SEPTUAGINT, 2 [ISBE]
- VI. Reconstruction of Septuagint Text; Versions, Manuscripts and Printed Editions.
The task of restoring the original text is beset with difficulties. The materials (MSS, VSS, patristic citations) are abundant, but none has escaped "mixture," and the principles for reconstruction are not yet securely established (Swete, Introduction, I, iv-vi; III, vi).
1. Ancient Versions Made from Septuagint:
Among the chief aids to restoration are the daughter versions made from the Septuagint, and above all the Old Latin (pre-Hieronymian) version, for the earliest (African) Old Latin version dates from the 2nd century AD, i.e. before Origen, and contains a text from which the asterisked passages in Hexaplaric manuscripts are absent; it thus "brings us the best independent proof we have that the Hexaplar signs introduced by Origen can be relied on for the reconstruction of the LXX" (Burkitt). The Old Latin also enables us to recognize the ancient element in the Lucianic recension. But the Latin evidence itself is by no means unanimous. Augustine (De Doctr. Christ., ii.16) speaks of the infinite variety of Latin VSS; though they may ultimately prove all to fall into two main families, African and European. Peter Sabatier's collection of patristic quotations from the Old Latin is still useful, though needing verification by recent editions of the Fathers. Of Old Latin manuscripts one of the most important is the codex Lugdunensis, edited by U. Robert (Pentateuchi e codex Lugd. versio Latin antiquissima, Paris, 1881; Heptateuchi partis post. versio Latin antiq. e codex Lugd., Lyons, 1900). The student should consult also Burkitt's edition of The Rules of Tyconius ("Texts and Studies," III, 1, Cambridge, 1894) and The Old Latin and the Itala (ibid., IV, 3, 1896).
Jerome's Vulgate is mainly a direct translation from the Hebrew, but the Vulgate (Jerome's Latin Bible, 390-405 A.D.) Psalter, the so-called Gallican, is one of Jerome's two revisions of the Old Latin, not his later version from the Hebrew, and some details in our Prayer-book Psalter are ultimately derived through the Vulgate (Jerome's Latin Bible, 390-405 A.D.) Psalter from the Septuagint. Parts of the Apocrypha (Wisdom, Ecclesiasticus, Baruch, 1 and 2 Maccabees) are also pure Old Latin, untouched by Jerome.
The early date (2nd century AD) once claimed for the Egyptian or Coptic versions (Bohairic, i.e. in the dialect of Lower Egypt, Sahidic or Upper Egyptian and Middle Egyptian) has not been confirmed by later researches, at least as regards the first-named, which is probably not earlier than the 3rd or 4th century AD. Rahlfs (Sept-Studien, II, 1907) identifies the Bohairic Psalter as the Hesychian recension. The Sahidic version of Job has fortunately preserved the shorter text lacking the later insertions from Theodotion (Lagarde, Mittheilungen, 1884, 204); this does not conclusively prove that it is pre-Origenic; it may be merely a Hexaplaric text with the asterisked passages omitted (Burkitt, EB, IV, 5027). The influence bf the Hexapla is traceable elsewhere in this version
The Ethiopic version was made in the main from the Greek and in part at least from an early text; Rahlfs (Sept. Stud., I, 1904) considers its text of S-K, with that of codex B, to be pre-Origenic.
The Vulgate (Jerome's Latin Bible, 390-405 A.D.) or Peshitta Syriac version was made from the Hebrew, though partly influenced by the Septuagint. But another Syriac version is of primary importance for the Septuagint text, namely, that of Paul, bishop of Tella (Constantine in Mesopotamia), executed at Alexandria in 616-17 and known as the Syro-Hexaplar. This is a bald Syriac version of the Septuagint column of the Hexapla, containing the Hexaplar signs. A manuscript of the poetical and prophetical books is in the Ambrosian Library at Milan and has been edited by Ceriani (Monumenta sacra et profana, 1874); fragments of the historical books are also extant (Lagarde and Rahlfs, Bibliothecae Syriacae, Gottingen, 1892). This version supplements the Greek Hexaplaric manuscripts and is the principal authority for Origen's text. For the original version of Daniel, which has survived in only one late MS, the Syro-Hexaplar supplies a second and older authority of great value.
The Armenian version (ascribed to the 5th century) also owes its value to its extreme literalness; its text of the Octateuch is largely Hexaplaric.
