Article (slightly modified) by Robert A. Kraft, from The Interpreter's Dictionary of the Bible, Supplementary Volume (Abingdon 1976) 811/ii - 815 (section "A" by Emanuel Tov). [Copyright Robert A. Kraft, 28 September 1998] [[811]] B. EARLIEST GREEK VERSIONS ("Old Greek"). Questions concerning the history and significance of the Greek text have received relatively less attention than those concerning the relation of the Greek to its Semitic parent. See TEXT, OT §C; §A above.

1. Terminology, technology, and text. Several closely interrelated problems have sometimes interfered with fruitful scholarly discussion of the Greek materials.

a. The term "Septuagint." As early as the second century CE, "Septuagint" was used as an umbrella term for the Christian collections of Jewish scriptures in Greek translation, which were believed to be inspired (see SEPTUAGINT §2) and thus superior to the "later" translations (see GREEK VERSIONS, MINOR[S]). This convenient but potentially misleading use of the term still prevails (e.g., "LXX" is sometimes used for Rahlfs' 1935 edition and its eclectic text; see SEPTUAGINT §9b). Since there is no homogeneity among the various translation units of this collection (see §3d below), it is more accurate to speak of the oldest recoverable Greek form of each section/book (OG = "Old Greek"), which in the Pentateuch is the LXX proper (see ARISTEAS). Discussion must proceed section by section, or book by book, according to identifiable translation units.

b. Ancient technology and modern OG study. The possibility of including the entire Jewish Scriptures in a single volume did not exist until. the development of the large-scale CODEX format in the fourth century. Prior to that, small-scale codices (second- third centuries) and rolls were used, which could hold only one or two major writings (e.g., Numbers-Deuteronomy). Even during the Middle Ages the practice of including the entire Greek OT under one cover was not widespread. Only about one percent of the approximately twelve hundred known MSS of portions of OG contain a majority of the OT writings.

c. Canon, order of contents. Identification of "canonical" writings (see CANON OF THE OT) must have depended less on comprehensive OT MSS than on traditional lists of canonical literature. (Perhaps repositories for storing the various individual scripture rolls together were used in early times.) The relative imprecision of the early lists, the varying order in the later lists, and the fluidity with which the term "scripture" is applied by certain early authors may to some extent be due to the developing technological situation.

d. Textual complications. Technical considerations must also have contributed to the complexity of the text itself. If alternative text types existed in a given locale, it would have been extremely difficult to preserve a homogeneous text type in a section of material requiring multiple scrolls or small codices. The process of copying [[812]] the smaller units in order to construct larger codices would tend to encourage the intermixing of materials with diverse textual histories. Relatively mechanical textual confusion of this sort may be reflected in the widespread appearance of "Theodotionic" Daniel texts (see DANIEL[S]), and doubtless helps explain other textual problems in presumably homogeneous OG sections (see §3b-c below, especially Samuel-Kings), including the fact that certain Greek MSS even change their textual affinities within a given book. In addition, all of the usual problems relating to the transmission of texts in antiquity apply as well to the history of OG materials (see TEXT, OT §D; TEXT, NT §A), along with the special problems relating to translated literature (see §3e below) and to Hexaplaric influence. See VERSIONS, ANCIENT §2a.

2. Working backward toward the earliest accessible stages. Including commentary MSS, lectionaries, and catenae, over two thousand direct witnesses to the Greek text of Jewish Scriptures are extant, dating from the sixteenth century or earlier. About ninety per cent of the actual OG MSS contain only a single book or a discrete, relatively small section -- e.g., the text of Psalms or of Psalms plus Odes is found by itself in more than 750 MSS! By means of TEXTUAL CRITICISM, scholars attempt to reconstruct as accurately as possible the form of the text as it first appeared. One of the first steps in that direction involves a careful comparison of the preserved witnesses, including versions and quotations as well as the aforementioned Greek materials.

