Journal of Confessional Bibliology — Scripture Alone, Preserved by Providence
Confessional Bibliology Article

Aquila, Symmachus, Theodotion, and the kaige Revisions

Aquila, Symmachus, Theodotion, and the kaige Revisions

A Greek reading that follows Hebrew closely can look primitive.

Its word order seems Semitic. One Greek word repeatedly answers to one Hebrew word. Small Hebrew particles receive wooden Greek equivalents. Beside it, a smoother Greek reading may look like paraphrase or corruption.

The natural conclusion is often backward.

Jewish readers repeatedly revised Greek biblical texts toward Hebrew forms current in their own communities. A rigidly Hebraizing reading may therefore be a later correction of freer Greek, not the earliest Greek translation. The phenomenon does not make the Hebrew secondary. It shows translators returning a version to its source.

The proposition of this article is:

Aquila, Symmachus, Theodotion, and the revisions grouped under the name kaige show that Greek Scripture contained competing attempts to represent Hebrew; each form is a historically located witness, not an independent authority.

Before a Greek reading is called “the Septuagint,” its layer must be identified.

Three Names Beside the Fifth Column

Origen placed the versions of Aquila, Symmachus, and Theodotion beside the Greek text associated with the Seventy. Ancient evidence and later scholarship assign them broadly to the first centuries of the Christian era, after the initial Greek translations of many Old Testament books.

Their translation profiles differ.

  • Aquila is associated with an intensely formal rendering. Hebrew forms, particles, and word relations are represented in Greek with unusual consistency, sometimes at the expense of ordinary Greek style.
  • Symmachus is generally associated with more idiomatic and elegant Greek, though still engaged closely with Hebrew.
  • Theodotion presents the most difficult historical problem. The version bearing his name is often revisionary rather than an entirely new translation, and “Theodotionic” readings appear in evidence that may precede the traditional date assigned to the historical Theodotion.

These descriptions are profiles, not mechanical rules. The versions survive mainly in fragments, citations, marginal readings, and Hexaplaric witnesses. We do not possess each translator’s complete signed edition.1

Nor should biographical traditions decide textual questions. Ancient writers disagree about the translators’ dates, communities, and religious histories. A reading is not Aquila’s merely because it is literal, or Symmachus’s because it is smooth. Attribution depends upon actual versional evidence.

What kaige Means

The word kaige comes from the recurring Greek expression kai ge, used to render a particular Hebrew combination meaning something like “and also” or “and indeed.” Scholars found that this rendering belonged to a wider cluster of consistent correspondences.

The name now identifies a revisionary phenomenon: Greek texts were systematically altered to agree more closely with a Hebrew form resembling the consonantal tradition later represented by the Masoretic Text.

The revision does not consist of one favorite phrase. Its profile includes repeated choices for Hebrew particles, verb forms, vocabulary, and word order. Dominique Barthélemy’s study of the “predecessors of Aquila” showed that such highly consistent Hebraizing revision existed before the mature version associated with Aquila.2

This corrected an older historical shortcut.

Scholars had found readings like those of Theodotion in sources thought earlier than the second-century translator. Calling them “proto-Theodotion” acknowledged the problem but did not explain it. The kaige model treats many of these readings as part of a broader Jewish revisionary activity whose relation to the later Theodotion tradition must be investigated rather than assumed.

Thus kaige-Theodotion can be a useful label for a textual relationship. It must not be turned into the claim that one historical person named Theodotion produced every reading in the group.

The Nahal Hever Control

The Greek Minor Prophets scroll from Nahal Hever, normally designated 8HevXIIgr, gives the discussion a physical anchor.

Fragments of the scroll preserve portions of the Twelve Minor Prophets in Greek. The hand is conventionally assigned around the turn of the era—often to the late first century BC or early first century AD. That remains a palaeographical assignment, not a production date supplied by a colophon.3

The scroll is important because its revisions are visible in an ancient Jewish artifact. Its Greek stands in a relationship to an older Greek form of the Twelve while repeatedly moving toward a Hebrew text close to the proto-Masoretic tradition.

