Method

How a plus becomes a verdict. Every step leaves a record in the published data.

1. The difference layer

The Septuagint and the Masoretic text were aligned from the CATSS parallel files, which the project uses as an index only; the Greek published here is Swete's. Each Addition was cut into units of one clause or one motif, and given a draft English translation and a one-line summary of what it contributes.

2. The bench

27 works, 2,125 paragraphs, pulled from Sefaria and Hebrew Wikisource at pinned versions, each paragraph carrying its source, licence and date. Composite works are dated per stratum.

3. Search

First a lexical pass: two to four Hebrew retroversions per unit, matched against the bench with weighted word-sequence overlap. Then three reading passes by language models under materially different briefs, each reading every bench paragraph against every unit and keeping anything plausible. Together they produced 1,249 candidate pairs.

4. Verification and rulings

A separate verification pass read each candidate against its unit and ruled it verbatim, motif or unrelated under written rejection criteria: material the Masoretic text already supplies, stock biblical idiom, long paragraphs matched on vocabulary alone, a shared word with a different referent, anything explicable from a different biblical book, a passage that contradicts the plus, and a copy of another witness. The editor spot-checked forty of its confident calls and ruled how each work transmitted what it carries. An adversarial reviewer then attacked every published claim; its findings were adopted in full. A second round re-read every pair the first could not settle (below).

Verbatim means the witness reproduces two or more of the plus's clauses, in the plus's order, with the same content and at least one rendering of a specific word of the plus that is neither the Masoretic text's nor stock. Language does not matter. Motif means a specific shared detail the Masoretic text does not supply.

Final count: 72 verbatim, 136 motif, 1041 unrelated.

5. How much to trust the motif tier

After the first round, not much: in a seeded sample of forty motif pairs the adversarial reviewer rejected 58% (95% interval about 42–72%), and the negative controls showed the same leak. So a second round re-ruled the whole tier. The brief was sharpened with worked examples of the commonest errors, and a fresh reading had to agree with the reviewer's rulings on at least 34 of the same forty pairs before anything else ran; it agreed on 35. Then all 509 uncertain pairs were re-read, each blind to its earlier ruling, and 366 (72%) were rejected. The 136 motif pairs that remain are a confirmed count, not a candidate list.

They are not perfect. When the confirmed pairs were fingerprinted (section 8), the readers, told that some pairs might share nothing, found no specific shared detail in 48 of the 136. That is a second, independent measure of the tier, published with it; no pair has been re-ruled on its strength.

6. Controls

Pairs fixed from the scholarly literature before any search, to test recall (positive), precision (negative), and the machinery itself.

7. Conduits

Each kept parallel carries a judgment about how the material reached the witness: a medieval retranslation from Latin, an Aramaic targum, oral tradition, a lost written intermediary, or a common Hebrew source. No pair on this bench carries the last.

8. Translation fingerprints

A detail that one version of Esther has and the others lack is a fingerprint of that version. Every unit was broken into its details (375, plus 188 that one version has and the Septuagint lacks), and each detail was looked for in six versions: the Septuagint, the Alpha-text (Lagarde), the Old Latin (Sabatier), the Vulgate, Josephus and Sefer Yosippon (Breithaupt). Five readers then traced each confirmed pair to the versions that carry its shared detail, blind, mixed with fifty controls. All ten Yosippon controls pointed to Yosippon and nine of ten Vulgate controls to the Vulgate; none of thirty negatives was assigned a route.

The result: the motifs show that the content of the Additions was known, but they rarely separate one version from another, because the versions share most of the Additions' details. The one ancient text that does separate them is Targum Sheni 8:13, whose letter follows a text of the Greek type (in Greek or in the Old Latin) and not the Vulgate, checked against Haelewyck's critical edition of the Old Latin. See the guide for the whole method in plain language.