The Septuagint, also known as the LXX, is a Greek translation of Jewish scripture that was prepared in Alexandria and Palestine. The Hebrew source of the LXX differed from the other textual witnesses (the Masoretic Text [MT] and many of the Qumran texts), and this accounts for its great significance in biblical studies. The LXX is the main ancient witness that occasionally reflects compositional stages of books of the Hebrew Bible that are different from the MT and from other sources. Moreover, the LXX is important as a reflection of early biblical exegesis, Greek-speaking Judaism, and the Greek language. Finally, the LXX is also of major importance for understanding early Christianity since much of the vocabulary and some religious ideas of the New Testament are based on it.

Name and Nature.

The name of the LXX reflects the tradition that seventy-two elders translated the Torah into Greek (thus Sof. 1.7 and parallels, and the Letter of Aristeas, a late first-millennium B.C.E. Jewish composition describing the origin of the LXX). In the first centuries C.E. this tradition was expanded to include all the translated biblical books, and finally it encompassed all the Jewish scriptures translated into Greek as well as several works originally composed in Greek.

The translation of the Torah may reflect an official translation, as narrated in the Letter of Aristeas and Jewish sources, but it was not created by seventy-two individuals as narrated in these sources. The books of the Torah were probably rendered by five different translators. The subsequent biblical books were similarly translated by different individuals, although some of them translated more than one book.

The collective name Septuagint(a) now denotes both the original translation of Hebrew and Aramaic scriptures into Greek and the collection of sacred Greek writings in their present, canonical form. Neither usage is completely accurate, since the collection contains original translations, late revisions (recensions) of those translations, and compositions originally written in Greek. For this reason, scholars usually use the “Septuagint” for the collection of sacred Greek writings and the Old Greek (OG) for the reconstructed original translation. The name is often put in quotation marks (“LXX”) when it is necessary to stress the diverse nature of the books included in the collection.


The “LXX” contains two types of books:

(a)The Greek translation of the twenty-four canonical Hebrew-Aramaic books. The translation of these books contributes significantly to biblical studies, in particular to the textual transmission and exegesis of the Bible.

(b)Books not included in the collection of the Hebrew scripture and subsequently named Apocrypha (the “hidden” books) in Greek and sefarim ḥiṣoniyyim (“the outside books”) in Hebrew. These books, considered deuterocanonical in the Catholic Church, consist of two groups:

(1) a Greek translation of books, whose Hebrew or Aramaic source has been lost or preserved only in part (e.g., Sirach and Baruch 1:1–3:8); and

(2) a few works originally composed in Greek (e.g., the Wisdom of Solomon).

Sequence of the Books.

The twenty-four books of the Hebrew Bible included in the LXX together with the so-called Apocrypha are arranged in a sequence different from that of the traditional Hebrew canon. Whereas its books are arranged in three sections (Torah, Prophets, Writings), probably reflecting their acceptance at different stages as authoritative, the Greek arrangement reflects their literary genre. The Greek canon may be conceived of as having at least three and as many as five divisions:

  • I (1) Legal and (2) historical books;
  • II (3) Poetical and (4) sapiential books;
  • III (5) Prophetic books.

The Greek tradition that places the Prophets at the end of the Greek canon is found in Codex B and many additional sources. According to Swete (1914, p. 219) this tradition reflects “the great majority of authorities both Eastern and Western.” On the other hand, the three-section division of codices A, S, and other sources reflects that of the traditional Hebrew canon. The sequence of the majority Greek tradition is usually presented as reflecting a Christian tradition. In this sequence, the Prophets, who according to Christian belief foretell the coming of Jesus, are placed immediately before the New Testament. The minority tradition of A, S, and other sources possibly reflects either a late approximation to the Hebrew tradition or the original Greek arrangement, since S is the earliest manuscript of the complete Greek scripture.

Within each section the Greek books are arranged in a sequence different from that of the twenty-four books of the Hebrew Bible. In reviewing earlier scholarship, Lust (2003) has shown that there is no evidence for the assumption that the LXX reflects an Alexandrian as opposed to a Palestinian canon. The differences between the Hebrew and Greek canons pertain to the following areas:

  • Major/Minor Prophets. Since the Minor Prophets were joined as one unit (“the Book of the Twelve”), there was no option to place some of them before the Major Prophets and others after them. The Hebrew tradition places the Major Prophets first because of their prominence, in defiance of chronological considerations that would have given preference to Amos as the earliest of the writing prophets (Amos prophesied shortly after 760 B.C.E.). This tradition is also reflected in Codex S of the LXX and some, chiefly Western, church fathers. However, in codices A, B, V and the majority of the patristic lists the Minor Prophets precede the Major Prophets.
  • • Internal sequence of the Minor Prophets. The main organizing principle behind the internal sequence of the Greek Minor Prophets probably was chronological: Hosea, Amos, Micah, Joel, Obadiah, Jonah, Nahum, Habakkuk, Zephaniah, Haggai, Zachariah, and Malachi. The last six books have the same position in the Hebrew tradition, while the first six occur in this sequence in MT: Hosea, Joel, Amos, Obadiah, Jonah, Micah. If the organizing principle of the common Hebrew-Greek sequence is chronological, the sequence of the LXX possibly reflects internal amendments in the first six books so as to arrange them by size.
  • Judges-Ruth. In the arrangement of the LXX, Ruth follows Judges. This sequence is secondary since the Hebrew and Greek canons contain the same Deuteronomistic block of historical books from Deuteronomy until the end of 2 Kings, in which the different literary genre of Ruth is out of place.

Additionally, the individual apocryphal books are integrated into the various sections of the LXX in accordance with their literary genre.

Original Form, Jewish Background, Place, and Date.

