01 Executive Summary — Matthew 1–28 Vietnamese Language Package

Phase 1 Translation Requirements Interface (TRI) Output

Curriculum: Matthew 1–28 | Core passage: Matthew 5:1–12 (the Beatitudes) | Destination language: Vietnamese

---

Why it matters

Matthew is the New Testament’s most explicitly Jewish, most kingdom-saturated, and most eschatologically dense Gospel — and Vietnam’s dominant religious landscape (Mahayana Buddhism, Confucian ethics, and thờ cúng tổ tiên ancestor veneration) collides with more of its core vocabulary than almost any other New Testament book translated so far in this pipeline. This Language Package extension gives Phase 2 translators an enforced, doctrine-aware glossary so that terms like “Kingdom of Heaven,” “Son of Man,” “bind and loose,” and “blood of the covenant” render with theological precision instead of drifting toward Buddhist cosmology, folk ancestor-worship, or Confucian filial-duty framing by default.

Key findings

• 87 total glossary terms now govern Matthew translation: 37 inherited verbatim from the Romans baseline, 50 newly introduced by Matthew’s distinctive theology (Kingdom of Heaven, Son of Man, keys/bind-loose, blood of the covenant, ransom, parousia, and more).
• 26 Critical-risk and 23 High-risk terms (49 of 87, ~56%) require non-negotiable enforcement and priority back-translation — a materially higher Critical/High density than the Romans baseline carried, driven by Matthew’s apocalyptic (ch. 24–25), sacramental (ch. 26), and church-authority (ch. 16, 18) material.
• 28 doctrines analyzed across all 28 chapters; per assets/doctrine_risk_registry.json, 14 are Critical and 10 are High risk — 24 of 28 doctrines (86%) require mandatory human theologian review, with only 3 requiring native-speaker-only review and 0 automated-only.
• Three collision risks are structurally unavoidable because they live in ordinary Vietnamese morphemes rather than specialized vocabulary: “thờ” (worship vs. ancestor veneration), “buộc/mở” (church discipline vs. folk binding-charms), and “phước” (Beatitudes’ “blessed” vs. forbidden “phước đức” merit language) — the last of these sits inside the curriculum’s own core passage and cannot be sidestepped by word choice alone.
• Full-book chapter sweep (Steps 1–8 prior outputs) confirms every chapter of Matthew has been reviewed; no chapter was silently skipped, including chapters (14, 17) that contribute no new terminology beyond reinforcing doctrines anchored elsewhere.

Risks

• Highest-severity risk: “Phước cho…” (Blessed are…, Matt 5:3–12) shares its root with the baseline-forbidden “phước đức” (Buddhist merit) and with folk fortune-seeking practice (“cầu phước”); it opens the curriculum’s theological anchor passage and has no alternative Vietnamese Bible-tradition rendering to substitute.
• Sacramental/ecclesial risk cluster: “huyết của sự giao ước” (blood of the covenant, 26:28) risks collapsing into “huyết thệ” (folk blood-oath brotherhood) if the “giao ước” qualifier is ever dropped; “buộc lại…mở ra” (bind/loose, 16:19, 18:18) uses the same ordinary verbs as “bùa chú” folk magical charms.
• Eschatological risk cluster: “sự tận thế,” “sự hình phạt đời đời,” and “sự đến (của Con Người)” all risk assimilation to cyclical Buddhist kalpa/mạt pháp cosmology or a temporary, karma-exhausted địa ngục, inverting Matthew’s linear, final, bodily-return theology.
• Discipleship risk cluster: “từ bỏ mình / mất mạng sống” (deny self/lose life) risks conflation with Buddhist “vô ngã” (no-self); “gươm” (sword, not peace, 10:34) is acutely sensitive against Vietnamese filial duty (đạo hiếu) and could be misheard as endorsing literal violence or family disrespect.
• Cross-curriculum internal risk: “sự công bình” now carries two distinct senses across Romans (declared/forensic) and Matthew (lived/practiced) using the identical Vietnamese term — a consistency risk that is easy to miss precisely because it doesn’t involve any external religious tradition.

Opportunities

• Vietnam’s century of continuous Protestant Bible translation (Bản Dịch Truyền Thống, 1926–present) already supplies settled, non-syncretistic vocabulary for nearly every Matthew concept; the task is disciplined enforcement and mandatory teaching notes, not new coinage — Phase 2 can proceed without inventing theological vocabulary.
• Judaism and Islam offer the closest relational analogues among non-Christian traditions to Matthew’s repentance, mercy, and righteousness vocabulary, useful as bridge-illustrations in accompanying teaching material without importing doctrinal error.
• Vietnamese “vua” (king) and dynastic-kingship history give the Son of David / King of the Jews material unusually strong native resonance, an asset largely absent for other translated curricula.
• The 15-item ranked ambiguity list in analysis/06_linguistic_gap_analysis.md gives Phase 2 QA a concrete, priority-ordered checklist rather than an open-ended risk surface.

Recommended actions

1. Route all 49 Critical/High-risk terms and all 24 Critical/High-risk doctrines to mandatory human theologian review before any Matthew segment is marked approved, per the escalation rules in analysis/12_ai_translation_requirements.md.
2. Enforce the mandatory teaching-note requirement at every occurrence (not just first occurrence) for the three morpheme-level structural risks: thờ phượng, buộc lại/mở ra, and phước cho.
3. Lock the Beatitudes (5:1–12), Peter’s confession (16:16), and the Great Commission (28:18–20) as verbatim-consistent anchor texts across all Phase 2 output, mirroring the Romans baseline’s treatment of Romans 1:16–17, 8:28, and 10:9–10.
4. Carry the “sự công bình” dual-sense teaching note forward into any future cross-curriculum indexing so Romans and Matthew study material never present the term inconsistently to the same learner.

---

This summary synthesizes analysis/04_comparative_theology.md, analysis/06_linguistic_gap_analysis.md, analysis/07_semantic_analysis.md, analysis/08_core_glossary.md, analysis/11_doctrine_analysis.md, assets/translation_memory.json, assets/bible_term_registry.json, and assets/doctrine_risk_registry.json. All cited counts are drawn directly from those artifacts.