A bare mention must suffice of the Arabic version (of which the prophetical and poetical books, Job excluded, were rendered from the Septuagint); the fragments of the Gothic version (made from the Lucianic recension), and the Slavonic (partly from Septuagint, also Lucianic) and the Georgian versions.
For a full description of the Greek manuscripts see Swete, Introduction, I, chapter V. They are divided according to their script (capitals or minuscules) into uncials and cursives, the former ranging from the 4th century (four papyrus scraps go back to the 3rd century; Nestle in Hauck-Herzog, Realencyklopadie fur protestantische Theologie und Kirche, XXIII, 208) to the 10th century AD, the latter from the 9th to the 16th century AD. Complete Bibles are few; the majority contain groups of books only, such as the Pentateuch, Octateuch (Gen-Ruth), the later historical books, the Psalter, the 3 or 5 "Solomonic" books, the Prophets (major, minor or both). Uncials are commonly denoted by capital letters (in the edition of Holmes and Parsons by Roman figures); cursives, of which over 300 are known, by Arabic figures; in the larger Cambridge Septuagint the selected cursives are denoted by small Roman letters.
The following are the chief uncials containing, or which once contained, the whole Bible: B (Vaticanus, at Rome, 4th century AD), adopted as the standard text in all recent editions; Codex Sinaiticus, at Petersburg and Leipzig, 4th century AD), discovered by Tischendorf in 1844 and subsequent years in Catherine's Convent, Mt. Sinai; A (Alexandrinus, British Museum, probably 5th century AD); C (Ephraemi rescriptus, Paris, probably 5th century), a palimpsest, the older Biblical matter underlying a medieval Greek text of works of Ephrem the Syrian. For the Octateuch and historical books: D (Cottonianus, British Museum, probably 5th or 6th century), fragments of an illuminated Gen, the bulk of which perished in a fire at Ashburnham House in 1731, but earlier collations of Grabe and others are extant, which for the lost portions are cited in the Cambridge texts as D (Dsil, i.e. silet Grabius, denotes an inference from Grabe's silence that the manuscript did not contain a variant); F (Ambro-sianus, Milan, 4th to 5th century), fragments of the Octateuch; G (Sarravianus, fragments at Leyden, Paris and Petersburg, 4th to 5th century), important as containing an Origenic text with the Hexaplar signs; L (Purpureus Vindobonensis, Vienna, 5th to 6th century), fragments of an illuminated manuscript Genesis on purple vellum; M (Coislinianus, Paris, 7th century), important on account of its marginal Hexaplaric matter. For the Prophets, Q (Marchalianus, Rome, 6th century) is valuable, both for its text, which is "Hesychian" (see above), and for its abundant marginal Hexaplaric matter. A curious mixture of uncial and cursive writing occurs in E (Bodleianus, probably 10th century), fragments of the historical books (to 3 R 16 28) preserved at Oxford, Cambridge (1 leaf), Petersburg and London; Tischendorf, who brought the manuscript from the East, retained the tell-tale Cambridge leaf, on which the transition from uncial to cursive script occurs, until his death. The long-concealed fact that the scattered fragments were part of a single manuscript came to light through Swete's identification of the Cambridge leaf as a continuation of the Bodleian fragment. Many of the cursives still await investigation, as do also the lectionaries. The latter, though the manuscripts are mainly late, should repay study. The use of the Septuagint for lectionary purposes was inherited by the church from the synagogue, and the course of lessons may partly represent an old system; light may also be expected from them on the local distribution of various types of text.