In the wake of de Lagarde's 1863 work, most scholars are now of the opinion that all LXX manuscripts derive from a single translation (for its reconstruction, see below) that was repeatedly revised to the proto-MT. The alternative model, suggested by Kahle in 1915, assumes multiple translations without specifying the relation between these translations.

Jewish Background.

The Jewish origin of the LXX, described in the Letter of Aristeas, rabbinic literature, and additional sources, is reflected in its terminology and exegesis. Several Hebrew words have been preserved in the LXX in their Hebrew or Aramaic form (at the time of the translation, Aramaic was more commonly spoken by Jews than Hebrew), such as sabbata (Sabbath) and pascha (Heb. pesaḥ, Passover). The Greek Torah also reflects neologisms in the Greek language that were meant to represent special Jewish customs or terms, such as the names of the festivals and Jewish concepts for which no words existed in the Greek language. Thus, holokautoma (“whole-burnt offering”) was probably coined to reflect the special meaning of the ‘olah offering.

Jewish exegesis is visible wherever a special interpretation of the LXX is paralleled by rabbinic literature. Such exegesis reveals the Palestinian background of the Pentateuch translators. The Jewish background of the Greek translation of the Torah is well established, while that of the post-Pentateuchal books is not. Nevertheless, there is little doubt that Jews translated these books in the third and second centuries B.C.E. There probably were no gentiles in Egypt, Palestine, or elsewhere who would have had the skills to make such a transcultural translation, or an incentive to do so.


Although often described as the “Alexandrian version of the Bible,” an Alexandrian origin of the LXX is likely only for the Torah and some additional books. There is now a growing understanding that several books were produced in Palestine. These are, in order of probability: Esther, Qoheleth (Aquila or kaige-Theodotion; see “Kaige-Theodotion” below), sections of the “LXX” of Samuel-Kings, Canticles, Lamentations, and Ruth (all: kaige-Theodotion).


According to the Letter of Aristeas, the Torah was translated in Egypt at the beginning of the third century B.C.E. This dating is supported by the early date of several papyrus and leather fragments of the Greek translation of the Torah from Qumran and Egypt, some of which date from the middle of the second century until the first century B.C.E (see “Direct Witnesses” below). The remaining biblical books were translated at different times. Some evidence for their dates is external, for example, quotations from the LXX in ancient sources, and some internal, for example, reflections of historical situations or events found in the translation.

Most of the post-Pentateuchal books use the vocabulary of the Torah, and the translations of the Latter Prophets, Psalms, and Sirach also quote from that translation. The Prophets and several of the Hagiographa—the Greek versions of which were known to the grandson of Ben Sira at the end of the second century B.C.E.—were likely translated at the beginning of that century or possibly earlier. It is difficult to date individual books because there is little explicit evidence: Chronicles is quoted by Eupolemos in the middle of the second century B.C.E., and Job by Pseudo-Aristeas in the beginning of the first century B.C.E. The Isaiah translation alludes to historical situations and events that point to the years 170–150 B.C.E. The “LXX” also contains revisions of original translations (see further below). These revisions were made from the first century B.C.E. until the beginning of the second century C.E., so we can infer that about four hundred years separate the translation of the Torah from the latest “LXX” translation.


The LXX is evidenced in direct witnesses, such as early papyrus fragments and manuscripts, and indirect witnesses, such as the translations made from the LXX (for an updated description of all the direct witnesses, see Rahlfs–Fraenkel 2004).

Direct Witnesses.

Numerous sources contain the LXX, either completely or in part, varying in date from the second century B.C.E. until the late Middle Ages:

  • (1) Early papyrus and leather texts, including both scrolls and codices dating from the second century B.C.E. onward, were discovered in the Judean desert and Egypt. The early date of these fragments, mainly of the Torah, helps us gain insights into the earliest periods since the Hexapla either replaced older text traditions or contaminated manuscripts from the third century C.E. onward (see “Hexapla” below). 4QLXXLeva and 4QpapLXXLevb (both first century B.C.E.) probably reflect the OG better than the manuscript tradition contained in the later uncial manuscripts. 4QLXXLeva contains a slightly freer translation than that found in the uncial manuscripts. This scroll presumably reflects the OG, revised toward the MT in the uncials. In 4QpapLXXLevb Lev 3:12, 4:27, IAW probably reflects the original, pre-Christian rendering of the Tetragrammaton preceding kyrios of the LXX. The translation vocabulary of both Qumran scrolls was not yet standardized as it was in the later uncials. The Chester Beatty/Scheide collection contains large sections of most biblical books; the papyri containing Daniel (numbered 967–968) are particularly important because they are the sole witnesses (except for the late Hexaplaric manuscripts) of the LXX of this book.
  • (2) Uncial or majuscule manuscripts from the fourth to the tenth century C.E., written with “capital” letters, are the main source for our knowledge of the LXX. Codices B, A, and S contain all or almost all the books of the LXX. The fourth-century Codex B (Cod. Vat. Gr. 1209, or “Vaticanus”) is the best complete manuscript of the LXX, serving as the base text for several editions. It is relatively uninfluenced by the revisions of the LXX, but in Isaiah it is Hexaplaric and in Judges it contains another type of revision. Codex A (B.M. Royal MS 1 D V–VIII, designated as “Alexandrinus”) dates from the fifth century. It often adapted the text to similar verses and it also added harmonizing details. Several books of this codex represent the Hexaplaric tradition faithfully. Codex S, also named א (B.M. Add. 43725, or “Sinaiticus”), dates from the middle of the fourth century and is the oldest manuscript that combined the LXX and NT. It usually agrees with Codex B, when the two reflect the OG translation, but it is also influenced by the later revisions of the LXX.
  • (3) Minuscule manuscripts, written with lowercase letters, from the ninth to the sixteenth centuries, are recorded in the Göttingen and Cambridge editions, while others are known from the edition of Holmes-Parsons 1798–1827. Despite their late date, minuscules often preserve ancient traditions. For example, the Lucianic tradition (LXXLuc) is known mainly from the b, o, c2, and e2 minuscules in the Cambridge editions (see “Editions” below).