The manuscript therefore establishes several limited but consequential facts:

  1. Jewish Greek biblical transmission was active and revisionary.
  2. Revision toward a Hebrew text close to the later Masoretic tradition occurred by the period assigned to the scroll.
  3. A Greek form more closely aligned with that Hebrew can be later than the freer Greek wording it replaces.
  4. “Greek Scripture before Christianity” was not one frozen edition.

The scroll does not establish that every Hebraizing reading in every book belongs to the same recension. It does not date Aquila’s complete version. It does not identify every earlier Greek reading. It does not turn the surviving Greek into an authority over Hebrew.

It is a control case, not a universal template.

Revision Is Not Ordinary Copying

An ordinary copyist attempts to reproduce an exemplar, though changes inevitably occur. A reviser intentionally compares the Greek with Hebrew and alters the Greek according to a policy.

That distinction matters because revision produces patterns.

Suppose an earlier translator rendered one Hebrew expression with several natural Greek equivalents according to context. A reviser may replace those alternatives with the same Greek equivalent wherever the Hebrew expression occurs. The result can be:

  • more predictable;
  • more useful for readers checking Greek against Hebrew;
  • less idiomatic in Greek;
  • and later than the readings it displaced.

Timothy Lee’s analysis of 8HevXIIgr emphasizes precisely this standardization of equivalents. Comparison of the revised scroll with reconstructed Old Greek helps expose the reviser’s method.4

One isolated agreement can be accidental. A repeated bundle of unusual equivalences is evidence of editorial policy.

This is why textual affiliation cannot be decided by counting agreements alone. A dozen ordinary agreements may tell less than three rare, diagnostic ones. The question is not only how many readings agree, but what kind of readings they are and whether they reveal a consistent revision.

A Later Greek Reading Can Preserve Earlier Hebrew

The chronological relations are not one straight line.

A Greek revision can be later than an older Greek translation while representing a Hebrew form earlier than both surviving Greek manuscripts. Four dates must remain separate:

Claim Question
Greek artifact date When was this physical copy produced?
Greek textual-layer date When did this wording enter the Greek tradition?
Hebrew exemplar date When did the Hebrew form used by the reviser circulate?
Present witness date When is each surviving manuscript first attested or documented?

Consequently, “late Greek” does not mean “late Hebrew.” Nor does “literal Greek” mean “earliest Greek.”

This complexity gives Greek revisions genuine historical usefulness. A reviser’s consistent rendering may help identify the Hebrew he read. Yet the inference still passes through a translator’s technique. The Greek artifact is extant; the exact Hebrew exemplar may have to be reconstructed.

Packet Five will address that inferential step in detail.

How Revision Enters Later Codices

Revised readings did not remain in sealed editions.

Readers copied them into margins. Scribes substituted them in running text. Hexaplaric signs were lost. One manuscript family adopted a correction while another retained an older rendering. A book could acquire a mixed text in which Old Greek, Jewish revision, Hexaplaric editing, and ordinary copying all contributed.

A modern critical apparatus may therefore report:

  • a reading attributed to Aquila, Symmachus, or Theodotion;
  • a reading supported by 8HevXIIgr;
  • a main-text reading judged to represent the oldest recoverable Greek;
  • a variant judged to be kaige or otherwise recensional;
  • disagreement among later codices.

The apparatus does not magically identify every layer. It records witnesses and presents editorial judgments that must be examined.

This is why “the Greek says” is often inadequate. Does Greek mean the Nahal Hever scroll, Vaticanus, Sinaiticus, a Hexaplaric marginal note, Rahlfs-Hanhart’s main text, or a Göttingen editor’s reconstruction?

Those objects can agree. They can also represent different moments in a long history.

The Strongest Objections

One objection says:

If Jewish revisers conformed Greek to a proto-Masoretic Hebrew text before or near the time of Christ, does that not prove every later Masoretic reading original?