3. Printed Texts:
Of the printed text the first four editions were (1) the Complutensian Polyglot of Cardinal Ximenes, 1514-17, comprising the Greek, Hebrew and Vulgate (Jerome's Latin Bible, 390-405 A.D.) texts, the last in the middle place of honor being compared to Jesus in the midst between the two thieves (!). The Greek was based on manuscripts from the Vatican and one from Venice; it exhibits on the whole the Lucianic recension, as the Hesychian is by a curious coincidence represented in (2) the Aldine edition of 1518, based on Venetian manuscripts. (3) The monumental Sixtine edition, published at Rome in 1586 under the auspices of Pope Sixtus V and frequently reprinted, was mainly based on the codex Vaticanus, the superiority of which text is justly recognized in the interesting preface (printed in Swete's Intro). (4) The English edition (Oxford, 1707-20) begun by Grabe (died 1712) was based on the codex Alexandrinus, with aid from other manuscripts, and had the peculiarity that he employed Origen's critical signs and different sizes of type to show the divergence between the Greek and the Hebrew. Of more recent editions three are preeminent. (5) The great Oxford edition of Holmes and Parsons (Oxford, 1798-1827, 5 volumes, folio) was the first attempt to bring together in a gigantic apparatus criticus all the evidence of uncial and cursire manuscripts (upward of 300), versions and early Citations from Philo and Josephus onward. As a monumental storehouse of materials "H. and P." will not be wholly superseded by the latest edition now (1913) in preparation. (6) The serviceable Cambridge "manual," edition of Swete (lst edition 1887-94, edition 3, 1901-7, 3 volumes, 8vo), is in the hands of all serious Septuagint students. The text is that of B, or (where B fails) of A, and the apparatus contains the readings of the principal uncial manuscripts. New materials discovered since the edition of H. and P., especially codex S, are employed, and greater accuracy in the presentation of the other evidence has been made possible by photography. The fact that the text here printed is but a provisional one is sometimes overlooked. Swete's edition was designed as a precursor to (7) the larger Cambridge Septuagint, of which three installments embracing the Pentateuch have (1913) appeared (The Old Testament in Greek, edition A.E. Brooke and N. McLean, Cambridge, 1911 pt. III. Numbers and Deuteronomy). The text is a reprint of Swete's except that from Ex onward a few alterations of errors in the primary manuscript have been corrected, a delicate task in which the editors have rejected a few old readings without sufficient regard to the peculiarities of Hellenistic Greek. The importance of the work lies in its apparatus, which presents the readings of all the uncials, versions and early citations, and those of a careful representative selection of the cursives. The materials of H (Law of Holiness, Lev. 17 through 26) and P (the Priestly Code) are brought up to date and presented in a more reliable and convenient form. Besides these there is (8) Lagarde's reconstruction of the Lucianic recension of the historical books, which, as stated, must be used with caution (see above)
4. Reconstruction of Original Text:
The task of reconstructing the Oldest text is still unaccomplished. Materials have accumulated, and much preliminary "spade-work" has been done, by Lagarde in particular (see his "axioms" in Swete, Introduction, 484, ff) and more recently by Nestle and Rahlfs; but the principles which the editor must follow are not yet finally determined. The extent to which "mixture" has affected the documents is the stumbling-block. Clearly no single Moabite Stone presents the oldest text. That of codex B, as in the New Testament, is on the whole the purest. In the 4 books of "Reigns" (1 Samuel through 2 Kings), e.g., it has escaped the grosser interpolations found in most manuscripts, and Rahlfs (Sept.-Studien, I, 1904) regards its text as pre-Origenic. It is, however, of unequal value and by no means an infallible guide; in Judges, e.g., its text is undoubtedly late, no earlier than the 4th century AD, according to one authority (Moore," Jgs," ICC). In relation to two of the 4th-century recensions its text is neutral, neither predominantly Lucianic nor Hexaplaric; but it has been regarded by some authorities as Hesychian. Possibly the recension made in the country which produced the Septuagint adhered more closely than others to the primitive text; some "Hesychian" features in the B text may prove to be original. Still even its purest portions contain marks of editorial revision and patent corruptions. Codex Alexandrinus presents a quite different type of text, approximating to that of the Massoretic Text. In the books of "Reigns" it is practically a Hexaplaric text without the critical signs, the additional matter being mainly derived from Aquila. Yet that it contains an ancient element is shown by the large support given to its readings by the New Testament and early Christian writers. Individual manuscripts must give place to groups. In order to reconstruct the texts current before Origen's time, it is necessary to isolate the groups containing the three 4th-century recensions, and to eliminate from the recensions thus recovered all Hexaplaric matter and such changes as appear to have been introduced by the authors of those recensions. Other groups brought to light by the larger Cambridge text have also to be taken into account. The attempt to Renetrate into the earlier stages of the history is the hardest task. The Old Latin version is here the surest guide; it has preserved readings which have disappeared from all Greek manuscripts, and affords a criterion as to the relative antiquity of the Greek variants. The evidence of early Christian and Jewish citations is also valuable. Ultimately, after elimination of all readings proved to be "recensional" or late, the decision between outstanding variants must depend on internal evidence. These variants will fall into two classes: (1) those merely affecting the Greek text, by far the larger number and presenting less difficulty; (2) those which imply a different Hebrew text. In adjudicating on the latter Lagarde's main axioms have to be borne in mind, that a free translation is to be preferred to a slavishly literal one, and a translation presupposing another Hebrew original to one based on the Massoretic Text.
