Executive Summary: Proverbs (French)

Why it matters

Proverbs is the first non-Greek, non-Christian-exclusive book added to this French translation pipeline. It is Hebrew Wisdom literature shared with Judaism, addressed to “my son” rather than a church, and organized around a single programmatic thesis — the fear of the LORD is the beginning of wisdom — repeated as a three-point inclusio (1:7 / 9:10 / 31:30). Because French already has a fixed, Critical-tier theological vocabulary from the Romans baseline (grâce, justice, Seigneur, loi, gloire), extending the pipeline to Proverbs means actively fencing several Hebrew terms away from those fixed French words, not simply adding new ones. Get the fences wrong, and Proverbs’ teaching will silently corrupt Romans’ doctrine (or vice versa) for any learner moving between the two curricula.

Key findings

• 64 terms are now tracked in the Proverbs translation_memory.json: 15 inherited from the Romans baseline (each with a new Proverbs-specific extension note) and 49 new terms unique to this book, covering all 31 chapters.
• The single highest-stakes decision in the whole curriculum is rendering YHWH as “l’Éternel,” never “le Seigneur” — because the baseline reserves “Seigneur” exclusively, at Critical risk, for Christ’s NT Lordship confession (Romans 10:9). This fence applies to roughly 90 occurrences across the book.
• 17 doctrines were analyzed (vs. the baseline’s Romans doctrine set), with risk distribution: 4 Critical, 5 High, 7 Medium, 1 Low. 9 doctrines require mandatory human theologian review (vs. 12 in the Romans baseline) — a comparable oversight burden despite Proverbs’ shorter, simpler surface prose.
• At the term level, the new translation memory carries 8 Critical-risk and 14 High-risk terms (22 total) requiring theologian sign-off, spanning the Divine Name, the Fear-of-the-Lord phrase, Wisdom Personified (especially the Arian-controversy proof text at 8:22), the corporal-discipline “rod” cluster, and the chen/chesed-vs-grâce fence.
• France-specific legal exposure is new to this curriculum: the 2019 French law banning “violences éducatives ordinaires” makes any literal rendering of Proverbs’ child-discipline “rod” texts (3:11-12; 13:24; 19:18; 22:15; 23:13-14; 29:15,17) pastorally and potentially legally fraught, requiring a standing figurative-rendering strategy with a mandatory translator’s note at every occurrence.
• French lacks Hebrew’s five-way fool taxonomy (kesil/eviyl/letz/peti/naval); a fixed, non-interchangeable five-term French assignment (sot/insensé/moqueur/le simple/insensé impie) is now mandated as a system-level fence, since any individual improvisation collapses a doctrinally significant graduated moral distinction.

Risks

• Divine Name collapse: any lapsed occurrence of “le Seigneur” for YHWH silently reopens the Lordship-title collision the entire Romans Christological architecture was built to prevent.
• Grace bleed-through: rendering chen or chesed as “grâce” would import Romans’ forensic, Trent-vs-Reformed grace controversy into an unrelated OT relational term.
• Cross-curriculum “justice” ambiguity: because “justice” is mandatorily reused for Proverbs’ tzedeq, every lesson spanning both curricula carries a permanent, structural risk of readers conflating this-worldly ethical conduct with Romans’ forensic righteousness by faith unless explicitly disambiguated.
• Arian-controversy reopening: a careless “m’a créée” for Proverbs 8:22’s qanani would reintroduce a fourth-century Christological heresy into contemporary catechesis.
• Legal/pastoral misfire: literal “rod” translations could be read as endorsing conduct now illegal in France, damaging the curriculum’s credibility with mainstream French readers.
• Marian/typological over-reach: French Catholic devotional culture’s Sedes Sapientiae tradition creates a live, real-world pull toward reading Proverbs 8–9’s personified Wisdom as a direct Marian or unqualified Christological text.

Opportunities

• Providence continuity: Proverbs’ repeated providence cluster (3:5-6; 16:1-9; 19:21; 21:1,30-31) reinforces the Romans baseline’s Romans 8:28 providence doctrine almost without divergence — a strong, ready-made cross-curriculum teaching bridge.
• Interfaith bridge points: speech ethics, diligence, economic generosity, and humility show unusually wide convergence across Jewish, Islamic, Buddhist, and Hindu ethical traditions, making these doctrines low-risk, high-value points for ecumenical or interfaith-aware teaching material.
• Deliberate correction as a teaching moment: departures from older French Bible convention (eshet chayil as “femme de valeur” rather than “femme vertueuse”; chen as “faveur” rather than “grâce”) are lexically well-grounded and can be presented to reviewers as accuracy corrections, strengthening the curriculum’s credibility with both Catholic and Protestant audiences.
• Jewish rabbinic precedent: existing Talmudic tradition already limits literal readings of the “rod” texts, offering a valuable, citable precedent for this curriculum’s own figurative-rendering strategy.

Recommended actions

1. Lock the Divine Name fence (“l’Éternel,” never “Seigneur”) as a non-negotiable, curriculum-wide validation check before any other Phase 2 work begins.
2. Route all 9 Critical/High-tier doctrines and all 22 Critical/High-tier terms through mandatory human theologian review per analysis/12_ai_translation_requirements.md’s escalation rules — do not allow automated or native-speaker-only sign-off for these.
3. Adopt the fixed five-term fool-taxonomy assignment as a single system-level decision recorded once, not re-derived per chapter or per translator.
4. Publish the standing translator’s note for the corporal-discipline cluster once, and require every occurrence (six passages) to cite it rather than improvise new language.
5. Brief all reviewers on the chen/chesed/grâce fence and the tzedeq/justice cross-curriculum disambiguation before Phase 2 translation begins, since both are silent-failure risks (the AI could produce fluent, natural-sounding French that is nonetheless doctrinally wrong).