Executive Summary: Matthew 1–28 Language Package (English → Korean)

Why it matters

Korean is the best-prepared language in this pipeline for a Bible curriculum: 130+ years of indigenous Bible translation history means almost every Matthew term already has settled, community-recognized vocabulary. But that maturity creates its own risk profile — the danger in Korean is rarely “no word exists,” it’s “a perfectly natural word already means something else.” Matthew intensifies three specific, well-documented Korean syncretism pressures beyond what Romans surfaced: (1) Buddhist-cosmology collision on judgment/afterlife vocabulary (지옥, 천국), absent from Romans entirely; (2) a Confucian homophone trap on Jesus’ own signature self-title (인자); and (3) live, contemporary Korean new-religious-movement claims that directly contest Matthew’s own second-coming warnings (24:23-26). This curriculum’s core passage, the Beatitudes (5:1-12), sits at the highest concentration point of Korea’s central risk term — 복이 있는 자 (“blessed”) uses the same 복 root the baseline already flags as Critical for gospel/grace, meaning the anchor passage itself requires the pipeline’s most careful handling.

Key findings

• Full-book coverage confirmed: all 28 chapters reviewed; every chapter maps to at least one of 36 registry doctrines; none required a “no content” placeholder (chapters 12, 22, 24, 26–27 reinforce rather than introduce doctrine, per 11_doctrine_analysis.md).
• Doctrine-level risk (from assets/doctrine_risk_registry.json): 13 Critical-risk doctrines, 17 High-risk doctrines, 5 Medium, 0 Low — 30 of 36 doctrines (83%) require mandatory human theologian review, versus 5 requiring native-speaker review and 0 automated-only. This is a markedly higher theologian-review burden than the Romans baseline.
• Term-level risk (from assets/bible_term_registry.json, Matthew-specific term set): approximately 20 Critical-risk terms and 44 High-risk terms newly introduced or elevated by Matthew, on top of the 46 terms inherited unchanged from the Romans baseline TM.
• New collision category not present in Romans: Buddhist-origin loanwords now carry direct doctrinal weight — 지옥 (Gehenna) is itself a Buddhist term for a temporary rebirth-realm punishment, colliding head-on with Matthew’s final, eternal judgment; 천국 (kingdom of heaven) has drifted toward the Buddhist Pure Land concept 극락 in everyday usage.
• Highest silent-misreading risk in the entire package: 인자 (“Son of Man,” Jesus’ ~30-occurrence self-title) is a perfect, unmarked Hangul homophone of 인자 (仁者, “a benevolent person,” a Confucian virtue-term) — a reader can process every occurrence without ever detecting the substitution.

Risks

• Critical: 지옥 (Gehenna/temporary-rebirth collision), 양과 염소 (Sheep and Goats read as works-salvation), 영벌 (eternal punishment softened to purgatorial), 변형되다 (Transfiguration read as shape-shifting/화신), 대속물 and 성찬 (ransom/Lord’s Supper read as shamanistic appeasement or ancestral food-offering), 죄를 사하는 권세 (authority to forgive sins softened), 성령을 모독하는 죄 (blasphemy against the Spirit mishandled pastorally), 아버지와 아들과 성령의 이름으로 (Trinitarian formula), and the elevated-in-Matthew Critical application of 의 (righteousness) at 5:20.
• High: 인자 homophone collision; 매다/풀다 (binding/loosing read as shamanistic ritual); 재림 vs. living-messiah claims of Korean new-religious movements; 천국 vs. 하나님의 나라 “two kingdoms” misreading; 멍에 (yoke’s oppressive colonial connotation working against the verse’s intended comfort); 상 (reward) sliding toward 기복신앙 or achievement-culture merit; 다윗의 자손 vs. 다윗의 혈통 conflation; 청함을 받은 자 vs. Romans’ 부르심 cross-document confusion; 유전 (tradition of the elders) touching the historically sensitive jesa/ancestor-rite martyrdom memory; 하나님의 아들 applied to Christ vs. believers without distinction.
• Structural risk: 83% theologian-review doctrine load will strain Phase 2 review-routing capacity if not proportionally staffed above Romans-curriculum levels.

Opportunities

• Korean’s settled Bible-translation tradition means almost no new coinage is needed — Phase 2 can lean on established KRV vocabulary and focus effort on teaching-note fencing rather than lexical invention, a lighter lift than most pipeline languages face.
• Korean already cleanly distinguishes ἐξουσία/권세 from δύναμις/능력 and 언약/covenant from 계약/contract — genuine structural strengths to be preserved and exploited in teaching material, not re-litigated.
• The vineyard-laborers parable (20:1-16) and Beatitudes’ reward language (5:12) are ideal built-in teaching passages for directly confronting the 기복신앙 transactional-blessing risk the baseline already identifies as Korea’s single most important internal church self-critique.
• 아빠/Abba’s near-cognate fit (already validated in Romans) and 11:28-30’s yoke/rest passage offer strong opportunities for warm, relationally resonant teaching once the yoke idiom’s negative connotation is directly named and corrected.

Recommended actions

1. Load assets/translation_memory.json (Matthew version, this package) as the sole enforced glossary for Phase 2 — it is a complete superset of the Romans baseline TM plus all new Matthew terms.
2. Route all 30 Critical/High doctrines and their associated Critical/High terms to mandatory human theologian review per the escalation rules in analysis/12_ai_translation_requirements.md; do not permit native-speaker-only sign-off on any of the 13 Critical doctrines.
3. Apply the mandatory, no-exception disambiguation note at every occurrence of 지옥 (Gehenna) and at every first occurrence of 인자 (Son of Man) per lesson — these are the two highest silent-misreading risks identified in 06_linguistic_gap_analysis.md.
4. Enforce the 의/의리 distinguishing footnote at every Sermon-on-the-Mount and chapter-23 occurrence of righteousness, not only in justification contexts, given Matthew 5:20’s programmatic role.
5. Explicitly teach 천국 = 하나님의 나라 (same referent) and 그리스도의 유일한 아들 되심 ≠ 신자의 양자 됨 (Christ’s unique Sonship ≠ believers’ derivative sonship) as standing cross-reference notes wherever either recurs.
6. Staff Phase 2 review capacity for Matthew above Romans-curriculum baseline levels, given the 83% theologian-review doctrine load documented above.