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

The bottom line: The Matthew curriculum is translatable into doctrinally faithful, culturally aware Dutch using the same HSV/NBV21-anchored register established by the Romans baseline — but Matthew carries substantially higher doctrinal density and denominational fault-line exposure than Romans did, requiring more theologian review capacity, not less, before Phase 2 begins.

---

Why it matters

Matthew is the most quoted, most liturgically central, and most confessionally contested Gospel in the Dutch church landscape — its “Zoon van David,” “Koninkrijk der hemelen,” “deze rots,” and “zalig” sit directly on top of the Netherlands’ own Reformed/Catholic geographic divide, its 1618-19 Synod of Dordrecht heritage, and its rapidly secularizing vocabulary. A single mistranslated word here (e.g., “gelukkig” for “zalig,” “heidenen” for the Great Commission’s “volken,” or “kerk” for “gemeente” at Matthew 16:18) does not merely reduce clarity — it can invert the doctrine the passage exists to teach.

Key findings

• Doctrinal density is more than double Romans’: 32 doctrines identified, of which 17 are Critical and 8 are High risk (25 of 32, or 78%, require mandatory human theologian review); only 7 require native-speaker-only review, and zero are automated-only. The Romans baseline, by comparison, required theologian review for 15 of 29 doctrines.
• Term-level risk is comparably concentrated: across the full assets/translation_memory.json (baseline-inherited + new Matthew terms), approximately 61 individual terms carry Critical or High doctrine_risk and require mandatory theologian-level back-translation checks — roughly 14 Critical, 47 High across baseline-inherited and newly introduced vocabulary combined.
• The core passage (5:1-12) is the single highest lexical-risk zone in the whole book: “zalig” (blessed) has drifted colloquially toward “delicious/blissful,” an almost exact parallel to the baseline’s flagged “zonde” drift, but landing directly on the curriculum’s theological anchor text.
• Four new, Matthew-specific denominational fault lines were identified beyond anything in the Romans baseline: (1) the rock/keys/binding-and-loosing (16:18-19; 18:18) — the sharpest Catholic/Protestant interpretive divide in the whole Gospel; (2) the paedobaptist/credobaptist baptism divide (3:6-17; 28:19); (3) the Reformation-era Eucharistic controversy embedded in the Lord’s Supper institution (26:26-28); (4) a growing Dutch mainline (PKN) universalist/annihilationist-sympathetic minority contesting 25:46’s eternal punishment.
• A distinctly Dutch Jehovah’s Witnesses (Nieuwe-Wereldvertaling) collision risk profile was newly catalogued, covering five specific renderings (“martelpaal” for cross, “Jehovah” restored into embedded OT quotations, “eeuwige afsnijding” for eternal punishment, “aanwezigheid” for the second coming, and a diminished “Zoon van God”) that must be actively and explicitly excluded from this curriculum’s vocabulary and framing.
• Full-book coverage is confirmed: all 28 chapters were swept for load-bearing vocabulary; chapters contributing no new terms (9, 14, 17, 19) are explicitly documented as reviewed, not silently skipped, per the PRD’s full-coverage mandate.

Risks

• Highest risk: “Zalig” (blessed) misread as “delightful” in the core passage — mandatory theologian review at every occurrence, no exceptions.
• Highest risk: The four-way Christological “son” cluster (Zoon van God / Zoon van David / Mensenzoon / kinderen van God) — no other doctrine in this curriculum has this many closely-spaced, easily-confused, doctrinally load-bearing filial titles; cross-confusion would corrupt Christology at its foundation.
• High risk: “Loon” (reward), the ordinary Dutch wage-word, appearing five separate times across the book (5:12,46; 6:1-18; 10:41-42; 20:8; 25:14-30) in direct tension with the baseline’s Critical Grace doctrine.
• High risk: “Heidenen” vs. “volken” — a single word-choice error at Matthew 28:19, 24:14, or 21:43 would silently narrow or ethnically re-color the entire Great Commission.
• High risk: Baptismal and Eucharistic wording (3:6-17; 28:19; 26:26-28) sitting directly atop live, unresolved contemporary Dutch ecclesial divides — these must be translated with denominational neutrality, not resolved by translator preference.

Opportunities

• Dutch’s existing dual-anchor translation ecosystem (HSV + NBV21) is well-suited to Matthew’s specific risk profile, since the formal-equivalence/dynamic-equivalence split maps almost exactly onto Matthew’s own Critical-term fault lines (e.g., HSV’s “zalig”/“de boze”/“gemeente” vs. NBV21’s “gelukkig”/“het kwaad”/occasional softer phrasing) — this curriculum can use that existing split as a built-in comparative check rather than building one from scratch.
• Dutch’s clean lexical distinction between “voorbede” and Catholic “voorspraak” (inherited from the baseline) and its dedicated “zending” vocabulary continue to reduce ambiguity relative to other language pairs in this batch, lowering translation risk for the Great Commission and prayer-related doctrines specifically.
• The Heidelberg Catechism’s Lord’s Day 31 (Q&A 83-85) gives Dutch Reformed-catechized readers unusually precise prior vocabulary for “the keys of the kingdom” — this can be leveraged in teaching material as a built-in doctrinal anchor rather than treated purely as a source of Catholic/Protestant risk.

Recommended actions

1. Load assets/translation_memory.json, assets/bible_term_registry.json, and assets/doctrine_risk_registry.json (Matthew) together with the baseline Romans equivalents before any Phase 2 segment translation begins — Matthew-specific risk-tier elevations documented in these files (e.g., church, lord, resurrection, incarnation, jesus, god, holy_spirit all elevated versus their Romans-baseline tier) govern review routing for this curriculum.
2. Budget theologian review capacity at roughly double the per-chapter rate used for Romans, concentrated on chapters 1, 5, 12, 16, 18, 20, 22, 24–26, and 28.
3. Apply the Matthew-specific forbidden-substitutions list in analysis/12_ai_translation_requirements.md at validation time for every segment, with particular priority on the five Jehovah’s Witnesses Nieuwe-Wereldvertaling collision points and the “zalig”/“gelukkig” core-passage flashpoint.
4. Treat the rock/keys/binding-and-loosing cluster (16:18-19; 18:18), the baptismal command (28:19), and the Lord’s Supper institution (26:26-28) as standing denominational-neutrality checkpoints in every editorial pass, not one-time decisions — confirm at each review cycle that no single tradition’s sacramental or ecclesiological theology has been silently imported into the wording.
5. Carry the YouVersion HSV (1394) / NBV21 (1968) version-ID recommendation forward as the standing Dutch-language citation standard for this and all subsequent curricula in this language pair, per analysis/05_translation_landscape.md.

---

This summary synthesizes analysis/04_comparative_theology.md, analysis/05_translation_landscape.md, analysis/06_linguistic_gap_analysis.md, analysis/07_semantic_analysis.md, analysis/08_core_glossary.md, analysis/11_doctrine_analysis.md, and the three Phase 1 asset files. It must be read alongside those documents, not in place of them, before Phase 2 translation work begins.