a. Grouping the witnesses by families. The witnesses frequently show textual relationships among themselves that encourage the construction of textual "family trees" which theoretically provide a route back to MS ancestors no longer extant. Sometimes the lost ancestor of a textual family might even be a conscious revision of the text (e.g., "Lucianic," "Hexaplaric"). Unfortunately, the incomplete nature of the preserved evidence (missing links) and the presence of cross- fertilization between MSS and families (e.g., through conscious "correction" and harmonization) seriously complicates this endeavor. See SEPTUAGINT §4.

b. Identifying lost archetypes. Prior to the tenth century CE MSS seldom provide explicit information regarding their own transcriber, location, date, or the MS from which they were copied. Occasionally it is possible to assign a given type of handwriting or format or system of notation to a precise locale and/or date, but the earlier the materials the more difficult such identification becomes. Sometimes the text reflected in one of the "daughter versions" made from the Greek or in quotations made by an identifiable author may provide clues to the origin of a MS or to a lost archetype -- e.g., a text displaying strong affinities with the Coptic-Sahidic version might have a middle Egyptian provenance, or affinities with quotations in Antioch- related church Fathers often receive the label "Lucianic." Other clues to a family's background are sometimes provided by the use of Origen's Hexaplaric signs within the text (see SEPTUAGINT §4a) , or by the appearance of a variety of marginal notations attempting to [[812 ii]] identify particular readings (e.g., "from the Samaritikon," "from Aquila," etc.).

c. Moving behind the archetypes, attested recensions, and unaligned witnesses. Unfortunately, very little concrete evidence is available for tracing the history and development of Greek scriptural texts in the centuries prior to Origen (ca. 250 CE) and the so-called Hesychian and Lucianic recensions (ca. 300; see SEPTUAGINT §4). While it is sometimes possible to detect the influence of Origen's Hexapla on later MSS, identification of the supposed "Lucianic" recension (and its background) is still widely debated (see GREEK VERSIONS MINOR[S] §§1c, 4), and there has been little success in isolating a specific "Hesychian" recension. Furthermore, the few preserved MSS and fragments that antedate Origen seem to defy classification by post-Origen criteria. With the exception of materials used to posit a "ProtoLucianic" textual stream going back at least to Josephus, the earliest witnesses have sometimes been termed "prerecensional" or "mixed texts" (i.e., not aligned with established recensions or families). But such terminology should not be allowed to obscure the probability that conscious revisional activity in Greek texts took place almost from the very beginning, both by way of refinement of style and vocabulary and through attempts to keep the Greek text in close conformity to the (changing) Hebrew/Aramaic textual situation (see TEXT, HEBREW, HISTORY OF[S] §1). Indeed, the traditions about later translation activity (see GREEK VERSIONS, MINOR[S]) support such suspicions. Furthermore, unintentional changes that also create family relationships among MSS were inevitable at every stage of copying. Thus in principle, the pre-Origen situation in OG texts should not be expected to differ from the postOrigen situation; there certainly were MS families and probably also conscious revisional activities of various sorts.

d. The earliest preserved OG materials. The growing body of relevant materials that antedate the Second Jewish Revolt (135 CE) permits a glimpse of early OG textual developments. Actual Greek fragments include: Approximate Date Siglum Identification 2nd BCE 801 4 QLXX Lev\a 2nd/1st BCE 957 P.Rylands 458 Deut 805 7 QLXX Exod 804 7 QLXX Epist.Jer 1st BCE 942 P.Fouad 266 Gen 848 P.Fouad 266 Deut #1 1st BCE/1st CE 802 4 QLXX Lev\b 803 4 QLXX Num 1st CE 847 P.Fouad 266 Deut #2 1st/2nd CE 814 P.Yale 1 Gen (codex) Early 2nd CE 963 Chester Beatty Num-Deut (codex) Extensive quotations and allusions in Greek are also preserved in Philo (ca. 30 CE), Paul (ca. 50), Josephus (ca. 80), 1 Clement (ca. 95), and a number of other Christian texts from the early [[813]] period. Determination of the exact wording of a quotation often must remain problematical, however, since the original text of the document in which it occurs must itself be reconstructed by careful textcritical methods. Finally, some later non-Greek versions and quotations (especially OL) can also be used, since they are clearly based upon Greek texts of an early date (see LATIN VERSIONS [SI). The general conclusion to be drawn is that considerable textual variety is present at every stage for which sufficient evidence is preserved. Any hopes to reconstruct with precision the original form of each OG translation by a relatively mechanical use of the preserved witnesses seems doomed to falter on the stormy sea of textual variations.