VII. Number, Titles and Order of Books.
In addition to the Hebrew canonical books, the Septuagint includes all the books in the English Apocrypha except 2 Esdras (The Prayer of Manasseh only finds a place among the canticles appended in some manuscripts to the Psalms) besides a 3rd and 4th book of Maccabees. Swete further includes in his text as an appendix of Greek books on the borderland of canonicity the Ps of Sol (found in some cursives and mentioned in the list in codex A), the Greek fragments of the Book of Enoch and the ecclesiastical canticles above mentioned. Early Christian writers in quoting freely from these additional books as Scripture doubtless perpetuate a tradition inherited from the Jews of Alexandria. Most of the books being original Greek compositions were ipso facto excluded from a place in the Hebrew Canon. Greater latitude as regards canonicity prevailed at Alexandria; the Pentateuch occupied a place apart, but as regards later books no very sharp line of demarcation between "canonical" and "uncanonical" appears to have been drawn.
Palestinian Jews employed the first word or words of each book of the Pentateuch to serve as its title; Genesis e.g. was denoted "in the beginning," Exodus "(and these are the) names"; a few of the later books have similar titles. It is to the Septuagint, through the medium of the Latin VSS, that we owe the familiar descriptive titles, mostly suggested by phrases in the Greek version. In some books there are traces of rival titles in the Ptolemaic age. Exodus ("outgoing") is also called Exagoge ("leading out") by Philo and by the Hellenist Ezekiel who gave that name to his drama on the deliverance from Egypt. Philo has also alternative names for Deuteronomy--Epinomis ("after-law") borrowed from the title of a pseudo-Platonic treatise, and for Judgess "the Book of Judgments." The last title resembles the Alexandrian name for the books of Samuel and Kings, namely, the four Books of Kingdoms or rather Reigns; the name may have been given in the first place to a partial version including only the reigns of the first few monarchs. Jerome's influence in this case restored the old Hebrew names as also in Chronicles (= Hebrew "Words of Days," "Diaries"), which in the Septuagint is entitled Paraleipomena, "omissions," as being a supplement to the Books of Reigns.
3. Bipartition of Books:
Another innovation, due apparently to the Greek translators or later editors, was the breaking up of some of the long historical narratives into volumes of more manageable compass. In the Hebrew manuscripts, Samuel, Kings, Chronicles, Ezra-Nehemiah form respectively one book apiece. In the Septuagint the first three of these collections are subdivided into two volumes as in modern Bibles; an acquaintance with the other arrangement is, however, indicated in Codex B by the insertion at the end of 1 R, 3 R, 1 Chronicles of the first sentence of the succeeding book, a reminder to the reader that a continuation is to follow. Ezra-Nehemiah, the Greek version (2 Esdras) being made under the influence of Palestinian tradition, remains undivided. Originally Chronicles-Ezra-Nehemiah formed a unit, as was apparently still the case when the oldest Greek version (1 Esdras) was made.
4. Grouping and Order of Books:
In the arrangement of books there is a radical departure from Palestinian practice. There were three main unalterable divisions in the Hebrew Bible, representing three stages in the formation of the Canon: Law, Prohets "Former" i.e. Joshua, Judges, Samuel, Kings, and "Latter") and "Writings." This arrangement was known at Alexandria at the end of the 2nd century BC (Sir, prol.) but was not followed. The "Writings" were a miscellaneous collection of history and poetry with one prophetical book (Daniel). Alexandrian scholars introduced a more literary and symmetrical system, bringing together the books of each class and arranging them with some regard to the supposed chronological order of their authors. The Law, long before the Greek translation, had secured a position of supreme sanctity; this group was left undisturbed, it kept its precedence and the individual books their order (Leviticus and Numbers, however, exchange places in a few lists). The other two groups are broken up. Ruth is removed from the "Writings" and attached to Judges. Chronicles and Ezra-Nehemiah are similarly transferred to the end of the historical group. This group, from chronological considerations, is followed by the poetical and other "Writings," the Prophets coming last (so in Codex Vaticanus, etc.; in Codex Sinaiticus and Codex Alexandrinus, prophets precede poets). The internal order of the Greek Hagiographa, which includes quasi-historical (Esther, Tobit, Judith) and Wisdom books, is variable. Daniel now first finds a place among the Prophets. The 12 minor prophets usually precede the major (Codex Sinaiticus and Western authorities give the four precedence), and the order of the first half of their company is shuffled, apparently on chronological grounds, Hosea being followed by Amos, Micah, Joel, Obadiah, Jonah, Jeremiah has his train of satellites, Baruch, Lamentation (transferred from the "Writings") and Epistle of Jeremiah; Susanna and Bel and the Dragon consort with and form integral parts of Daniel. Variation in the order of books is partly attributable to the practice of writing each book on a separate papyrus roll, kept in a cylindrical case; rolls containing kindred matter would tend to be placed in the same case, but there would be no fixed order for these separate items until the copying of large groups in book-form came into vogue (Swete, Introduction, 225 f, 229 f).
