept. B apxwv); 202-5 (Sept. dpxwv); K. 117 (Sept. Baoiλeús);

2 K. 2310 (Sept. Móλox); Jer. 3235 (Sept. Moλòx Baσiλeús).

2 Talm. Shabboth, 115 b, 116a; Midrash Bereshith Rabba, LXIV. fol. 71 d, Num. 1035; Vayyikra Rabba, Lev. 91; Rashi on Prov. 91.

8 The section of the Bush ñì тoû ẞárov Mk. 1226, referring to Ex. 3, and ev 'Hλela Rom. 112, referring to the story of Elijah, 1 K. 19, are the only two known to the New Testament.

4 See Chap. XIV. pp. 362, 363.

Romans in 70 A.D. There was indeed a codex of the Law in the temple, which was taken by Titus to Rome among the spoils. But the ancient Greek Version, the ancient Syriac Version, the earliest Aramaic Targums, and the citations in the New Testament, the Book of Jubilees,2 and other writings of the first and second centuries B.C. and the first century A.D., make it evident that there was no official Hebrew text until the second century A.D.

After the destruction of Jerusalem the scribes made a rally at Jamnia, where they established a school and held several assemblies. They determined the extent of the Canon and occupied themselves with fixing the text of the manuscripts which had been saved from the wreck of war. There can be no doubt that Rabbi Akiba and his associates at Jamnia not only fixed the Canon of the Old Testament, but also established the first official Hebrew text of the Canon. There is a fixture in the consonantal text of Hebrew manuscripts from the second century onwards, which can be accounted for only by the establishment at that time of such an official text.5 This text was established in troublous times, when it was impossible to give the time and painstaking required for such an undertaking. There was no leisure to correct even the plainest mistakes. It was made by the comparison of a few manuscripts. Tradition speaks of three, in cases of disagreement the majority of two always determining the correct reading.

1 Josephus, B. J., VII. 5, § 5. This is said to have been given by the Emperor Severus, about 220 A.D., to a synagogue built by him at Rome. Ginsburg, (l.c., pp. 410 seq.) gives a list of thirty-two readings said to have been taken from this codex.

2 The Book of Jubilees, or Little Genesis, as it is sometimes called, testifies to a text somewhat different from that of the Sopherim. See Dillmann, Beiträge aus. d. Buch d. Jubiläen z. Kritik. d. Pentateuch-Textes, Sitzungsberichte d. König. Preus. Akad. der Wissenschaften, 1883. The same is true with reference to other pseudepigrapha.

3 See pp. 130, 131.

4 See Bacher, Hebr. Sprachwissenschaft, 1892, s. 2.

5 Olshausen, Psalmen, s. 18; Lagarde, Anm. z. Griech. Uebersetzung d. Proverbien, 1863, s. 444 seq.; Kuenen, Gesammelte Abhandlungen, 1894, s. 83 seq. This is denied by Hermann Strack, in Semitic Studies in Memory of A. Kohut, 1897, p. 571, on the ground that he has found in ancient manuscripts a very great number of various readings which are unknown to scholars.

6 Cornill, Ezechiel, 1886, s. 10.

The Sopherim found in the court of the temple the codex 190,
In one they found

D; and they accepted two, and rejected one. In one

הוא and the codex ,זעטוטי and the codex מענה Deut. 33), and in two written) מעון אלהי קדם written

Ex. 245), and) וישלח את זעטוטי בני ישראל they found written and they accepted ; וישלח את נערי בני ישראל in two written

two, and rejected one. In one they found written nine times instead of 7, and in two written eleven times ; and they accepted the two, and rejected the one.1

Some scholars think that all manuscripts varying from the official text were ruthlessly destroyed.2 Whether this was so or not, it is altogether probable that the destruction of manuscripts during the war of Hadrian (132-135 A.D.) would so reduce the number of competing manuscripts, that the official manuscripts of the scribes would gain the supremacy.