3. Reconstructing OG origins. Although ultimately the available evidence is inadequate for the task, there is some value in attempting to indicate fruitful avenues of approach to the problems of OG origins.

a. Tradition, theory, and data. First, it is hoped that current discussion can move beyond the heated dispute carried on in the names of Lagarde (one original OG existed from which the textual families, recensions, etc. developed) and Kahle (from independent competing Greek translations, a Christian Greek OT was formed by various selective processes; see SEPTUAGINT §5). The generalized polemic positions are too simple to do justice to the evidence; what may be true for the OG Pentateuch is not necessarily true for OG Judges or Tobit and vice versa. Secondly, the relation between the ARISTEAS legend of Pentateuchal LXX origins and the question of the origin of the other OG sections calls for careful reexamination. The frequent assumption that virtually every OG section was translated in Alexandria, although not at the same early date as the Pentateuch, requires careful evaluation in light of the full range of alternative possibilities (including Palestine itself; see Barthe/lemy). While more satisfactory criteria are being developed for identifying place of origin and date of Greek translations from Semitic originals, Alexandria should not be allowed to dominate by default.

b. Sections of Scripture with a relatively narrow textual base. The extant witnesses to several OG sections attest a relatively small range of textual variation, which suggests that a single OG translation may underlie each (unless evidence of ancient alternative translations has disappeared, as almost happened in Daniel) : the Pentateuch (with some special problems in Exodus and Deuteronomy), Joshua, Ruth, Isaiah, Job (later supplemented from Theodotion), Psalms, Proverbs, Koheleth (probably in Aquila's version), Chronicles, 1 Maccabees. A few other books have a relatively narrow base because they were probably written originally in Greek: 2-4 Maccabees, Wisdom of Solomon.

c. Sections with a problematical textual base. The extant witnesses to other portions of the Greek OT show such a wide range of divergence that it is difficult to determine whether a single [[813 ii]] original OG translation (which later underwent corruption and revision) lies behind them, or whether two or more relatively independent Greek translations of the same section should be posited. Especially noteworthy are Sirach (complex Hebrew and Greek textual situations), Daniel ("Theodotion" versus non-Theodotionic text), Tobit (text preserved by most MSS versus that of Sinaitictis and OL; perhaps a third version in a few witnesses), Judges (majority text versus that of Vaticanus and a few allies), Ode of Habakkuk (Hab. 3; majority text versus "Barberini" version). Samuel-Kings also deserves mention here, with at least two different translation techniques (haige and nonkaige) interwoven in the preserved witnesses (see SAMUEL, I AND II[S] §2c). The probable translation block encompassing Jeremiah-Baruch + Ezekiel + Minor Prophets also exhibits some textual problems requiring thorough analysis -- possibly extensive recensional activity on a single OG translation produced the situation. Similar textual problems occur also in Ezra-Nehemiah (see OG "I Esdras"), Esther, and Judith.

