Executive Summary — Matthew 1–28 (English → French)

Curriculum: Matthew · Core passage: Matthew 5:1–12 (the Beatitudes) · Language pair: English → French

Why it matters

Matthew is the largest and most doctrinally dense curriculum built on this French Language Package to date. It introduces 67 new theological terms beyond the 38 baseline Romans terms it reuses, and — because it is narrative rather than epistolary — it carries doctrinal weight in recurring titles, formulas, and idioms (Fils de l’homme, Royaume des cieux, clés du Royaume, “Ceci est mon corps”) that cannot be resolved by a single glossary choice the way a Romans doctrinal noun often can. France’s mixed Catholic-majority, Protestant-minority, and strongly secular laïque readership means nearly every high-stakes term collides with an existing French institutional, devotional, or idiomatic frame that this package must actively fence rather than ignore.

Key findings

• 105 tracked terms in the Matthew translation memory: 22 Critical, 31 High, 42 Medium, 10 Low — requiring theologian review for 53 terms (50%), native speaker review for 42 (40%), automated review for only 10 (10%).
• 48 doctrines in the Matthew doctrine risk registry: 11 Critical, 18 High, 15 Medium, 4 Low — 29 doctrines (60%) require mandatory human theologian review, the highest proportion of any curriculum built on this package so far.
• The single most consequential ambiguity identified across all Phase 1 analysis is the Fils de l’homme / Fils de Dieu conflation risk: French readers routinely and plausibly misread “Fils de l’homme” as emphasizing only Jesus’ humility, missing its Daniel 7 authority-and-judgment claim, and risk blurring it with the unrelated Critical title “Fils de Dieu.”
• Thirteen of twenty-one identified “crowded semantic neighborhood” collisions are with French Catholic institutional or devotional culture specifically (Église, les saints, lier/délier, clés du Royaume, la Tradition, mystère, repentir, miséricorde, dettes) — confirming and sharpening the pattern the Romans baseline first documented.
• Three verified Jehovah’s Witnesses (Traduction du monde nouveau) sect-renderings are in wide French circulation and must be explicitly excluded: “rendirent hommage” (not “adorer”) at the worship climaxes; “retranchement éternel” (not “châtiment éternel”) at 25:46; lower-case impersonal “esprit saint.”
• Genuinely missing French vocabulary (not just crowded neighborhoods) exists for exactly two concepts: the imperative “make disciples” (μαθητεύσατε — no single French verb exists) and σκάνδαλον’s original interpersonal-sin sense (fully displaced by modern “scandale” = media controversy).

Risks

• Critical-tier concentration in chapters 1, 16, 24–26, and 28 (Incarnation, Sonship, Church-founding, Judgment/Parousia, Words of Institution, Trinitarian formula, Great Commission) creates a reviewer-bandwidth bottleneck if not proactively queued ahead of lower-risk chapters.
• Two doctrinally load-bearing but publicly memorized French texts — the Lord’s Prayer’s “dettes” vs. the liturgical “offenses,” and the Beatitudes’ “heureux” vs. the sound-alike but doctrinally loaded Catholic beatification title “Bienheureux” — create a live risk of a technically correct translation reading as wrong to readers measuring it against what they already pray or know.
• France’s colonial missionary history and transatlantic/colonial slave-trade memory (2001 Taubira law) make the Great Commission (“faites… des disciples de toutes les nations”) and the servant/slave metaphor (20:26-28, “esclave”) uniquely sensitive in a French destination context relative to other target languages in this package family.
• Etymological contamination uniquely afflicts this book: modern French “avoir du talent” and “l’avènement d’une nouvelle ère” derive from or resemble the very texts (the parable of the talents; the parousia) being translated, actively working against correct comprehension rather than merely being neutral or absent vocabulary.

Opportunities

• French uniquely preserves the Pierre/pierre wordplay at Matthew 16:18, an asset not available in every target language, worth highlighting in teaching material even though the surrounding papal-primacy debate must remain unresolved in the base text.
• The 2017 French ecumenical liturgical revision (Lord’s Prayer’s “ne nous laisse pas entrer en tentation,” “le Mauvais”) gives this curriculum a ready-made, already-memorized consensus wording to align with rather than compete against, reducing reader friction at one of the book’s highest-traffic passages.
• The established Amen/Alléluia/Abba/Hosanna transliteration convention from the Romans baseline extends cleanly to Matthew’s own in-text-glossed terms (Emmanuel, Éli Éli lema sabachthani), giving Phase 2 a stable, low-risk pattern to apply without new design work.
• Segond 21 (S21) is confirmed as the recommended base YouVersion reading version, satisfying both the baseline’s anti-archaism register requirement and this curriculum’s need to avoid the two documented sect-adjacent renderings (worship, eternal punishment) at their exact points of doctrinal sensitivity.

Recommended actions

1. Lock assets/translation_memory.json (Matthew edition, this version) and analysis/12_ai_translation_requirements.md as the mandatory Phase 2 pre-flight load before any segment translation begins.
2. Queue Matthew 1, 16, 24–26, and 28 for priority theologian review scheduling given their Critical-tier density, rather than processing chapters in simple sequential order.
3. Confirm the Segond 21 numeric YouVersion version_id programmatically before generating any hyperlinked cross-reference in learner-facing material (see analysis/05_translation_landscape.md Section 6).
4. Build the mandatory OT/background cross-reference footnote apparatus (fulfillment formulas, Fils de David, Fils de l’homme, l’abomination de la désolation, clés du Royaume) into the Phase 2 templating layer now, since this is a structural requirement, not a per-segment lexical choice.
5. Brief all Phase 2 reviewers explicitly on the three documented Traduction du monde nouveau sect-renderings before work begins, since their surface formality makes them the highest-probability accidental-echo risk in this specific language pair.

---

See analysis/11_doctrine_analysis.md, assets/doctrine_risk_registry.json, assets/translation_memory.json, and assets/bible_term_registry.json for the full data underlying this summary.