  1. For the ancient notices and surviving Hexaplaric evidence, see Frederick Field, ed., Origenis Hexaplorum quae supersunt, vol. 1 (Oxford: Clarendon, 1875); Natalio Fernández Marcos, The Septuagint in Context (Leiden: Brill, 2000), chaps. 7–9. ↩
  2. Dominique Barthélemy, Les devanciers d’Aquila (Leiden: Brill, 1963). For a recent account that distinguishes the kaige phenomenon from simplistic identification with the historical Theodotion, see Anneli Aejmelaeus, “The Origins of the Kaige Revision”. ↩
  3. Israel Antiquities Authority, 8HevXIIgr manuscript record and images. The assigned date should always be identified as palaeographical. See also Emanuel Tov, Robert A. Kraft, and Peter J. Parsons, The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr), Discoveries in the Judaean Desert VIII (Oxford: Clarendon, 1990). ↩
  4. Timothy Lee, “Standardizing Translation Equivalents in the Kaige Revision: The Evidence of 8HevXIIgr,” Journal of Septuagint and Cognate Studies 56 (2023). Lee analyzes revision as the systematic standardization of translation equivalents. ↩

It proves less and more.

It proves more than a theory of late medieval standardization allows: Hebrew forms closely related to the Masoretic tradition were guiding Jewish Greek revision in antiquity. It proves less than the identity of every consonant, vowel, accent, and later printed detail. A scroll of the Minor Prophets cannot directly authenticate every reading in every Old Testament book.

A second objection says:

If a Greek reading disagrees with the Masoretic tradition and appears freer, must it preserve the Old Greek and therefore an earlier Hebrew?


  1. For the ancient notices and surviving Hexaplaric evidence, see Frederick Field, ed., Origenis Hexaplorum quae supersunt, vol. 1 (Oxford: Clarendon, 1875); Natalio Fernández Marcos, The Septuagint in Context (Leiden: Brill, 2000), chaps. 7–9. ↩
  2. Dominique Barthélemy, Les devanciers d’Aquila (Leiden: Brill, 1963). For a recent account that distinguishes the kaige phenomenon from simplistic identification with the historical Theodotion, see Anneli Aejmelaeus, “The Origins of the Kaige Revision”. ↩
  3. Israel Antiquities Authority, 8HevXIIgr manuscript record and images. The assigned date should always be identified as palaeographical. See also Emanuel Tov, Robert A. Kraft, and Peter J. Parsons, The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr), Discoveries in the Judaean Desert VIII (Oxford: Clarendon, 1990). ↩
  4. Timothy Lee, “Standardizing Translation Equivalents in the Kaige Revision: The Evidence of 8HevXIIgr,” Journal of Septuagint and Cognate Studies 56 (2023). Lee analyzes revision as the systematic standardization of translation equivalents. ↩

No. It may be Old Greek, a translator’s idiom, an interpretive rendering, a scribal alteration, or a witness to another Hebrew exemplar. The translator’s profile and the external evidence must decide among those possibilities. Disagreement is a question, not a verdict.

A third objection says:

Calling readings “revisions” merely protects the Hebrew from contrary evidence.


  1. For the ancient notices and surviving Hexaplaric evidence, see Frederick Field, ed., Origenis Hexaplorum quae supersunt, vol. 1 (Oxford: Clarendon, 1875); Natalio Fernández Marcos, The Septuagint in Context (Leiden: Brill, 2000), chaps. 7–9. ↩
  2. Dominique Barthélemy, Les devanciers d’Aquila (Leiden: Brill, 1963). For a recent account that distinguishes the kaige phenomenon from simplistic identification with the historical Theodotion, see Anneli Aejmelaeus, “The Origins of the Kaige Revision”. ↩
  3. Israel Antiquities Authority, 8HevXIIgr manuscript record and images. The assigned date should always be identified as palaeographical. See also Emanuel Tov, Robert A. Kraft, and Peter J. Parsons, The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr), Discoveries in the Judaean Desert VIII (Oxford: Clarendon, 1990). ↩
  4. Timothy Lee, “Standardizing Translation Equivalents in the Kaige Revision: The Evidence of 8HevXIIgr,” Journal of Septuagint and Cognate Studies 56 (2023). Lee analyzes revision as the systematic standardization of translation equivalents. ↩