VIII. Characteristics of the Version and Its Component Parts.
Notwithstanding the uncertain state of the text, some general characteristics of the version are patent. It is clear that, like the Hebrew itself, it is not a single book, but a library. It is a series of versions and Greek compositions covering well-nigh 400 years, since it includes a few productions of the 2nd century AD; the bulk of the translations, however, fall within the first half of the period (Sirach, prolegomena).
1. Grouping of Septuagint Books on Internal Evidence:
The translations may be grouped and their chronological order approximately determined from certain characteristics of their style. (1) We may inquire how a Hebrew word or phrase is rendered in different parts of the work. Diversity of renderings is not an infallible proof that different hands have been employed, since invariable uniformity in translation is difficult of attainment and indeed was not the aim of the Pentateuch translators, who seem rather to have studied variety of expression. If, however, a Hebrew word is consistently rendered by one Greek word in one portion and by another elsewhere, and if each of the two portions has other features peculiar to itself, it becomes highly probable that the two portions are the work of different schools. Among "test-words" which yield results of this kind are "servant" in "Moses the servant of the Lord," "Hosts" in "Lord of Hosts," "Philistines" (Swete, Introduction, 317 f; Thackeray, Grammar of the Old Testament, 7 ff). (2) We may compare the Greek with that of dated documents of the Ptolemaic age. The translations were written in the koine or "common" Greek, most of them in the vernacular variety of it, during a period when this new cosmopolitan language was in the making; the abundant dated papyri enable us to trace some stages in its evolution. The Petrie and Hibeh papyri of the 3rd century BC afford the closest parallels to the Greek Pentateuch. The following century witnessed a considerable development or "degeneracy" in the language, of which traces may be found in the Greek of the prophetical books. Beside the vernacular Greek was the literary language of the "Atticistic" school which persistently struggled, with indifferent success, to recover the literary flavor of the old Greek masterpieces. This style is represented in the Septuagint by most of the original Greek writings and by the paraphrases of some of the "Writings." (3) We may compare the Greek books as translations, noting in which books Iicense is allowed and which adhere strictly to the Hebrew. The general movement is in the direction of greater literalism; the later books show an increasing reverence for the letter of Scripture, resulting in the production of pedantically literal VSS; the tendency culminated in the 2nd century AD in the barbarisms of Aquila. Some of the "Writings" were freely handled, because they had not yet obtained canonical rank at the time of translation. Investigation on these lines goes to show that the order of the translation was approximately that of the Hebrew Canon. The Greek Hexateuch may be placed in the 3rd century BC, the Prophets mainly in the 2nd century BC, the "Writings" mainly in the 2nd and 1st centuries BC.
(1) The Hexateuch.
The Greek Pentateuch should undoubtedly be regarded as a unit: the Aristeas story may so far be credited. It is distinguished by a uniformly high level of the "common" vernacular style, combined with faithfulness to the Hebrew, rarely lapsing into literalism. It set the standard which later translators tried to imitate. The text was more securely established in this portion and substantial variant readings are comparatively few. The latter part of Exodus is an exception; the Hebrew had here not reached its final form in the 3rd century BC, and there is some reason for thinking that the version is not the work of the translator of the first half. In Deuteronomy a few new features in vocabulary appear (e.g. ekklesia; see Hort, Christian Ecclesia, 4 ff). The Greek version of Josephus forms a link between the Pentateuch and the later historical books. The text was not yet fixed, and variants are more abundant than in the Pentateuch. The earliest VS, probably of selections only, appears from certain common features to have been nearly coeval with that of the Law.
(2) The "Latter" Prophets.