d. Classifying translation types and techniques. More precise features of the respective OG versions in each translation section are sometimes ascertainable with a high degree of probability, if the investigator moves beyond mere eclectic use of preserved textual witnesses to cautious, controlled conjectural restorations. If it can be assumed that (1) the OG translators) usually followed specific patterns, and (2) used a Hebrew/Aramaic text roughly similar to (or reconstructable from) available witnesses; and, if significant evidence of such patterns is still embedded in the extant witnesses to the Greek, it is often possible to identify other passages in which the pattern is expected but has now disappeared. The more self-consciously consistent the translator, the easier the task of conjectural reconstruction. Conversely, inconsistent or extremely free translation efforts are less susceptible to this sort of controlled conjecture. The science of identifying translation patterns in this literature is still in its infancy (see Barthe/lemy, Tov, Wevers, Martin; see also GREEK VERSIONS, MINOR[SI). The following tentative classifications of types of translations, according to the translator's relative skills in representing the Semitic parent language and producing an acceptable Greek version, may provide some impression of the relationship between the identifiable translation blocks in Jewish Greek Scriptures:

1. Literal translation reflecting closely the Semitic text:

a. Relatively more focus on parent text:

(1) mechanical: Aquila, OG Koheleth (=? Aquila) ;
(2) relatively wooden/stilted: "Kaige," Vaticanus text of Judges.

b. Relatively more focus on producing acceptable Greek:

(1) relatively stilted: Psalms, Jeremiah-Baruch+Ezekiel+Minor Prophets, text A of Judges(-Ruth), Chronicles, Canticles (?), Ezra-Nehemiah (?), Joshua (?), "Theodotionic" Daniel (?); [[814]]
(2) more idiomatic Greek: the Pentateuch, Symmachus, Esther (part), "Theodotionic" Job and Daniel texts (?), 1 Maccabees (?).

2. Free translation less concerned with the parent language/text:

a. nonparaphrastic free renderings: Isaiah, non-"Theodotionic" Daniel, Barberini Ode of Habakktik, 1 Esdras (?) ;

b. free paraphrase: Job, Proverbs.

e. Synthesis of results. It is not always possible to make absolute distinctions between OG and "later translators" (revisors?), or to distinguish with any confidence between an independent translation (prepared de novo from an available Hebrew/Aramaic base text), a recension (conscious large-scale revision of an existing Greek translation), and sporadic revision/interpolation/glossing (conscious spot "correction" of existing Greek texts). There is abundant evidence of various Conscious editorial efforts prior to Origen's Hexapla -- sometimes bringing a Greek text closer to a known Semitic base, sometimes incorporating (or substituting) material from other known Greek renderings or smoothing apparent problems. Special textcritical criteria are necessary for dealing with translation Greek, especially since the Hebrew/Aramaic materials were also undergoing textual modification over a long period in which Greek translations already were available. The old textcritical "rule" to prefer a "freer" reading to a more literal one is only of limited use. With relatively free OG renderings, the expected revisional tendency would often be toward literalistic correction rather than away from it. But with originally hyperliteral versions, the tendency could be just the opposite. The more general question about whether the development went from relatively free translations toward more literal ones as concepts of inspired text (and canon) became selfconscious is oversimple and not always applicable. Until it can be determined more precisely who made which translations for what purposes (e.g., private use versus public liturgy; see Brock), speculation about unilinear development in translation types and techniques must remain highly tentative. Even with our present state of knowledge, both Aquila (hyperliteral) and Symmachus (much more idiomatic) stand relatively near each other chronologically. The situation might have been similar for the earliest translations.

4. Influence of OG on later developments. In a variety of ways, OG materials have influenced subsequent language and literature, both in Greek and in translation.

a. On Greek vocabulary and syntax. The OG translations not only introduced some new words into Greek but also often modified the meanings of extant vocabulary. The extent to which they reflect spoken (Jewish) Greek dialects remains an open question. Sometimes it is impossible to determine whether a preserved Greek writing (or section thereof) was composed in Greek strongly influenced (consciously or unconsciously) by the available translation idiom, or was itself translated from Semitic. Criteria for judging this sort of problem are being developed (see R. A. Martin, [[814 ii]] Syntactical Evidence of Semitic Sources in Greek Documents [1974]), and should receive a strong assist from contemporary, computer technology.

b. Quotations/allusions drawn from OG. Knowledge of Jewish Scriptures and scriptural tradition in the Greek-speaking world usually was mediated through OG and related translations. Quotations/allusions not only provide a potentially valuable indication of textual developments (see §2b above), but also often indicate how particular passages were interpreted and applied at various times and places (see EXErESIS[S]). The "Centre d' Analyse et de Documentation Patristiques" at Strasbourg is building a central file of patristic biblical quotations: see J. Allenbach et al., Biblia Patristica: Index des citations et allusions bibliques dans la litte/rature patristique, 1 (1975).