The Nahal Hever scroll prevents that charge. Revision is not invented to escape a variant; it is a historically attested activity with detectable patterns. The classification can still be wrong in a particular passage, which is why the evidence must be published and tested.

The Confessional Boundary

Aquila, Symmachus, Theodotion, and the kaige revisers did not create a second inspired Old Testament.

Their work can clarify how Jewish readers understood Hebrew. It can preserve indirect evidence about ancient Hebrew exemplars. It can show that comparison with Hebrew repeatedly governed the correction of Greek.

That is ministerial work. It is not magisterial authority.

The confessional order does not require every ancient Greek witness to be homogeneous or every difference to be worthless. It requires the translated text to remain accountable to the language in which the Old Testament was immediately inspired.

The revisionary history supplies an ironic warning to anyone who would reverse that relation. Ancient Jewish revisers made Greek answer more closely to Hebrew. Modern critics may study what they did, but they may not turn the revised stream into the judge of its fountain.

Article 18 will bring this layered model to the two codices most often made to stand for “the Septuagint”: Vaticanus and Sinaiticus.

Notes and Further Reading

For the relationship between the Nahal Hever scroll, the Old Greek Twelve Prophets, and revision toward Hebrew, see the Oxford description of the Discoveries in the Judaean Desert volume, The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr). For a general introduction, see Karen H. Jobes and Moisés Silva, Invitation to the Septuagint, 2nd ed. (Grand Rapids: Baker Academic, 2015), chaps. 5–6.

Source-control note: The surviving fragments and photographs of 8HevXIIgr are direct artifact evidence (E1). Its late-first-century-BC or early-first-century-AD placement is an assigned palaeographical range (E3). The identification of Old Greek, kaige equivalents, and relationships among Aquila, Theodotion, and other revisions is a scholarly reconstruction from recurring textual patterns (E5). Neither every Masoretic agreement nor every Greek disagreement can be assigned to a recension without proof (E6). The final authority of the received Hebrew is a confessional judgment (E7).

Previous in the series: “Origen’s Hexapla and the Layered Greek Tradition.”

Next in the series: “Vaticanus and Sinaiticus Are Witnesses, Not ‘the Septuagint.’”


  1. For the ancient notices and surviving Hexaplaric evidence, see Frederick Field, ed., Origenis Hexaplorum quae supersunt, vol. 1 (Oxford: Clarendon, 1875); Natalio Fernández Marcos, The Septuagint in Context (Leiden: Brill, 2000), chaps. 7–9. ↩
  2. Dominique Barthélemy, Les devanciers d’Aquila (Leiden: Brill, 1963). For a recent account that distinguishes the kaige phenomenon from simplistic identification with the historical Theodotion, see Anneli Aejmelaeus, “The Origins of the Kaige Revision”. ↩
  3. Israel Antiquities Authority, 8HevXIIgr manuscript record and images. The assigned date should always be identified as palaeographical. See also Emanuel Tov, Robert A. Kraft, and Peter J. Parsons, The Greek Minor Prophets Scroll from Nahal Hever (8HevXIIgr), Discoveries in the Judaean Desert VIII (Oxford: Clarendon, 1990). ↩
  4. Timothy Lee, “Standardizing Translation Equivalents in the Kaige Revision: The Evidence of 8HevXIIgr,” Journal of Septuagint and Cognate Studies 56 (2023). Lee analyzes revision as the systematic standardization of translation equivalents. ↩
Discussion

Comments

Share a question or response to this article.

Leave a Comment

Your email address will not be published. Required fields are marked.