There is little doubt that the next books to be translated were the Prophets in the narrower sense, and that Isaiah came first. The style of the Greek Isaiah has a close similarity, not wholly attributable to imitation, to that of the Pentateuch: a certain freedom of treatment connects it with the earlier translation period: it was known to the author of Wisdom (Isa 3:10 with Ottley's note). The translation shows "obvious signs of incompetence" (Swete), but the task was an exacting one. The local Egyptian coloring in the translation is interesting (R. R. Ottley, Book of Isaiah according to the Septuagint, 2 volumes, Greek text of A, translation and notes, Cambridge, 1904-6, with review in JTS, X, 299). Jeremiah, Ezekiel and the Minor Prophets were probably translated en bloc or nearly so. The Palestinian Canon had now been enlarged by a second group of Scriptures and this stimulated a desire among Alexandrian Jews to possess the entire collection of the Prophets in Greek. The undertaking seems to have been a formal and quasi-official one, not a haphazard growth. For it has been ascertained that Jeremiah and Ezekiel were divided for translation purposes into two nearly equal parts; a change in the Greek style occurs at the junctures. In Jeremiah the break occurs in chapter 29 Septuagint order); the clearest criterion of the two styles is the twofold rendering of "Thus saith the Lord." The last chapter (Jer 52) is probably a later addition in the Greek. The translator of the second half of Jer also translated the first half of Baruch (1:1-3:8); he was incompetent and his work, if our text may be relied on, affords flagrant examples of Greek words being selected to render words which he did not understand merely because of their similar sound. Ezekiel is similarly divided, but here the translator of the first half (chapters 1 through 27) undertook the difficult last quarter as well (chapters 40 through 48), the remainder being left to a second worker. An outstanding test is afforded by the renderings of the refrain, "They shall know that I am the Lord." The Greek version of "the twelve" shows no trace of a similar division; in its style it is closely akin to the first half of Ezekiel and is perhaps by the same hand (JTS, IV, 245, 398, 578). But this official version of the Prophets had probably been preceded by versions of short passages selected to be read on the festivals in the synagogues. Lectionary requirements occasioned the earliest versions of the Prophets, possibly of the Pentateuch as well. Two indications of this have been traced. There exists in four manuscripts a Greek version of the Psalm of Habakkuk (Hab 3), a chapter which has been a Jewish lesson for Pentecost from the earliest times, independent of and apparently older than the Septuagint and made for synagogue use. Similarly in Ezekiel of the Septuagint there is a section of sixteen verses (36:24-38) with a style quite distinct from that of its context. This passage was also an early Christian lesson for Pentecost, and its lectionary use was inherited from Judaism. Here the Septuagint translators seem to have incorporated the older version, whereas in Hab 3 they rejected it (JTS, XII, 191; IV, 407).
(3) Partial Version of the "Former" Prophets.
The Greek style indicates that the history of the monarchy was not all translated at once. Ulfilas is said to have omitted these books from the Gothic version as likely to inflame the military temper of his race; for another reason the Greek translators were at first content with a partial version. They omitted as unedifying the more disastrous portions, David's sin with the subsequent calamities of his reign and the later history of the divided monarchy culminating in the captivity. Probably the earliest versions embraced only (1) 1 R, (2) 2 R 1 1 through 11 1 (David's early reign), (3) 3 R 2 12 through 21 13 (Solomon and the beginning of the divided monarchy); the third book of "Reigns" opened with the accession of Solomon (as in Lucian's text), not at the point where 1 Kings opens. These earlier portions are written in a freer style than the rest of the Greek "Reigns," and the Hebrew original differed widely in places from that translated in the English Bible (JTS, VIII, 262).
(4) The "Writings."
The Hagiographa at the end of the 2nd century BC were regarded as national literature. (Sirach, prolegomena "the other books of our fathers"), but not as canonical. The translators did not scruple to treat these with great freedom, undeterred by the prohibition against alteration of Scripture (Dt 4:2; 12:32). Free paraphrases of extracts were produced, sometimes with legendary additions. A partial version of Job (one-sixth being omitted) was among the first; Aristeas, the historian of the 2nd century BC, seems to have been acquainted with it (Freudenthal, Hellenistische Studien, 1875, 136 ff). The translator was a student of the Greek poets; his version was probably produced for the general reader, not for the synagogues. Hatch's theory (Essays in Biblical Greek, 1889, 214) that his Hebrew text was shorter than ours and was expanded later is untenable; avoidance of anthropomorphisms explains some omissions, the reason for others is obscure. The first Greek narrative of the return from exile (1 Esdras) was probably a similar version of extracts only from Chronicles-Ezra-Nehemiah, grouped round a fable of non-Jewish origin, the story of the 3 youths at the court of Darius. The work is a fragment, the end being lost, and it has been contended by some critics that the version once embraced the whole of Chronicles-Ezra-Nehemiah (C. C. Torrey, Ezra Studies, Chicago, 1910). The Greek is obviously earlier than Esdras B and is of great value for the reconstruction of the Hebrew. The same translator appears from peculiarities of diction to have produced the earliest version of Dnl, treating it with similar freedom and incorporating extraneous matter (the Song of Three Children, Susanna, Bel). The maximum of interpolation is reached in Esther, where the Greek additions make up two-thirds of the story. The Greek Proverbs (probably 1st century BC) includes many maxims not in the Hebrew; some of these appear to be derived from a lost Hebrew collection, others are of purely Greek origin. This translator also knew and imitated the Greek classics; the numerous fragments of iambic and hexameter verse in the translation cannot be accidental (JTS, XIII, 46). The Psalter is the one translation in this category in which liberties have not been taken; in Ps 13 (14):3 the extracts from other parts of Psalms and from Isaiah included in the B text must be an interpolation possibly made before Paul's time (Rom 3:13 ff), or else taken from Romans. The little Ps 151 in Septuagint, described in the title as an "autograph" work of David and as "outside the number," is clearly a late Greek production, perhaps an appendix added after the version was complete.