Indirect Witnesses.

In the first centuries C.E. the LXX was the official source of the Bible for the Christian church. Because of this, many translations of it were needed for the churches in the East and West: Armenian, Coptic (Sahidic, Bohairic, Akhmimic), Georgian, Old Slavonic, Ethiopic, Gothic, and Arabic. Several of these translations are important for our knowledge of the LXX and its revisions. The Vetus Latina (VL or “the Old Latin” translation) is of particular interest. It preserved many important Greek readings, and was sometimes the only witness, often in conjunction with the LXXLuc and (for Samuel) 4QSama. VL, the first Christian translation of the LXX, derived directly from the Greek. However, certain “Hebraizing” elements in it may have come directly from a Hebrew source. The textual evidence for VL is very complex, and its various manuscripts may evidence several Greek sources of different nature. If all these complications are taken into consideration, one may extract valuable data from the VL not only for the reconstruction of the OG, but also for its Hebrew Vorlage, especially in Samuel and Kings.


Almost all the uncial manuscripts of the LXX have been published in separate unaltered (“diplomatic”) unaltered editions. The two major editions are Holmes–Parsons 1798–1827 and Brooke–McLean–Thackeray, 1906–1940, (“the Cambridge Septuagint”). The Göttingen Septuagint (1926–), containing the reconstructed (“eclectic”) “original” text and a critical apparatus of variants, is the most precise and thorough critical edition of the LXX. An abridged version of the work according to the Göttingen system was published by Rahlfs 1935 and Rahlfs–Hanhart 2006.

Auxiliary Tools.

The bilingual concordance of Hatch–Redpath 1897–1906 lists the “formal” Hebrew and Aramaic equivalents for most of the LXX words in the canonical books, Sirach, and 1 Esdras—for the other Apocrypha the Greek words are listed without equivalents. Muraoka 2010 provides Greek–Hebrew/Aramaic reverse indexes. Scholarly translations of the LXX are contained in A New English Translation of the Septuagint (NETS) and Septuaginta Deutsch. The Computer Assisted Tools for Septuagint/Scriptural Study (CATSS) MT/LXX database allows for advanced searches.


Linguistic exegesis is an essential part of translation. This exegesis consists of linguistic identifications that identify all forms in the source language and its syntax, semantic exegesis of all the words in the source language, and determining the equivalents of the source language in the target language. All translations reflect at least these three levels of linguistic exegesis. Nevertheless, only a few units in the LXX (such as 2 Kings = 4 Kingdoms in the LXX) are limited to such exegesis.

Most LXX units reflect exegesis beyond the linguistic exegesis described above. Some exegetical elements form a necessary part of the translation process, while others personalize the text to such an extent that its plain meaning is completely obscured. For example, some translators felt free to include allusions to other biblical verses or to insert their own reflections into the translation. While these translators seem to have strayed far from the simple meaning of the verses, they were, nevertheless, reflecting what they considered to be the message of the text.

As a rule, translations also reflect various forms of contextual exegesis and exegesis that is theologically or historically motivated. Theological exegesis—often expressed through small or large changes in words and verses, expansions or omissions of “offensive” ideas, and theologically motivated translation equivalents—may relate to the description of God and God's acts, the Messiah, Zion, and the exile, as well as concepts such as repentance. The translation also contains other forms of exegesis, such as historical, geographical, and cultural exegesis.

Translation Technique.

Each language has its own internal logic, which complicates translation when categories of the source language do not exist in the target language. In such cases, the translators exhibited their inventiveness within the framework of what is usually called translation technique. Translators had to devise ways of representing the Hebrew verbal system, construct formation, conjunctions, and particles as well as formulations unique to Hebrew, such as ʾăšer…bô (literally: “which/that…upon which”). As a result of the LXX's heterogeneous character, the translation techniques visible in the various books are extremely diverse, especially in the post-Pentateuchal books.

The study of the translation technique of the LXX was developed in the twentieth century with the particular aim of better assessing the text-critical value of that version. Scholars make judgments on individual readings and complete translation units based on manual and computer-assisted studies of the translation units, especially if the translation was made faithfully.

Retroversion of the Parent Text.

Scholars are divided concerning the feasibility of reconstructing the Hebrew text used by the ancient translators (usually named the Vorlage or “parent text”). Some think that it is possible to reconstruct words or sentences, while others emphasize the difficulties involved. Scholars such as Tov 1997 have formulated general rules for reconstruction, but they are of limited value because even the use of one particular rule or another is subjective. The support of the Samaritan Pentateuch (SP) and the Qumran scrolls for specific reconstructions from the LXX gives further confidence to the reconstruction procedure as a whole (see “Greek and Hebrew Scrolls from the Judean Desert” below).

Reconstructions of individual elements in the LXX are recorded in monographs, commentaries, and critical editions of the Bible; Polak–Marquis 2002; and the CATSS database. In addition, Tov 1997, pp. 126–127 lists large-scale retroversions of complete chapters (e.g., Jeremiah 27) and books (e.g., Ezekiel, Esther).

Hebrew Source of the LXX.

Among the ancient translations, the LXX holds pride of place for textual critics since it reflects a greater number of variants than all the other translations put together. Textual differences between the LXX and the other witnesses of the Hebrew Bible, reconstructed as variants in the underlying Hebrew Vorlagen, are extant in all the books. They are especially remarkable in Samuel, since its Masoretic Text is often corrupt.

