Executive Summary — Hebrews (English → Thai)

Why it matters

Hebrews is the most theologically concentrated curriculum this pipeline has processed for Thai. Its core passage (9:11-28) — Christ’s once-for-all high-priestly sacrifice — sits directly on top of Thailand’s dominant religious grammar: repeatable merit-making (บุญ/กรรมดี), a royal-Brahmin priesthood (ปุโรหิตหลวง), and Buddhist technical vocabulary for enlightenment (ตรัสรู้) and final rest (นิพพาน). Get the vocabulary right here and Hebrews lands as intended: a categorical, once-for-all rescue by a personal God. Get it wrong, even in one high-frequency term, and the letter’s central argument silently inverts into “a superior ritual within the merit system it was written to refute.”

Key findings

• 18 doctrines identified across all 13 chapters (full-book coverage per PRD mandate); 8 Critical + 8 High = 16 of 18 require mandatory human theologian review, with only 2 Medium-tier doctrines routed to native-speaker review and zero doctrines eligible for automated-only review — a first for this language pair, reflecting how densely Hebrews concentrates high-stakes Christological and soteriological argument compared to Romans/Galatians.
• 91 new Hebrews-specific terms cataloged (14 Critical, 34 High, 30 Medium, 13 Low), plus 23 baseline terms inherited exactly from the Romans/Galatians package (8 Critical, 10 High, 5 Medium) and 20 proper names requiring established or newly-fixed transliterations.
• Five collision points carry the highest combined severity and probability of error: (1) ἀρχιερεύς/ἱερεύς (“high priest”/“priest”) vs. the Thai royal court’s Brahmin priesthood (ปุโรหิตหลวง) — recurs in nearly every chapter; (2) τέλειος/τελειόω (“perfect/made perfect”) vs. บารมี (accumulated merit across lifetimes) — the single highest-frequency Critical term, occurring at least nine times; (3) ἄγγελοι (“angels”) vs. เทวดา (Buddhist-Brahmanic devas) — foundational to the entire chapter 1 argument; (4) διαθήκη’s covenant/testament wordplay at 9:16-17, a structural linguistic gap Thai cannot carry without added bridging wording; (5) φωτισθέντας (“enlightened”) vs. รู้แจ้ง/ตรัสรู้ — the Buddha’s own technical attainment vocabulary.
• One unresolved cross-curriculum coordination gap: ἱλαστήριον (“mercy seat,” Hebrews 9:5) is the same Greek word rendered “propitiation” at Romans 3:25, but the Romans baseline has no explicit translation_memory entry reconciling it. This must be fixed once, jointly, before either passage is translated in Phase 2 — otherwise the pipeline will produce two inconsistent Thai renderings of one Greek term.

Risks

• Highest risk (Critical, core-passage-specific): “Once for all” (ครั้งเดียวเป็นพอ) and “to die once” (ตายเพียงครั้งเดียว), both inside Hebrews 9:11-28, directly contradict Thailand’s two most deeply embedded religious intuitions — repeatable merit-accrual and the death-and-rebirth cycle. A single softened rendering or an unguarded surrounding paraphrase can silently reverse the passage’s entire argument without triggering an obvious translation error in ordinary QA.
• High cumulative-exposure risk: the ปุโรหิต/มหาปุโรหิต royal-Brahmin collision recurs across nearly every chapter (2, 3, 4, 5, 6, 7, 8, 9, 10), meaning a single missed translator-training note creates a standing ambiguity for the whole book, not an isolated verse.
• Register risk: access-to-God vocabulary (เข้าเฝ้า, “draw near”) is structurally opposite in connotation (formal court distance/protocol) to the boldness (parrhēsia) Hebrews explicitly pairs with it — a risk that can pass unnoticed because both individual words are “correct” in isolation.
• Cross-curriculum risk: the ἱλαστήριον/propitiation gap with Romans 3:25 (see Key Findings) is a structural, not merely lexical, exposure that will surface in any lesson cross-referencing both books.

Opportunities

• Hebrews’ extensive royal-court and throne imagery (throne of grace, drawing near, access through blood) can be taught as a positive bridge for Thai audiences already fluent in honor/audience/hierarchy categories — provided every occurrence is anchored to Hebrews’ distinctive boldness (parrhēsia) so it does not collapse into ordinary court protocol.
• The letter’s five warning passages and its “faith hall of fame” (ch. 11) give strong material for native-speaker-reviewed pastoral application lessons once the Critical-tier vocabulary (enlightened, rest, reward) is locked down, since these chapters’ doctrinal terms are High rather than Critical risk once the核心 vocabulary fence is in place.
• Because Hebrews reuses 23 baseline terms exactly, Phase 2 translator onboarding can lean heavily on existing Romans/Galatians training material, concentrating new calibration effort on the ~10-15 highest-priority Hebrews-specific terms identified above rather than retraining from zero.

Recommended actions

1. Resolve the ἱλαστήριον/Romans 3:25 cross-curriculum gap before any Phase 2 work begins on either Hebrews 9:5 or a future Romans 3:25 segment. This is a blocking dependency, not a routine glossary decision.
2. Treat Hebrews 9:11-28 as the mandatory Phase 2 translator-calibration passage, since it concentrates nearly every Critical risk category in the curriculum (once-for-all, die-once, covenant/testament wordplay, high priest, mercy seat, blood of Christ, conscience-cleansing).
3. Institute a standing (not one-time) translator note requirement for ปุโรหิต/มหาปุโรหิต and for τέλειος/τελειόω, flagged at every occurrence given their unusually high recurrence across the book, rather than relying on a first-occurrence footnote alone.
4. Route 100% of segments in the five warning passages, all of Hebrews 1, and all of 9:11-10:18 to mandatory human theologian review with no sampling, consistent with the escalation rules in 12_ai_translation_requirements.md.
5. Reuse Romans/Galatians-trained translators where possible, supplementing with focused calibration on the five highest-severity Thai-specific collision points identified above, to accelerate Phase 2 onboarding without re-deriving baseline competency.

---

See assets/translation_memory.json, assets/bible_term_registry.json, assets/doctrine_risk_registry.json, analysis/11_doctrine_analysis.md, and analysis/12_ai_translation_requirements.md for full supporting detail.