Executive Summary: Galatians (English → Hebrew)

Why it matters

Galatians is Paul’s most polemically compressed letter, and Hebrew is the one destination language where nearly every key term he argues against — Torah, circumcision, the yoke, the covenant — is also the destination culture’s most beloved vocabulary. Unlike lower-resource target languages, this is not a vocabulary-gap problem; it’s a collision-management problem. Get the framing wrong anywhere in six chapters, and the letter can read as an attack on Judaism itself rather than Paul’s actual, narrower argument: that Gentile justification does not require Torah-observance or circumcision. This package extends the baseline Romans Language Package with the term database, doctrine registry, and AI instruction set needed to prevent that misreading at scale, across every chapter.

Key findings

• 117 total glossary terms now govern Phase 2 Galatians translation: 33 inherited exactly from the baseline Romans package, 84 newly introduced by this book.
• 66 of those 117 terms (56%) are Critical or High risk — 33 Critical, 33 High — meaning a majority of this letter’s working vocabulary requires mandatory human theologian review, not automated or native-speaker-only review.
• 29 doctrines span the full six chapters; 25 require human theologian review (16 Critical + 9 High), 4 require native speaker review, and zero are automated-only — the highest theologian-review burden of any curriculum in this pipeline to date.
• The core passage, Galatians 2:15–21, packs seven independently Critical/High-risk terms into seven verses (works of the law, justification ×3, faith of/in Christ, crucified with Christ, gave himself up, righteousness, grace) — the single densest doctrinal span identified across either curriculum studied so far. It now carries a mandatory dedicated theologian review pass, over and above normal per-segment routing.
• Fifteen ranked “highest-risk ambiguities” have been catalogued (Linguistic Gap Analysis §E), topped by Galatians 2:16’s triple justification-verb repetition and 5:1’s “yoke of slavery” collision with the positive Rabbinic “yoke of Torah.”

Risks

• Reverse risk profile: Hebrew almost never lacks a word for a Galatians concept — it has too good a word, one already carrying a cherished, positive covenantal sense (Torah, brit milah, ol Torah) that runs counter to Paul’s argument. Five terms carry this collision at Critical severity: works of the law, circumcision, yoke of slavery, elemental principles/calendar observance, and the Hagar-Sarah allegory.
• Highest-stakes single phrase: “The Israel of God” (Galatians 6:16) lands with maximal personal and national identity weight for any Hebrew-speaking reader. Any rendering implying Church-replaces-Israel would compound centuries of documented supersessionist harm.
• Two genuinely new Critical-tier gaps not present in the Romans baseline: “crucified with Christ” (no Rabbinic category for participatory union, plus the cross symbol’s Crusade/pogrom history) and “baptized/immersion” (mikveh associations risk flattening once-for-all incorporation into one more purificatory rite).
• Known bad-actor renderings in circulation (Hebrew Roots/“One Law” reframings of 4:9-10 and 5:1; Sacred Name spellings; Jehovah’s Witness deity-of-Christ evasions; anti-missionary polemical retranslations of 3:13/3:16) must be actively screened out, not merely avoided by omission.

Opportunities

• Authentic historical vocabulary already exists for the letter’s central polemical term: “works of the law” (מַעֲשֵׂי הַתּוֹרָה) is independently attested at Qumran (4QMMT) — a genuine, teachable asset showing Paul’s Hebrew-equivalent phrase is not an artificial import but real Second-Temple Hebrew.
• Positive intertextual bridges strengthen rather than merely constrain several renderings: Isaiah 53:12 for “gave himself up,” Deuteronomy 27 for the anathema formula, Deuteronomy 21:23 for the “tree” citation, Leviticus 19:18 for “love your neighbor,” and the Talmud’s own “leaven in the dough” (Berakhot 17a) parallel for Paul’s leaven warning.
• Hebrew’s own grammar is an asset Paul’s English readers don’t get for free: זֶרַע (seed) is already number-ambiguous, transparently supporting the singular/plural “seed of Abraham” argument (3:16, 3:29) more visibly than in English.
• Five terms are recommended for promotion into the shared cross-curriculum translation memory (baptism, flesh, cross/crucified, redeem/redemption, eternal life) — closing gaps the Romans baseline left latent and benefiting every future NT epistle curriculum in this pipeline.

Recommended actions

1. Route Galatians 2:15–21 as a standalone theologian review unit in Phase 2 Step 17, independent of normal segment-level flagging, given its unmatched Critical-term density.
2. Enforce the crowded-neighborhood fencing rule (new Glossary Enforcement Priority tier 4 in 12_ai_translation_requirements.md): a lexically correct rendering without disambiguating framing is an incomplete translation, not a stylistic option — apply this rigorously to Torah, circumcision, yoke, and Israel-of-God language above all.
3. Lock the forbidden-substitution list before batch processing begins: never מוֹרֶה for the law’s guardian role, never חֵרֶם for anathema, never גְּאוּלָה for the commercial “redeem,” never any qualifier on “the Israel of God.”
4. Confirm the YouVersion Hebrew NT version_id (recommended: Bible Society in Israel’s HaBrit HaChadashah, 1991 revision) against the live catalog before generating any batch hyperlinks.
5. Escalate the five cross-curriculum-promotion candidates (baptism, flesh, cross/crucified, redeem, eternal life) to baseline maintainers for formal incorporation, so future NT curricula inherit them rather than re-deriving them.
6. Brief all human theologian reviewers on the letter’s reversed risk profile before review begins: the task is rarely “find the missing word” and almost always “verify the correct word has been fenced against its own cherished, positive Jewish meaning.”

---

This summary reflects the full Phase 1 Galatians analysis set: 04_comparative_theology.md, 05_translation_landscape.md, 06_linguistic_gap_analysis.md, 07_semantic_analysis.md, 08_core_glossary.md, 11_doctrine_analysis.md, assets/translation_memory.json, assets/bible_term_registry.json, and assets/doctrine_risk_registry.json. It extends, and does not contradict, the baseline Romans Language Package executive findings.