The analysis of the character of the LXX in the various books indicates that they share only a limited number of features; it is therefore not appropriate to speak about a Septuagintal text-type, Septuagintal features, or the like. The main element shared by the Hebrew Vorlagen of the books of the LXX is that they were chosen to be rendered into Greek. The only textual feature recognizable in the LXX as a whole is a large number of small harmonizing pluses in the Torah/Pentateuch similar to those of the SP and its forerunners found at Qumran (the so-called pre-Samaritan scrolls).

Together with smaller textual variants created in the course of the textual transmission of the biblical books, the LXX reflects a sizeable number of editorial differences created in the course of the literary growth of the books. In the following books such variations have been recognized: Genesis (chronology; sequence differences in ch. 31); Exodus (internal sequence and content of chs. 35–40); Numbers (small details); Joshua (significant transpositions; pluses, e.g. after the end of the book; minuses, e.g., 20:4–6; theological differences); 1–2 Samuel (significant editorial differences between the MT, the LXX, and 4QSama, especially in the Song of Hannah and 1 Samuel 16–18); 1–2 Kings (pluses, minuses, and transpositions; different chronology—1 Kings LXX usually later than the MT); Jeremiah (different sequence and a much shorter text); Ezekiel (slightly shorter version); Psalms (a few editorial differences; inclusion of Psalm 151); Proverbs (pluses and minuses of verses, sequence differences, especially in chapters 24–31); Esther, EstherA-Text, and Daniel (rewritten compositions involving major additions and omissions); Ezra-Nehemiah and Chronicles (small differences).

The editorial differences between MT and the LXX are much greater than those between MT and the Qumran scrolls. The special character of the Vorlage of the LXX seems to be related either to one of two factors or to a combination thereof: (1) the idiosyncratic Hebrew scrolls used for the Greek translation were not used by the circles that fostered the MT; and/or (2) the relatively early date of the translation enterprise (275–150 B.C.E.), involving still earlier Hebrew scrolls, which could reflect vestiges of earlier editorial stages of the biblical books.

Greek and Hebrew Scrolls from the Judean Desert.

The Greek and Hebrew scrolls found in the Judean desert are of major importance for our understanding of the background and early history of the LXX. Likewise, the combined evidence of these two sources is of great importance for understanding the early history of the biblical text.

The Greek Torah fragments from Qumran may reflect the OG better than the uncial manuscripts, while the Minor Prophets Scroll from Naḥal Ḥever provides the best early evidence of early Jewish revisions of the OG (see “Evidence” above).

The Hebrew Qumran manuscripts, especially those from Cave 4, provide extensive evidence of sources agreeing with the LXX in small details. From the beginning of the publication of the scrolls, such agreements in details with the LXX have been recorded in the scholarly literature and the critical editions. However, several such agreements are irrelevant when the two texts share the same exegetical tradition. In other cases, the agreement may be significant contextually even if this is a mere occasional instance not pointing to any consistent pattern of agreement between the two sources.

In some cases a distinct proximity between a scroll and the LXX is recognized. Most convincing are scrolls that agree with the LXX in the latter's characteristic features, either in editorial differences or in a large percentage of meaningful variants. In all instances disagreements with the OG are also taken into consideration. While 4QJerb when extant and reconstructed is almost identical to the reconstructed Vorlage of the LXX, a few other scrolls are very close to that version, sometimes in its characteristic features: (1) 4QJerb agrees with the LXX in almost all its details against MT, including the LXX's characteristic editorial deviations from MT: shortness (both are significantly shorter than MT, in the case of the LXX by one-sixth) and sequence (different sequence in 10:5–12). (2) 4QJerd agrees with the major feature of the LXX (shortness) in personal names. The scroll also differs from the LXX in seven details. (3) 4QDeutq agrees with the LXX against MT in the addition of two significant stichs in Deuteronomy 32:43 that give a polytheistic flavor to the song. The two also agree in four small details, and they differ in three small details. (4) 4QSama agrees often with the LXX against MT in significant readings while disagreeing with it in equally significant readings. This scroll is probably the most difficult scroll to assess because of its separate agreements with either LXX or LXXLuc and because of the changing nature of the LXX which reflects the kaige-Theodotion revision in 2 Samuel 11:1–1 Kings 2:11 and the OG in all other chapters of 1–2 Samuel. 4QSama is often very close to the OG when disagreeing with the MT, leading to the assumption that these two sources were closely related at an early stage while major differences were created in both texts subsequent to their separation from one another. (5) 4QSamb, one of the earliest Qumran scrolls (ca. 250 B.C.E.), is closely related to the LXX in both superior and inferior readings. (6) 4QNumb often agrees with the LXX, especially in several medium-size harmonizing pluses, but it also often disagrees with that translation. At the same time, the first feature that comes to mind in the characterization of this scroll is its great similarity to SP, especially in its major editorial pluses based on Deuteronomy. 4QNumb should therefore be regarded as close to both SP and the LXX. (7) 11QPsa col. XXVIII is closely related to Psalm 151 LXX of which it presents a longer version. Since the Hebrew Vorlage of the LXX version of that Psalm has logical shortcomings in the flow of ideas, it is probable that segments were removed from the scroll editorially.

The texts that are close to the LXX do not form a close-knit textual family like the MT-group or the SP-group, nor were they produced by a scribal school. The individual scrolls happen to be close to the Hebrew texts from which the LXX was translated, but since the Vorlage of each book in the LXX was a single biblical scroll, the recognition of Hebrew scrolls that were close to the Vorlage of the LXX does not contribute to our understanding of the relationship among the Hebrew texts.