The official text of the Hebrew Bible in the second Christian century was composed of consonantal letters alone. Even the quiescent letters,3 which were used in ancient times, before the invention of vowel points, to indicate the vowel in difficult words, were not used with any precision; and later scribes were free to exercise their own judgment in the use of them. And so the Massoretic text perpetuates a great lack of uniformity and even inaccuracy of usage. The text used by the translators of the Septuagint was without separation of words and without the final letters, and also with occasional abbreviations; but the Sopherim of the second and third centuries made the separation of words, introduced the five final letters, and removed all abbreviations.5 The work of the Sopherim continued until the sixth century, when the Massorites began their labours. The work of the Sopherim, as described in the Talmud and early Rabbinical commentaries, was:

(1) the fixing of the pronunciation of certain words;

(2) the removal of certain superfluous particles from the text;

1 Jerusalem Taanith, IV. 2; Sopherim, VI. 4. See Ginsburg, Introduction

to Hebr. Bible, pp. 408, 409, who gives text and translation.

2 Nöldeke, Hilgenfeld's Zeitschrift, 1873, s. 444 seq.; W. R. Smith, Old Testament in the Jewish Church, 2d ed., pp. 62 seq.

יוהא 3

* Ginsburg, Introduction, pp. 137 seq.; Perles, Analekten, s. 35.

5 Ginsburg, Introduction, pp. 297 seq.

(3) the mention of words which, though not written, yet ought to be read, and the designation of words which, though written, ought not to be read.

The Babylonian Talmud gives these three under the technical

קריין ולא (3) ; עיטור סופרים (2) ; מקרא סופרים (1) :terms ארץ,As examples of the first are וכתיבן ולא קריין כתיבן The second .מצרים,שמים,when alone or preceded by the article

gives five instances in which the conjunction Waw, and, is to be omitted (Gen. 185, 245; Nu. 312; Pss. 367, 686). The third mentions that , Euphrates, is to be inserted (2 Sam. 8); N, man (2 Sam. 16); D, they are coming (Jer. 31); , to her (Jer. 50); (Ruth 2");, to me (Ruth 35, 17); and the following words are not to be read: ) (2 K. 518); 1 (Jer. 32"); 77, let him bend (Jer. 513); W, five (Ezek. 4816); and N, if (Ruth 319). Nedarim, 37 b-38 a. These are only specimens of a larger number of instances in these departments which are given in later times. (4) Extraordinary points were placed above letters or words to indicate that they were spurious.

The Siphri, the earliest Midrash, or commentary on Numbers, gives ten of these, Nu. 910; Ge. 165, 189, 193, 334, 3712; Nu. 210, 39, 2915; Deut. 2928, all in the Pentateuch. They were subsequently increased to fifteen by adding four from the Prophets, 2 Sam. 1920; Is. 44°; Ezek. 4120, 4622,- and one from the Writings, Ps. 27131

(5) Letters were suspended in order to express doubt as to their propriety.

, in Jud. 1830, changes Moses to Manasseh in order to remove reproach from the name of Moses. , in Ps. 8011, indicates a doubtful reading, as between, the Nile, and, forest; and a preference for the latter with possibly a reference to Rome instead of the original reference to Egypt. The other two instances (Job 38 13, 15) indicate a preference for Dy over DW¬, in order not to offend the dignity of David and of Nehemiah.2

(6) The letter Nun was inverted before and after a clause, in order to indicate bracketed material, which was, in the opinion of the scribes, out of place.3

1 See Ginsburg, l.c., pp. 319 seq., who gives the original, a translation, and comments on the fifteen examples. 2 Sanhedrin, 10, 3 b.

3 Numbers 1035, 36; Ps. 10723, 24, 25, 26, 27, 28, 29 ; so Siphri on Nu. 1035, Talm. Sabbath, 115 b-116 a; Sopherim, VI. 1.

N