Executive Summary: Matthew (English → Thai)

Why it matters

Matthew is the entry point to the New Testament for most readers and the most doctrinally dense book yet processed in this Thai Language Package family — denser than Romans or Galatians in raw Critical-risk surface area, because it fuses four convergent Christological titles (Son of David, Son of God, Son of Man, Christ/Messiah), Matthew’s own 32-occurrence organizing category (“kingdom of heaven”), and a sustained, book-length confrontation with Thailand’s public merit-making (บุญ) culture in the Sermon on the Mount. The core passage, Matthew 5:1-12, sits at the intersection of three of the highest-risk doctrine clusters in the whole book — kingdom of heaven, righteousness, and reward — making it a uniquely demanding anchor text, not a narrow slice of the translation problem.

Key findings

• 26 doctrines identified, extending the Romans/Galatines baseline: 15 Critical, 7 High, 2 Medium, 2 Low — matching doctrine_risk_registry.json exactly. 22 of 26 doctrines require mandatory human theologian review; only 2 route to native-speaker-only review and 2 to automated review.
• 96 terms are recorded in translation_memory.json: 27 inherited unchanged from the Romans/Galatians baseline and 69 new to Matthew. Of the new and inherited terms combined in bible_term_registry.json, 23 are Critical risk and 35 are High risk — 58 terms requiring mandatory theologian review, more than double the combined Critical+High term count carried by the Romans baseline alone.
• Three collision clusters recur across the entire book, not isolated verses: (1) สวรรค์ (heaven) vocabulary against Buddhist temporary heavenly rebirth-realms — kingdom of heaven, treasure/reward in heaven, keys of the kingdom, mysteries of the kingdom; (2) บุญ/merit vocabulary against Matthew’s own direct critique of public religious display (ch.5-6, 19-20, 25); (3) อวตาร/rebirth vocabulary against incarnation, transfiguration, resurrection, and the Son of Man’s second coming (ch.1, 3, 12, 17, 24, 27-28).
• Two title-distinctness pairs are uniquely load-bearing to Matthew: Son of Man (บุตรมนุษย์) vs. Son of God (พระบุตรของพระเจ้า) — two titles for one person that must never collapse into each other — and the believers’ plain บุตร (5:9, 45) vs. Christ’s own พระบุตร, a distinction absent from the Romans/Galatians baseline entirely.

Risks

• Highest combined severity × frequency: “Kingdom of heaven” (32 occurrences) risks being heard as a desirable but impermanent post-mortem tier reached by accumulated merit rather than God’s present-and-coming sovereign reign — the single most pervasive doctrinal exposure in the book.
• Highest single-verse stakes: Blasphemy against the Holy Spirit (12:31-32) and the cry of dereliction (27:46) both carry maximal doctrinal consequence from a single mistranslation or unexplained literal rendering.
• Highest culturally-specific confrontation: Matthew 6’s almsgiving/fasting/prayer teaching does not merely resemble Thailand’s public dāna culture — it directly names and rejects that pattern’s logic, meaning a mistranslation here doesn’t just confuse a doctrine, it can invert Jesus’ own argument into an endorsement of the very practice he condemns.
• Highest structural-consistency risk: the Great Commission’s Trinitarian baptismal formula (28:19) depends on explicit surrounding teaching, not lexical marking, since Thai grammar cannot itself distinguish “one name, three persons” from “three separate names.”

Opportunities

• Genuine structural bridges exist and should be used carefully in translator training: the rabbinic “bind and loose” idiom (Jewish tradition), voluntary total devotion in guru-disciple relationships (Hindu/Buddhist and Jewish rabbi-disciple traditions), and Islamic/Jewish shared belief in a final judgment and bodily resurrection all offer genuine, if partial, common ground — never equivalence — for teaching Matthew’s more difficult claims.
• Thailand’s existing formal-equivalence THSV11 tradition already supplies settled, non-ambiguous renderings for most shared baseline vocabulary, meaning Phase 2 effort concentrates on fencing existing fluent Thai words against Buddhist/folk misreadings rather than coining new vocabulary from scratch.
• Matthew’s own text frequently makes the needed doctrinal distinction explicit within the passage itself (e.g., the workers-in-the-vineyard parable at 20:1-16 for reward-not-merit; the eunuchs-for-the-kingdom aside at 19:12 for celibacy-not-devaluation), giving translator notes strong textual anchors rather than requiring external argumentation.

Recommended actions

1. Load both the baseline Romans/Galatians and the new Matthew translation_memory.json and bible_term_registry.json for every Phase 2 segment; never allow a Matthew-specific rendering to diverge from an already-established baseline term for the same underlying concept.
2. Route all 22 Critical/High Matthew doctrines and all 58 Critical/High terms to mandatory human theologian review per the escalation rules in 12_ai_translation_requirements.md; treat the core passage (5:1-12) and the closing commission (28:16-20) as standing, always-escalated segments regardless of individual sub-term risk tier.
3. Brief theologian reviewers on the three cross-cutting collision clusters (สวรรค์, บุญ, อวตาร/rebirth) as a single training module, since a translator who avoids one collision within a cluster but reaches for another within the same cluster has not actually reduced risk.
4. Enforce the title-distinctness validation rule (Son of Man vs. Son of God vs. Son of David vs. believers’ plain sonship) as a standing, automated pre-acceptance check across all 28 chapters, not merely a one-time note at first occurrence.