Revisions of the Septuagint.

A given textual tradition is considered a revision (recension) of the LXX if two conditions are met:

  • (1) The LXX and the revision share a common textual basis. If a common basis cannot be established, the two sources are considered separate translations rather than a source and its revision.
  • (2) The revision corrects the LXX in a particular way, usually as a more precise reflection of its Hebrew source.

Emergence of the Revisions.

At least three factors were instrumental in the creation of new Jewish Greek versions, which included revisional elements, but probably the main reason for their emergence was the first in the list that follows:

  • (1) Differences between the LXX and the current Hebrew text. The Greek-speaking Jews of Palestine and Egypt required a Greek translation that would faithfully reflect the proto-Masoretic Text that was current from the first century B.C.E. until the second century C.E. Such a text was necessary for their religious needs and, later, for their polemics with Christians.
  • (2) Disregard of the LXX. Jews stopped using the LXX and began new translations because the LXX started to be frequently used by Christians.
  • (3) Jewish exegesis. New Jewish-Greek versions were needed to reflect Jewish exegesis. However, the LXX reflects more elements of Jewish exegesis than do the revisions.

Nature of the Revisions.

The LXX revisions changed the text in a variety of ways. The earliest revisions contain just slight corrections while the later ones contain more extensive and consistent changes. Most of the revisions clearly intended to present scripture more precisely and consistently than the OG.

Several of the revisions contained the entire Jewish scripture. Thus, kaige-Theodotion included at least Baruch and the expanded version of Daniel, while Origen and Lucian included most of the Apocrypha. For most revisions, however, we do not know how many of the biblical books were included. Some of the revisions, such as kaige-Theodotion and Aquila, were widely circulated.

The Hexapla, which is of particular importance for the textual history of the LXX, is central in the classification of the revisions. It can be subdivided into the following three groups: pre-Hexaplaric revisions, the Hexapla, and post-Hexaplaric revisions.

Pre-Hexaplaric Revisions.

The revisions of Aquila, Symmachus, and Theodotion—in that order—are referred to as the “Three.” Several elements from these three revisions have been preserved among the remnants of the Hexapla, in various papyrus fragments, in marginal notes in Hexaplaric manuscripts of the LXX, and in quotations by the church fathers.


Kaige-Theodotion is the modern name of an early anonymous revision of the OG, dating to the middle of the first century B.C.E., at first identified in 8ḤevXII gr, the Greek Minor Prophets Scroll from Naḥal Ḥever. This text contains an early revision of the OG, and was named kaige. Barthélemy (1963) chose this name because one of its distinctive features is that the Hebrew word gam (also) is usually translated with kaige (at least) apparently in accordance with one of the thirty-two rabbinic hermeneutical rules, or middot, of R. Eliezer ben Yose ha-Gelili named “inclusion and exclusion.” To what extent kaige-Theodotion followed rabbinic exegesis in other details as well (as claimed by Barthélemy) remains a matter of debate.

A similar revision of the OG is also found in several segments of the “LXX” in Samuel-Kings (2 Sam 11:1 [10:1?]–1 Kgs 2:11 and 1 Kings 22:1–2 Kings), the B text of the “LXX” of Judges, and the “LXX” of Ruth and Lamentations. Barthélemy also assigned to this group the sixth column (attributed to Theodotion) and the Quinta, or fifth, Greek column of the Hexapla.

In antiquity, this anonymous revision was associated with Theodotion, who apparently lived at the end of the second century C.E., and was probably from Ephesus. Because Theodotion's translations belong to this group of revisions, the whole collection came to be known as kaige-Theodotion even though its various attestations are not uniform in character and accordingly different individuals may have been involved. The central parts of the presumed kaige-Theodotion revision are the above-mentioned segments of Samuel–Kings and 8ḤevXII gr. (For an updated review of the literature on kaige-Theodotion, see Kraft 2004 and Hugo 2006).


Aquila's revision dates to approximately 125 C.E. He issued two different editions of his revision for some biblical books, but the relation between them is unclear. Aquila's translation system is the most slavishly literal of the translators, creating a translation that is often not understandable without knowledge of the Hebrew source text. He believed that every letter and word in the Bible is meaningful, and therefore attempted to accurately represent every word, particle, and morpheme. For example, he translated the Hebrew nota accusativi ’et (the sign of the direct object) separately with syn (with) on the basis of the other meaning of ’et (namely “with”). In his linguistic approach toward translation, Aquila paid much attention to the etymology of the Hebrew words, and this, more than any presumed rabbinic exegesis, characterizes his version.

Some scholars, such as Friedmann (1896) and Silverstone (1931), think that Aquila is the “Onqelos the proselyte” mentioned in the Talmud (b. Meg. 3a and elsewhere) as the author of the Targum of the Torah. However, although the names Aquila and Onqelos are similar, there is no evidence that the same person translated the Torah into Aramaic and revised the LXX. While both translations are exact, the amount of adherence to the MT of the Greek translation is greater than that of the Aramaic one. Hyvärinen (1977, pp. 88–89) notes the closeness, not identity, between Aquila and the translator of Qohelet.


Symmachus's revision is usually dated to the end of the second century or the beginning of the third century C.E. His background and religious affiliation is unclear. According to Epiphanius, Symmachus was a Samaritan who had become a proselyte. Eusebius and Jerome state that he belonged to the Jewish-Christian Ebionite sect. The scholars Geiger (1910) and Salvesen (1991, p. 297) think that Symmachus was Jewish, while Barthélemy (1974) identified him with Somchos, a disciple of R. Meir, mentioned in b. ‘Erub. 13b. Salvesen (p. 296) suggests that Symmachus was active in Caesarea.