Bibliography. A period of renewed activity in the study of OG and related materials began in 1968 with the founding of the International Organization for Septuagint and Cognate Studies (IOSCS). Its annual Bulletin chronicles current work, and detailed listings of literature published through 1969 appear in A Classified Bibliography of the LXX, compiled by S. P. Brock, C. T. Fritsch, and S. Jellicoe (1973).

Textcritical materials, editions, studies: The main center of textcritical activity on OG remains the Go%ttingen "Septuaginta- Unternehmens"; especially noteworthy among related projects is the continuing work on OL biblical material by the Benton Institute. The latest Go%ttingen eds. are IX.2-3=Maccabacorum liber If (1959) and III (1960), ed. R. Hanhart; XII.1-2= Sapientia Salomonis (1962) and Sapientia Iesu Filii Sirach (1965), ed. J. Ziegler; VIII.3=Esther (1966), and VIII.I=Esdrae liber I (1974), ed. R. Hanhart; and I=Genesis (1974), ed. J. Wevers. Corrected reprints of several earlier volumes have also been issued. New MSS and fragments (esp. papyri) continue to appear -- see esp. K. Treu, "Christliche Papyri I-IV," Archiv fu%r Papyrusforschung, 19-22 (1969-73). Important recent studies of textcritical importance include: D. Barth6lemy, Les Devanciers d'Aquila (1963); Proceedings of IOSCS Symposium on Samuel-Kings, ed. R. A. Kraft, LXX and Cognate Studies, 11 (1972); P. Walters [Katz], The Text of the LXX: Its Corruptions and Their Emendation, ed. D. W. Gooding (1973); E. Tov, The LXX Translation of Jeremiah and Baruch (1975); J. W. Wevers, Text History of the Greek Genesis (1974); D. W. Gooding, Relics of Ancient Exegesis: A Study of the Miscellanies in 3 Reigns 2 (1975); R. Klein, Textual Criticism of the OT (1974). A new ed. of the Letters of Aristeas, by A. Pelletier, has appeared as Sources Chrdtiennes, 89 (1962).

Lexicography, tools, and studies: J. Reider's long awaited An Index to Aquila is finally available (ed. N. Turner, 1966), and various efforts to enhance the usefulness of the Hatch-Redpath concordance and its Hebrew-Greek index have appeared: e.g., X. Jacques, List of LXX Words Sharing Common Elements: Supplenient to Concordance or Dictionary (1972), E. C. Dos Santos, An Expanded Hebrew Index (1973). Plans for a special "LXX Lexicon" are under way through IOSCS auspices: see R. A. Kraft, ed., Septuagintal Lexicography (1972), for background work. Specific investigations of vocabulary include S. Daniel, Recherches sur le vocabulaire du culte dans la Septante (1966); D. Hill, Greek Words and Hebrew Meanings (1967).

General works, collected essays, etc.: Swete's invaluable Introduction to the OT in Greek (1900 [2nd ed., 19021) has been updated by S. Jellicoe, The LXX and Modern Studies (1968). On ancient translation[[815]] activities, see S. P. Brock, "The Phenomenon of the LXX," OTS, XVII (1972), 11-36. Pertinent essay collections include J. Ziegler, Sylloge: Gesammelte Aufsdtze Zur LXX (1971); J. Schreiner, ed., Wort, Lied und Gottesspruch (1972); S. Jellicoe, ed., Studies in the LXX: . . . Selected Essays (1974). R. A. KRAFT (University of Pennsylvania) //end//