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

Why it matters

The Matthew curriculum extends the Romans Language Package into markedly higher-risk territory. Where Romans required careful handling of a defined set of doctrinal terms, Matthew’s narrative form — genealogy, parable, exorcism, apocalyptic discourse, and passion narrative — surfaces 29 distinct doctrines and 92 catalogued glossary terms, with 55 of those terms (60%) rated Critical or High risk. Japan’s dual Shinto-Buddhist cultural substrate, its specific 20th-century political-religious history (State Shinto, the 1946 Humanity Declaration), and its recent religious-violence trauma (Aum Shinrikyo, 1995) all intersect directly with core Matthew content — Kingdom language, judgment/end-times language, and spiritual-warfare language — in ways Romans never touched. Getting this right is not optional polish; several of these collisions (Gehenna/地獄, Son of God/現人神, end-of-age/世の終わり) can invert or destroy the doctrine if mishandled.

Key findings

• 29 doctrines identified across all 28 chapters (full-book coverage confirmed, no chapter silently skipped): 15 Critical, 12 High, 2 Medium, 0 Low. 27 doctrines require mandatory human theologian review; 2 (Adoption/Corporate Sonship of Believers; Providence and the Father’s Care) require native speaker review only.
• 92 glossary terms now under enforcement: 34 inherited exactly from the Romans baseline, 58 newly introduced by Matthew. Combined risk profile: 22 Critical, 33 High, 30 Medium, 7 Low — meaning 55 terms (Critical + High) require mandatory theologian-level scrutiny before Phase 2 output is trusted.
• The core passage (Matthew 5:1-12, the Beatitudes) is the single densest concentration of new-term risk in the book — 义/righteousness, 天の御国/kingdom of heaven, 清い/pure in heart (with its 聖い homophone), あわれみ/mercy, 平和をつくる者/peacemaker, and 報い/reward all cluster in twelve verses, several rated Critical.
• Two entirely new risk classes emerge in Matthew that did not exist in Romans: (1) kanji-homophone collisions unique to Japanese orthography (清い/聖い; 赦す/許す; 預言/予言) that can silently invert meaning via IME auto-conversion or unwritten spoken teaching; (2) modern national-trauma resonance (世の終わり/再臨 vs. Aum Shinrikyo’s 1995 sarin attack) requiring historical-contextual framing at every occurrence, not just first use.

Risks

• Highest-severity, no-margin-for-error terms: 天の御国 (kingdom of heaven, State Shinto imperial-descent resonance), 人の子 (Son of Man, under-loaded — reads as merely “a human being”), ゲヘナ vs. 地獄 (must never use the Buddhist-cosmology-laden alternative), 世の終わり (Aum Shinrikyo trauma resonance), 贖い (ransom, no baseline atonement precedent, commercial-flattening risk), 偽善者 (hypocrite, direct collision with the normalized tatemae/honne social convention), 清い/聖い and 赦す/許す (homophone pairs at the Beatitudes and Lord’s Prayer respectively).
• Systemic risk: Matthew’s Critical/High term density (60%) is proportionally higher than Romans, meaning Phase 2 review capacity must scale accordingly or theologian-review bottlenecks are likely.
• Compounding risk: Several Critical terms cluster in the same short passages (e.g., 5:1-12 core passage; 28:16-20 Great Commission), meaning a single review pass must catch multiple simultaneous Critical-tier decisions rather than one at a time.

Opportunities

• Japan’s adult-heir adoption practice (養子縁組) remains a strong positive bridge for adoption/sonship language, inherited unchanged from Romans.
• Genuine, historically grounded teaching bridges exist and should be used deliberately (with required qualification): the Kirishitan-persecution/26-Martyrs memory for the persecution beatitude (5:10-12) and the Great Commission; the Buddhist shukke (renunciation) tradition for costly discipleship; Pure Land Buddhism’s tariki (other-power) structure for salvation-by-grace teaching, exactly as the Romans baseline already models.
• The 建前/本音 (tatemae/honne) social convention, while a genuine collision risk for “hypocrite,” is also a uniquely powerful illustrative tool if used deliberately and correctly: no other target-culture concept surveyed makes Jesus’ point about heart-level righteousness more vivid once properly fenced.

Recommended actions

1. Route all 27 Critical/High-routed doctrines and all 55 Critical/High glossary terms through mandatory human theologian review before Phase 2 output leaves draft status — no exceptions, per assets/doctrine_risk_registry.json and assets/translation_memory.json.
2. Add a dedicated kanji-homophone verification pass (清い/聖い; 赦す/許す; 預言/予言) to the Phase 2 validation pipeline as a standing, risk-tier-independent check, since this failure mode is purely orthographic and can slip past content-level review.
3. Mandate explicit historical-contextual framing text (not optional footnotes) at every occurrence of 世の終わり and 再臨, given the Aum Shinrikyo resonance — treat this as a non-negotiable content requirement, not a stylistic preference.
4. Scale theologian-review capacity for Matthew above Romans-era levels, given the 60% Critical/High term proportion versus Romans’ lower baseline proportion.
5. Treat Matthew 5:1-12, 6:9-13, 16:13-20, 20:28, 26:26-28, and 28:16-20 as fixed, top-priority, verbatim-consistent anchor texts across all curriculum documents, analogous to Romans 1:16-17, 8:28, and 10:9-10 in the baseline package.

---

See analysis/11_doctrine_analysis.md for the full doctrine matrix and chapter-by-chapter coverage confirmation, analysis/08_core_glossary.md for the full term inventory, and analysis/12_ai_translation_requirements.md for the enforceable Phase 2 instruction set.