Symmachus was very precise (his revision, like Aquila's, was based on kaige-Theodotion), but, he also very often translated ad sensum rather than representing the Hebrew words with fixed equivalents or stereotyped renderings.


The Hexapla (or “sixfold” edition) was an edition of the Bible arranged in six columns. The work was completed by Origen in the middle of the third century C.E., and included the Hebrew text, its transliteration in Greek, and four Greek translations. The fifth column, which contains the LXX, included a notation of the quantitative differences between the LXX and the Hebrew text: elements extant in Greek but not in Hebrew were denoted with an obelos (÷), while elements extant in Hebrew but not in the LXX, which were added to the fifth column mainly from the sixth column, kaige-Theodotion, were denoted with an asteriskos (⊗).

This composition was mainly intended for the internal requirements of the church (thus Brock 1974 based on Origen, Ep. ad Afr. 5), although Origen probably had other motives as well. Readings from the Hexapla are known from the Syro-Hexapla (the translation of the Hexapla into Syriac, prepared in the seventh century by Paul of Tella), a small number of preserved fragments, quotations by church fathers, and marginal notes in LXX manuscripts.

Post-Hexaplaric Revisions.

The most important post-Hexaplaric revision is that of Lucian of Antioch (d. 312 C.E.). The revision was rediscovered in the nineteenth century in the four minuscules in the Cambridge editions. It is also known from Greek and Latin sources antedating Lucian's time. Especially noteworthy are the agreements between LXXLuc and some Hebrew texts from Qumran (in particular 4QSama, c), the biblical quotations in Theodoret and Theophilus of Antioch, and the Vetus Latina, antedating the historical Lucian by 100–350 years. These early readings have been preserved in LXXLuc because the base layer of Lucian's revision, the Antiochene Greek text, probably constituted the OG text (itself based on a Hebrew text often deviating from the MT). That text has been lost in the sections in 1–4 Kingdoms that have been replaced with the kaige-Theodotion revision (2 Sam 11:1–24:25 and 2 Kings). Because of this historical coincidence, LXXLuc reflects in several sections of the historical books important Greek variants and through them important Hebrew variants.

The Greek Versions and Christianity.

At one time, the LXX was considered to be inspired scripture by both Jews and Christians, but now it is sacred only for the Eastern Orthodox Church, while it still has an important, though not canonical, position within other Christian churches. The translation started off as a Jewish enterprise (see above), and was accepted by the early Christians when they were still part of Judaism. From that time onward, when Jews had already abandoned the LXX, Christianity held on to the LXX as scripture until the time of the Vulgate, which assumed official status in the Western Church some time after its creation. At an earlier stage Judaism had changed its approach toward the Jewish-Greek translation, when Jews turned their back on the LXX in the pre-Christian period, and to an even greater extent after the rise of Christianity. Already in the first century B.C.E. it was realized that the Greek translation did not reflect the Hebrew Bible current in Palestine, and at that time the process of revision of the OG toward the proto-Masoretic Text started to take shape.

In the first century C.E., when the NT writers quoted the earlier scripture, they used the wording of the LXX. That was a natural development since the NT was written in Greek, and under normal circumstances its authors would quote from earlier scripture written in the same language.

At the same time, as a result of Jews abandoning the Jewish-Greek translation in the first centuries C.E., that translation was held in contempt in its own environment (Sof. 1.7 and parallels) in spite of its being a Jewish biblical version. Whether or not rabbinic Judaism officially rejected the LXX is unclear, but it was definitely disregarded since the rabbis did not quote from it.

The LXX lost its central position in Judaism from the first century C.E. onward. Subsequently, this process was accelerated when that translation was used as the official source for scripture in the writings of early Christianity. Christians accepted the LXX generally without changing its wording. At the same time, they inserted some changes in the external features of scripture: Christian scribes adopted the codex instead of the scroll and they introduced abbreviations for the Greek divine names (KS [kyrios “Lord”], ThS [theos “God”], and ChS [christos “Christ”], etc.).

The LXX influenced the NT at various levels because early Christianity adopted the LXX as its scripture. The influence of the LXX is visible in the areas of the language, terminology, and theological foundations of the NT, as well as in its frequent quotations.


As an extension of the use of the LXX language in the NT, several LXX words became technical terms in the NT. Thus, christos, originally a rendering of Hebrew mašî(a)ḥ (“anointed”) became the central name and title of Jesus.

Theological Foundations.

Quotations from the LXX in the NT are meant to prove that the message of the Hebrew scriptures (mediated through its Greek translation) is being fulfilled in the new writings. From the point of view of the NT, Hebrew scripture as a whole is thus considered prophetic writing. Some of these quotations pertain to the NT's theological foundations, which are based on the exact wording of the LXX. Thus, the idea of the Messiah's birth to a parthenos, “virgin” (Matt 1:23 and the parallel story in Luke 1:31) is based on an idiosyncratic equivalent created by the Greek translator of Isaiah 7:14.

Quotations of the LXX in the NT.

The main source of influence of the LXX on the NT is through its frequent quotations, especially from Isaiah, Psalms, and the Pentateuch, directly influencing the language and terminology of the NT as discussed above.

Remarkably, often the biblical text quoted in the NT reflects the main LXX uncials, although there are no statistics for the NT as a whole. Closest to the wording of the LXX are probably the citations in the Gospel of John, Luke-Acts, and the Catholic Epistles, as well as individual quotations in the other books. The text of these quotations is often close to codex A of the LXX or, more generally, Alexandrian witnesses.

