01 Executive Summary — Matthew 1–28 (English → Bengali)

Why it matters

Bengali Bible translators now have a full-book Language Package for Matthew, built on and never contradicting the existing Romans baseline. Matthew introduces a Christologically denser and narratively fuller text than Romans — a virgin conception, a fully narrated crucifixion and resurrection, an explicit Trinitarian baptismal formula, and Jesus’ own most extended teaching discourse (the Sermon on the Mount, anchored by the core passage, Matthew 5:1-12). Every one of those features sits directly on a live fault line with Bengal’s two dominant religious traditions — Hindu avatar/svarga/karma cosmology in West Bengal and Qur’anic Isa-Christology in Bangladesh — so the translation requirements built here exist specifically to keep the gospel’s hardest claims intact rather than quietly softened by default Bengali vocabulary.

Key findings

• 34 doctrines analyzed across all 28 chapters (no chapter silently skipped): 13 Critical, 13 High, 6 Medium, 2 Low risk — 26 doctrines require mandatory human theologian review, 6 require native speaker review, and only 2 are automated-review-only.
• 111 total glossary terms now govern Phase 2 translation: 44 terms inherited unchanged from the Romans baseline plus 67 new Matthew-specific terms (per assets/translation_memory.json), of which the new-term set alone carries 13 Critical and 20 High risk entries requiring the same theologian-level enforcement precision as the baseline’s Critical/High terms.
• One term is formally unresolved and blocking: “ransom” (λύτρον, Matthew 20:28; 26:28) has no safe Bengali rendering — both candidates either collide with the baseline-forbidden মুক্তি (Hindu rebirth-liberation) root or lose the essential “release from captivity” sense. Phase 2 must not process either verse until a human theologian resolves it.
• The core passage (5:1-12) is lexically well-served but doctrinally load-bearing: its highest risk is not missing vocabulary but crowded-neighborhood collisions already flagged elsewhere in the curriculum (ধন্য vs. ভাগ্যবান; বিশুদ্ধ vs. পবিত্র/শুচি; স্বর্গরাজ্য vs. Hindu svarga).
• The guru-shishya paradigm (শিষ্য) is an unavoidable, non-eliminable risk: no unloaded Bengali word for “disciple” exists, and this word carries maximum doctrinal weight at the Great Commission (28:19), Matthew’s — and the whole curriculum’s — capstone verse.

Risks

• Critical Christological risk concentration at Matthew 1 (virgin conception, Immanuel), 16 (Peter’s confession, church founding), and 28 (resurrection, Great Commission, Trinitarian formula) — three chapters alone carry a majority of the book’s Critical-tier doctrines.
• Sonship confusion risk: ঈশ্বরের পুত্র (Christ’s unique Sonship) and ঈশ্বরের সন্তান (believers’ adoptive sonship) must never be interchanged; a single crossover error at 5:9/5:45 would compound the pre-existing Muslim-background sensitivity around “Son of God” as a rejected biological claim.
• Atonement/sacrifice collision risk: রক্ত (blood) at the Last Supper (26:26-29) and the unresolved ransom term both sit directly against Bengal’s active Kali/Shakta বলি ritual-blood-sacrifice tradition; unglossed renderings risk being read as a repeatable propitiatory offering rather than Christ’s unique, once-for-all sacrifice.
• Do-not-echo landscape risk: Muslim-idiom “Isa Masih” contextualized New Testaments, the Jehovah’s Witnesses Bengali NWT, and Ahmadiyya apologetic literature all circulate in Bengali and actively dispute exactly the claims this curriculum must render without softening (deity of Christ, Trinitarian formula, bodily resurrection after a real death).

Opportunities

• Reuse discipline pays off immediately: 44 of 111 terms require zero new translation decisions — the Romans baseline’s hardest-won terms (ঈশ্বর, যীশু, খ্রীষ্ট, পরিত্রাণ, পুনরুত্থান, ধার্মিকতা, ঈশ্বরের পুত্র) transfer directly into Matthew, giving Phase 2 a large head start on consistency.
• Full-book chapter coverage surfaces reusable doctrine anchors early: the “Son of David” title, the Kingdom of Heaven’s স্বর্গরাজ্য/svarga distinction, and the মনুষ্যপুত্র (Son of Man) Danielic background are each introduced once and then simply cross-referenced dozens of times, reducing per-chapter translation-decision overhead.
• A single theologian decision on “ransom” unlocks two verses at once (20:28 and 26:28), since both share the same underlying term — a small, well-scoped escalation with outsized downstream benefit.
• Mandatory translator notes convert several unavoidable risks into managed risks: the guru-shishya (শিষ্য), bhakti-romantic (প্রেম), and svarga (স্বর্গরাজ্য) collisions cannot be eliminated lexically, but standing translator-note requirements already specified in this package neutralize their worst-case misreadings without requiring new vocabulary invention.

Recommended actions

1. Convene human theologian review immediately for the 26 Critical/High doctrines flagged “Human theologian” in assets/doctrine_risk_registry.json, prioritizing the unresolved ransom term (20:28; 26:28) since it blocks two Phase 2 segments outright.
2. Lock the Sonship-distinction validation rule (ঈশ্বরের পুত্র vs. ঈশ্বরের সন্তান) into automated pre-submission checks before any human review queue, given its recurrence across at least seven passages (1:23; 3:17; 5:9,45; 14:33; 16:16; 17:5; 27:43,54; 28:19).
3. Draft the standing translator notes now (guru-shishya discipleship, bhakti-prem love, svarga-kingdom) as reusable boilerplate for Phase 2, rather than re-deriving them per segment, since 06_linguistic_gap_analysis.md confirms these are permanent, non-eliminable lexical risks rather than one-time footnote items.
4. Reconfirm the YouVersion version ID/abbreviation for the Bengali O.V. edition with the engineering team before hard-coding any MAT.* deep links, per 05_translation_landscape.md Part D.
5. Proceed to Phase 2 for all non-blocked chapters — only segments touching Matthew 20:28 and 26:28 are gated; the remaining 27 chapters and the core passage may begin translation immediately under this Language Package.