(5) The Latest Septuagint Translations.
The latest versions included in the Septuagint are the productions of the Jewish translators of the 2nd century AD; some books may be rather earlier, the work of pioneers in the new school which advocated strict adherence to the Hebrew. The books of "Reigns" were now completed, by Theodotion, perhaps, or by one of his school; the later portions (2 R 11 2 through 3 R 2 11, David's downfall, and 3 R 22-4 R end, the downfall of the monarchy) are by one hand, as shown by peculiarities in style, e.g. "I am have with child" (2 R 11 5) = "I am with child," a use which is due to desire to distinguish the longer form of the pronoun 'anokhi ("I," also used for "I am") from the shorter 'ani. A complete version of Jdg was now probably first made. In two cases the old paraphrastic versions were replaced. Theodotion's Daniel, as above stated, superseded in the Christian church the older version A new and complete version of Chronicles-Ezra-Nehemiah was made (Esdras B), though the older version retained its place in the Greek Bible on account of the interesting legend imbedded in it; the new version is here again possibly the work of Theodotion; the numerous transliterations are characteristic of him (Torrey, Ezra Studies; theory had previously been advanced by Sir H. Howorth). In the Greek Ecclesiastes we have a specimen of Aquila's style (see McNeile's edition, Cambridge, 1904). Canticles is another late version
2. General Characteristics:
A marked feature of the whole translation is the scrupulous avoidance of anthropomorphisms and phrases derogatory to the divine transcendence. Thus Ex 4:16, "Thou shalt be to him in things pertaining to God" (Hebrew "for" or "as God"); 15:3, "The Lord is a breaker of battles" (Hebrew "a Man of war"); 24:10, "They saw the place where the God of Israel stood" (Hebrew "they saw the God of Israel"); 24:11, "Of the elect of Israel not one perished and they were seen in the place of God" (Hebrew "Upon the nobles .... He laid not His hand, and they beheld God"). The comparison of God to a rock was consistently paraphrased as idolatrous, as was sometimes the comparison to the sun from fear of sun-worship (Ps 83 (84):12, "The Lord loves mercy and truth" for Hebrew "The Lord is a sun and shield"). "The sons of God" (Gen 6:2) becomes "the angels of God." For minor liberties, e.g. slight amplifications, interpretation of difficult words, substitution of Greek for Hebrew coinage, translation of place-names, see Swete, Introduction, 323 ff. Blunders in translation are not uncommon, but the difficulties which these pioneers had to face must be remembered, especially the paleographical character of the Hebrew originals. These were written on flimsy papyrus rolls, in a script probably in a transitional stage between the archaic and the later square characters; the words were not separated, and there were no vowel-points; two of the radicals (waw and yodh) were also frequently omitted. Add to this the absence at Alexandria, for parts at least of the Scriptures, of any sound tradition as to the meaning. On the other hand the vocalization adopted by the translators, e.g. in the proper names, is of great value in the history of early Semitic pronunciation. It must further be remembered that the Semitic language most familiar to them was not Hebrew but Aramaic, and some mistakes are due to Aramaic or even Arabic colloquialisms (Swete, Introduction, 319).
IX. Salient Differences between Greek and Hebrew Texts.
Differences indicating a Hebrew original other than the Massoretic Text affect either the sequence or the subject-matter (compare Swete, Introduction, 231 ff).