Many NT quotations differing from the LXX are closer to the MT than to the LXX ad loc. This situation is recognized especially when the LXX ad loc. differs from the MT because of its different Hebrew Vorlage or its free translation character. In the case of the free translation of the LXX of Isaiah, we can easily recognize these relations. In such cases we can often identify the versions that are quoted in the NT, especially the revision of kaige-Theodotion preceding the writing of the NT books (see above). This line of research was initiated by Barthélemy (1963) within the realm of LXX studies, and was continued within NT studies by such scholars as Koch (1986), Menken (2004), and Wilk (2005). It is now clear that Matthew and Paul often quoted from kaige-Theodotion and other revisions of the OG. There is no reason to assume that Matthew or Paul themselves produced these literal translations, because the agreements between the quotations and known revisions such as kaige-Theodotion are too obvious. A well-known example of a quotation agreeing with a non-Septuagintal source is the one from Isaiah 25:8 in 1 Corinthians 15:54 quoting not the LXX but kaige-Theodotion.

It seems that the LXX (OG) was quoted in most writings of the NT, and that the use of kaige-Theodotion by Matthew and Paul pertains to a minority of the quotations. It remains intriguing that Paul used both the OG version and the kaige-Theodotion revision for the same biblical book (Isaiah), even in the same epistles (Romans, 1 Corinthians). Paul likewise quotes from LXX revisions in 1 Kings (3 Reigns) and Job, but he quotes more frequently from the OG. Paul apparently quoted from different Greek versions without distinguishing between his sources or possibly he revised some of his own writings according to different LXX manuscripts.

The case of Matthew's Bible is similar and different at the same time. Matthew reflects both the OG and an early revision, but these two sources probably derived from different layers of Matthew's compositional process (see Menken).

The use of different Greek versions by the same NT authors reflects the diverse textual situation in Palestine of that time, as known from the finds from the Judean desert. The Minor Prophets Scroll from Naḥal Ḥever (mid-first century B.C.E.) reflects the kaige-Theodotion revision, and the Qumran finds (first century B.C.E.) reflect the OG (see above). These very different Greek sources were used in different socioreligious environments, the OG in the Qumran community that practiced an open textual approach toward Hebrew scripture, and the kaige-Theodotion revision of the Minor Prophets (= MT) among the freedom fighters of Bar-Kochba who adhered strictly also to the MT for Hebrew scripture, as visible in all the find sites in the Judean desert except for Qumran (see Tov 2008).



Primary Sources

  • Brooke, Alan E., Norman McLean, and Henry St. John Thackeray, eds. The Old Testament in Greek According to the Text of Codex Vaticanus. Cambridge, U.K.: Cambridge University Press, 1906–1940.
  • Computer-Assisted Tools for Septuagint Studies (CATSS). Philadelphia/Jerusalem, directed by Robert A. Kraft & Emanuel Tov.
  • Göttingen Septuagint: Septuaginta, Vetus Testamentum graecum auctoritate societatis litterarum gottingensis editum. Göttingen: Vandenhoeck & Ruprecht, 1926.
  • Hatch, Edwin, and Henry Adeney Redpath, eds. A Concordance to the Septuagint and the Other Greek Versions of the Old Testament. Oxford: Clarendon, 1897–1906. Reprint, Graz: Akademische Druck- u. Verlagsanstalt, 1954; 2d ed.: Kraft, Robert A., and Emanuel Tov, eds. Grand Rapids, Mich.: Baker, 1998.
  • Holmes, Robert, and James Parsons, eds. Vetus Testamentum graecum cum variis lectionibus. 5 vols. Oxford: Clarendon, 1798–1827.
  • Pietersma, Albert, and Benjamin G. Wright III, eds. A New English Translation of the Septuagint and the Other Greek Translations Traditionally Included Under That Title. Oxford: Clarendon, 2007.
  • Rahlfs, Alfred. Septuaginta, id est Vetus Testamentum graece iuxta LXX interpretes. Stuttgart: Württemberger Bibelanstalt, 1935.
  • Rahlfs, Alfred. Septuaginta, id est Vetus Testamentum graece iuxta LXX interpretes. Edited by Robert Hanhart. 2d ed. Stuttgart: Deutsche Bibelgesellschaft, 2006.
  • Rahlfs, Alfred, and Detlef Fraenkel. Verzeichnis der griechischen Handschriften des Alten Testaments, I, 1, Die Überlieferung bis zum VIII. Jahrhundert. Göttingen: Vandenhoeck & Ruprecht, 2004.
  • Septuaginta Deutsch.

