Ezekiel — Executive Summary (Hmong Language Package, Phase 1)

The bottom line: Ezekiel is the highest-collision-density book this pipeline has yet processed for Hmong. Every one of its seven assigned doctrines carries Critical or High doctrinal-risk status, and the book’s core passage (37:1-14) contains the single most linguistically complex untranslatable wordplay in the curriculum. This package is ready for Phase 2 generation only under strict, non-negotiable enforcement of the terms and rules below.

---

Why it matters

Ezekiel speaks directly into the exact cultural terrain Hmong traditional religion occupies — spirit-possession trance, ancestor bone/grave veneration, household-spirit appeasement, and trans-generational fate (hmoov) — more densely and more specifically than any other book in this pipeline to date. Where Romans occasionally brushed against these categories, Ezekiel’s central images (dry bones, departing/returning presence, inherited guilt) sit directly on top of live Hmong cultural practice. Getting this package right is a precondition for teaching Ezekiel in Hmong without unintentionally re-baptizing existing animistic categories in biblical language.

Key findings

• 7 of 7 doctrines are Critical or High risk (5 Critical: Glory/Sovereignty, Individual Responsibility, New Heart/New Spirit, Restoration/Dry Bones, Presence Departing/Returning; 2 High: Judgment on Jerusalem/Nations, Vision of the New Temple) — 100% require mandatory human theologian review before publication, per assets/doctrine_risk_registry.json.
• 68 terms logged in the core glossary; 29 are Critical or High risk (11 Critical + 18 High), requiring human theologian review before publication; 21 are Medium-risk (native-speaker review); 18 are Low-risk (automated review sufficient).
• The compound divine title tus Tswv Vajtswv (Adonai YHWH) recurs roughly 217 times — the single highest-volume fixed-phrase enforcement requirement in the entire pipeline, with a permanent, zero-exception prohibition on any transliterated divine-name substitution.
• The Hebrew ruach word-play (wind → breath → Spirit) across the core passage has no lexical solution in Hmong — it requires three distinct terms (cua / pa / Vaj Ntsuj Plig) plus a mandatory footnote, making Ezekiel 37:1-14 the single most translation-engineering-intensive passage yet handled in this pipeline.
• 32 baseline Romans TM terms are inherited exactly into this package (e.g., Vajtswv, Vaj Ntsuj Plig, dab mlom, kev npau taws ntawm Vajtswv, hloov siab tshiab), several of which now serve as mandatory CONTRAST anchors for new Ezekiel terms that share roots but diverge in valence (e.g., Ezekiel’s positive “heart of flesh” versus Romans’ negative moral “flesh”).
• 2 proper names remain formally unresolved: Gog/Magog and the Zadokite priestly office (ch.44-46) have no confirmed permanent Hmong rendering and are flagged [TRANSLATOR REVIEW NEEDED] pending native-speaker and theologian input.

Risks

• Sharpest single risk: “Dry bones” and “opened graves” (37:1-14, 12-13) sit directly on top of Hmong ancestor bone-and-grave veneration custom — the belief that correct bone/grave treatment secures an ancestor’s peace. Left unaddressed, God’s unilateral resurrecting act could be heard as one more ritual-efficacy category.
• Second-sharpest risk: God’s presence departing (ch.8-11) and returning (ch.43, 48:35) almost exactly mirrors the traditional pattern of an offended household/village spirit withdrawing its protection until appeased by correct ritual. Both poles require explicit, sustained contrast-teaching or the doctrine’s entire narrative arc risks being absorbed into an existing animistic framework.
• Cross-curriculum consistency risk: Ezekiel 18’s individual-accountability thesis (“the son shall not bear the father’s iniquity”) could be misread as contradicting Romans 5’s federal Adam-Christ headship doctrine if the two are not explicitly taught as compatible, distinct-register truths.
• Reversed-caution trap: dawb huv (holiness) requires opposite handling instructions in Romans-sanctification contexts versus Ezekiel’s temple-architecture contexts (ch.40-46) — a pattern-matching failure here produces an internal doctrinal inconsistency rather than a traditional-religion collision, but is no less serious.
• Volume risk: the compound divine title’s ~217 occurrences and the recognition formula’s 60+ occurrences make consistency-under-volume itself a named risk category distinct from lexical-collision risk — any drift in either fixed formula propagates across most of the book.

Opportunities

• Islamic and Jewish comparative bridges: the Qur’an’s own dry-bones resurrection imagery (36:78-79) and Jewish techiyat ha-metim tradition offer unusually strong positive cross-tradition bridges for teaching Ezekiel 37 to audiences with any Islamic or Jewish contact, per analysis/04_comparative_theology.md.
• Agrarian/lineage resonance: land-inheritance imagery (ch.47-48) and clan-adoption-style covenant language resonate positively and safely with existing Hmong lineage-inheritance custom, requiring only comprehension-level (not collision-level) handling.
• Reinforcement of the existing Romans baseline: because this package inherits 32 Romans TM terms exactly, the two packages can be taught in sequence with genuine cross-reinforcement (e.g., the indwelling Spirit promise of Ezekiel 36:27/37:14 as the direct Old Testament anticipation of Romans 8) rather than requiring learners to master two disconnected vocabularies.
• Full-book coverage discipline: all 48 chapters have been logged (21 explicitly marked “reviewed — no new terms” rather than silently skipped), giving Phase 2 generation a complete, audit-ready map of exactly where new vocabulary risk is concentrated versus merely reinforced.

Recommended actions

1. Lock the translation memory (assets/translation_memory.json) as provisional-but-enforced immediately — treat all “NEW — provisional” entries with full Critical/High enforcement weight in Phase 2 generation pending formal TM-revision review.
2. Route 100% of doctrine-touching content to human theologian review before publication — there is no Medium/Low-risk doctrine in this package eligible for a lighter review path.
3. Prioritize theologian sign-off on the core passage (37:1-14) and Ezekiel 18 contrast-teaching notes first, since these carry the highest combined severity ranking in analysis/06_linguistic_gap_analysis.md §6.
4. Resolve the two open [TRANSLATOR REVIEW NEEDED] proper-name/office items (Gog/Magog transliteration; Zadokite priestly-office rendering) via native-speaker and theologian consultation before any ch.38-39 or ch.44-46 content is generated.
5. Build the dawb huv reversed-caution check into the Phase 2 generation pipeline as an explicit context-detection step, not a stylistic reminder, given its unique status as the one term in the combined Romans/Ezekiel package requiring opposite handling by context.

---

End of analysis/01_executive_summary.md