Executive Summary: Haggai — English → French

Why it matters

Haggai is short (38 verses, 2 chapters) but introduces a translation-risk profile the Romans baseline never had to face: the Hebrew original names God by the Tetragrammaton, not by NT Christological titles. Getting the divine name wrong — or letting it visually blur with the baseline’s NT “Seigneur” — would destabilize the OT-covenant-name / NT-Lordship-confession distinction across every future French curriculum built on this translation memory. Haggai also contains the book’s single highest-stakes phrase-level decision in the whole pipeline to date: the grammatically plural “treasures of all the nations” (2:7), long mistranslated in English devotional tradition as a singular messianic title. Both issues require resolution before any Phase 2 segment translation begins.

Key findings

• 65 terms are now fixed in assets/translation_memory.json: 24 inherited exactly from the Romans baseline (grâce, saint, gloire, paix, loi, Église, élection, Esprit Saint, David, Israël, etc.), 41 newly established for Haggai (le SEIGNEUR, le SEIGNEUR des armées, grand prêtre, mon Esprit, les trésors de toutes les nations, sceau, bénir/bénédiction, and others).
• 16 doctrines are catalogued in the Haggai doctrine risk registry, matching the four curriculum-level doctrines (Priorities, Obedience and Blessing, Presence/Spirit, Coming Glory) plus 12 supporting doctrines identified across the full book.
• Full-book coverage confirmed: every oracular unit (1:1-11, 1:12-15, 2:1-9, 2:10-19, 2:20-23) contributes new load-bearing vocabulary; no chapter is doctrinally silent.
• The divine name and the 2:7 “treasures/desire of nations” clause are the two Critical-tier doctrines in this curriculum (matching the registry’s divine_name_yhwh and coming_glory_of_the_temple entries) — a risk category the Romans baseline registry did not need at all for its NT-only text.

Risks

• Critical-tier terms: 3 — lord_yhwh (“le SEIGNEUR”), lord_of_hosts (“le SEIGNEUR des armées”), desire_of_nations (“les trésors de toutes les nations”). All three require zero-exception, document-wide consistency; a single lapse on the divine name recurs dozens of times per chapter and compounds instantly.
• Critical-tier doctrines: 2 — The Divine Name (YHWH) and God’s Sovereign Identity; The Coming Glory of the Temple. Both require human theologian review for every occurrence, per doctrine_risk_registry.json.
• High-tier terms: 8 (word of the LORD, high priest, glory [elevated from baseline Medium], I am with you, my Spirit [divine], priest, unclean, signet) plus the baseline’s inherited High-tier terms reused in this curriculum (Seigneur, Fils de Dieu, Église, loi, élection, obéissance de la foi).
• High-tier doctrines: 6 — Priorities: God’s House vs. Self-Interest; Obedience and Blessing; Presence and Spirit of God; Priesthood and Ritual Purity Typology; The Day of the LORD and Eschatological Shaking; Davidic Covenant Restoration and Messianic Hope.
• Total requiring theologian oversight: 8 doctrines (2 Critical + 6 High) and 11 terms (3 Critical + 8 High), per the doctrine risk registry and core glossary risk summaries.
• Total requiring native-speaker review: 7 doctrines; total automated-only: 1 doctrine (Divine Sovereignty over Creation and Resources).
• Single largest procedural risk: “grand prêtre”/“prêtre” collides with contemporary French Catholic ordained-clergy associations at four separate points across the book (1:1,12,14; 2:2,4,11-13), the highest occurrence count of any High-risk collision term.
• Single largest typographic risk: the “mon Esprit” (divine, majuscule) / “l’esprit de” (human, minuscule) capitalization fence is invisible in speech and vulnerable to careless copy-editing.

Opportunities

• Haggai’s short length allows the entire divine-name and treasures-clause resolution to be locked in Phase 1, before any Phase 2 segment work begins — a cleaner sequencing than would be possible on a longer book.
• The “bénir/bénédiction” term pair, newly established here (absent from the Romans baseline), can now be reused as a fixed precedent for any future OT curriculum in this translation-memory lineage.
• Haggai’s concrete, image-based doctrine (house, treasure, signet, blessing) is well-suited to vivid, accessible French teaching material for the curriculum’s biblically-unchurched secular audience segment, once the Deuteronomic covenant-curse background is supplied via cross-reference notes.
• The Zerubbabel signet oracle (2:23) offers a strong, underused Davidic-messianic teaching bridge to Matthew 1:12-13 that can enrich cross-curriculum continuity with the Romans package’s own messianic and Davidic-covenant terms.

Recommended actions

1. Lock the divine-name typography convention (“le SEIGNEUR” small caps) in every Phase 2 export pipeline and confirm it survives markdown/HTML conversion before any document is marked complete.
2. Route Haggai 2:7 and 2:23 to mandatory theologian review in full, regardless of surrounding-text routine appearance, per the escalation rules in 12_ai_translation_requirements.md.
3. Attach standing teaching-note glosses to every occurrence of “grand prêtre”/“prêtre,” “mon Esprit”/“l’esprit de,” “sceau,” and “bénédiction” rather than relying on translators to supply ad hoc context each time.
4. Brief all theologian reviewers in advance on the 2:7 grammatical divergence from the familiar English “Desire of All Nations” devotional tradition, so the textually correct rendering is not mistakenly flagged as a doctrinal downgrade.
5. Carry both translation-memory files (Romans baseline + Haggai extension) into every Phase 2 worker to guarantee cross-curriculum term consistency for readers moving between the two.