Secondary Sources


    • Bickerman, Elias. Studies in Jewish and Christian History, Part One. Leiden, Netherlands: Brill, 1976.
    • Dorival, Gilles, Marguerite Harl, and Olivier Munnich. La Bible grecque des Septante—Du judaïsme hellénistique au christianisme ancien. Paris: Cerf/C.N.R.S, 1988.
    • Jellicoe, Sidney. The Septuagint and Modern Study. Oxford: Clarendon, 1968.
    • Jobes, Karen H., and Moises Silva. Invitation to the Septuagint. Grand Rapids, Mich.: Baker Academic., 2000.
    • Kahle, Paul. “Untersuchungen zur Geschichte des Pentateuchtextes.” Theologische Studien und Kritiken 88 (1915): 399–439 = Kahle, Paul. Opera Minora. Leiden, Netherlands: Brill, 1956: 3–37.
    • Lagarde, Paul de. Anmerkungen zur griechischen Übersetzung der Proverbien. Leipzig: Brockhaus, 1863.
    • Lust, Johan. “Septuagint and Canon.” In The Biblical Canons, edited by Jean-Marie Auwers & Henk Jan de Jonge, pp. 39–55. Bibliotheca ephemeridum theologicarum lovaniensium 163. Leuven: University Press/Peeters, 2003.
    • Muraoka, Takamitsu. A Greek ≈Hebrew/Aramaic Two-Way Index to the Septuagint. Louvain: Peeters, 2010.
    • Swete, Henry Barclay. An Introduction to the Old Testament in Greek. 2d ed. Cambridge U.K.: Cambridge University Press, 1914.
    • Tov, Emanuel. “The Greek Biblical Texts from the Judean Desert.” In Hebrew Bible, Greek Bible, and Qumran: Collected Essays, by Emanuel Tov, pp. 339–364. Texte und Studien zum antiken Judentum 121. Tübingen: Mohr Siebeck, 2008.
    • Tov, Emanuel. “Reflections on the Septuagint with Special Attention Paid to the Post-Pentateuchal Translations.” In Die Septuaginta–Texte, Theologien, Einflüsse: 2. Internationale Fachtagung veranstaltet von Septuaginta Deutsch (LXX.D), Wuppertal 23.–27.7 2008, edited by Wolfgang Kraus and Martin Karrer, pp. 3–22. Wissenschaftliche Untersuchungen zum Neuen Testament 252. Tübingen: Mohr Siebeck, 2010.
    • Tov, Emanuel. The Text-Critical Use of the Septuagint in Biblical Research. 2d ed. Jerusalem Biblical Studies 8. Jerusalem: Simor, 1997.
    • Veltri, Giuseppe. Eine Tora für den König Talmai: Untersuchungen zum Übersetzungsverständnis in der jüdisch-hellenistischen und rabbinischen Literatur. Texte und Studien zum antiken Judentum 41. Tübingen: Mohr [Siebeck], 1994.

    Retroversion and Translation Technique

    • Olofsson, Staffan. The LXX Version: A Guide to the Translation Technique of the Septuagint. Coniectanea biblica: Old Testament Series 30. Lund: Almqvist & Wiksell, 1990.
    • Polak, Frank H., and Galen Marquis. A Classified Index of the Minuses of the Septuagint, Part I: Introduction; Part II: The Pentateuch. Computer Assisted Tools for Septuagint Studies Basic Tools 4, 5. Stellenbosch:, 2002.
    • Tov, Emanuel. The Text-Critical Use of the Septuagint in Biblical Research. 2d ed., revised and enlarged. Jerusalem Biblical Studies 8. Jerusalem: Simor, 1997.
    • Tov, Emanuel, and Benjamin G. Wright. “Computer-Assisted Study of the Criteria for Assessing the Literalness of Translation Units in the LXX.” In The Greek and Hebrew Bible: Collected Essays on the Septuagint, by Emanuel Tov, pp. 219–237. Vetus Testamentum Supplements 72. Leiden, Boston, and Cologne: Brill, 1999.


    • Barthélemy, Dominique. Les devanciers d'Aquila. Vetus Testamentum Supplements 10. Leiden, Netherlands: Brill, 1963.
    • Barthélemy, Dominique. “Qui est Symmaque?” Catholic Biblical Quarterly 36 (1974): 451–465.
    • Brock, Sebastian. “Origen's Aims as a Textual Critic of the Old Testament.” In Studies in the Septuagint: Origins, Recensions, and Interpretation, edited by Sidney Jellicoe and Harry M. Orlinsky, pp. 343–346. New York: Ktav, 1974.
    • Friedmann, Meïr ben Jeremiah. Onkelos und Akylas. Vienna: Israel.-theol. Lehranstalt, 1896.
    • Geiger, Abraham. Gesammelte Abhandlungen. Warsaw: Tushiyah, 1910. See pp. 51–59.
    • Hugo, Philippe. “Le grec ancien des livres des Règnes: Une histoire et un bilan de la recherche.” In Sôfer Mahîr: Essays in Honour of Adrian Schenker Offered by Editors of Biblia Hebraica Quinta, edited by Yohanan A. P. Goldman, Arie van der Kooij, and Richard D. Weis, pp. 113–141. Vetus Testamentum Supplements 110. Leiden, Netherlands, and Boston: Brill, 2006.
    • Hyvärinen, Kyösti. Die Übersetzung von Aquila. Coniectanea biblica: Old Testament Series 10. Lund: LiberLäromedel-Gleerup, 1977.
    • Kraft, Robert A. “Reassessing the Impact of Barthélemy's Devanciers, Forty Years Later.” Bulletin of the International Organization for Septuagint and Cognate Studies 37 (2004): 1–28.
    • Salvesen, Alison. Symmachus in the Pentateuch. Journal of Semitic Studies Monograph 15. Manchester, U.K.: University of Manchester, 1991.
    • Silverstone, A. E. Aquila and Onkelos. Manchester, U.K.: Manchester University Press. 1931.

    The Greek Versions and Christianity

    • Hengel, Martin. The Septuagint as Christian Scripture: Its Prehistory and the Problem of Its Canon. Edinburgh and New York: T. & T. Clark, 2002.
    • Koch, Dietrich-Alex. Die Schrift als Zeuge des Evangeliums. Untersuchungen zur Verwendung und zum Verständnis der Schrift bei Paulus. Beiträge zur historischen Theologie 69. Tübingen: Mohr, 1986.
    • Menken, Maarten J. J. Matthew's Bible—The Old Testament Text of the Evangelist. Bibliotheca ephemeridum theologicarum lovaniensium 173. Leuven: University Press; Dudley, Mass.: Peeters, 2004.
    • Wilk, Florian. “The Letters of Paul as Witnesses to and for the Septuagint Text.” In Septuagint Research: Issues and Challenges in the Study of Greek Jewish Scriptures, edited by Wolfgang Kraus and R. Glenn Wooden, pp. 253–271. Society of Biblical Literature Septuagint and Cognate Series 53. Atlanta: Scholars, 2005.

Emanuel Tov