The most extensive discrepancies in arrangement of materials occur in (1Ex 35 through 39, the construction of the Tabernacle and the ornaments of its ministers, (2) 3 R 4 through 11, Solomon's reign, (3) Jeremiah (last half), (4) Proverbs (end). (1) In Exodus the Septuagint gives precedence to the priests' ornaments, which in the Hebrew follow the account of the Tabernacle, and omits altogether the altar of incense. The whole section describing the execution of the instructions given in the previous chapters in almost identical words is one of the latest portions of the Pentateuch and the text had clearly not been finally fixed in the 3rd century BC; the section was perhaps absent from the oldest Greek version In Ex 20:13-15 Codex B arranges three of the commandments in the Alexandrian order (7, 8, 6), attested in Philo and in the New Testament. (2) Deliberate rearrangement has taken place in the history of Solomon, and the Septuagint unquestionably preserves the older text. The narrative of the building of the Temple, like that of the Tabernacle, contains some of the clearest examples of editorial revision in the Massoretic Text (Wellhausen, Hist of Israel, 67, 280, etc.). At the end of 3 R Septuagint places chapters 20 and 21 in their proper order; Massoretic Text reverses this, interposing the Naboth story in the connected account of the Syriac wars and justifying the change by a short preface. (3) In Jeremiah the chapter numbers differ from the middle of chapter 25 to the end of chapter 51, the historical appendix (chapter 52) concluding both texts. This is due to the different position assigned to a group of prophecies against the nations: Septuagint places them in the center, Massoretic Text at the end. The items in this group are also rearranged. The diversity in order is earlier than the Greek translation; see JTS, IV; 245. (4) The order of some groups of maxims at the end of Proverbs was not finally fixed at the time of the Greek translation; like Jeremiah's prophecies against the nations, these little groups seem to have circulated as late as the 2nd or 1st century BC as separate pamphlets. The Psalms numbers from 10 to 147 differ by one in Septuagint and Massoretic Text, owing to discrepancies in the lines of demarcation between individual psalms.
2. Subject Matter:
Excluding the end of Exodus, striking examples of divergence in the Pentateuch are few. Septuagint alone preserves Cain's words to his brother, "Let us go into the field" (Gen 4:8). The close of Moses' song appears in an expanded form in Septuagint (Dt 32:43). Similarly Hannah's song in 1 R 2 (? originally a warrior's triumph-song) has been rendered more appropriate to the occasion by the substitution in verse 8c of words about the answer to prayer, and enlarged by the insertion of a passage from Jeremiah; the changes in both songs may be connected with their early use as canticles. In Joshua the larger amount of divergence suggests that this book did not share the peculiar sanctity of the Law. But the books of "Reigns" present the widest differences and the fullest scope for the textual critic. The Septuagint here proves the existence of two independent accounts of certain events. Sometimes it incorporates both, while the Massoretic Text rejects one of them; thus Septuagint gives (3 R 2 35a ff,46a ff) a connected summary of events in Solomon's personal history; most of which appear elsewhere in a detached form, 3 R 12 24a-z is a second account of the dismemberment of the kingdom; 16:28a-h a second summary of Jehoshaphat's reign (compare 22 41 ff); 4 R 1 18a another summary of Joram's reign (compare 3 1 ff). Conversely in 1 R 17 through 18, Massoretic Text has apparently preserved two contradictory accounts of events in David's early history, while Septuagint presents a shorter and consistent narrative (Swete, Intro, 245 f). An "addition" in Septuagint of the highest interest appears in 3 R 8 53b, where a stanza is put into the mouth of Solomon at the Temple dedication, taken from "the Song-book" (probably the Book of Jashar); the Massoretic Text gives the stanza in an edited form earlier in the chapter (8 12 f); for the reconstruction of the original Hebrew see JTS, X, 439; XI, 518. The last line proves to be a title, "For the Sabbath--On Alamoth" (i.e. for sopranos), showing that the song was set to music for liturgical purposes. In Jeremiah, besides transpositions, the two texts differ widely in the way of excess and defect; the verdict of critics is mainly in favor of the priority of the Septuagint (Streane, Double Text of Jeremiah, 1896). For divergences in the "Writings" see VIII, above; for additional titles to the Psalms see Swete, Introduction, 250 f.
The most important works have been mentioned in the body of the article. See, further, the very full lists in Swete's Introduction and the bibliographies by Nestle in PRE3, III, 1-24, and XXIII, 207-10 (1913); HDB, IV, 453-54.
H. St. J. Thackeray