01 — Executive Summary: Haggai Translation Requirements (Mandarin)

Why it matters

Haggai is short — 2 chapters, 38 verses — but it is doctrinally dense and sits on some of the sharpest term-collision terrain in the entire Mandarin Bible-translation landscape. The book’s central rebuke (“my house lies in ruins while you each busy yourselves with your own house,” 1:9) is the curriculum’s core doctrine, and its resolution (“from this day I will bless you,” 2:19) uses the single character most exposed to Chinese folk fortune-seeking vocabulary (福) anywhere in this Language Package. Getting Haggai’s Mandarin rendering right protects four load-bearing doctrines — God’s House vs. Self-Interest, Obedience and Blessing, the Presence and Spirit of God, and the Coming Glory of the Temple — from being quietly re-read through Confucian ancestral-hall duty, Buddhist merit-accounting, folk drought-ritual, or Mandate-of-Heaven political-legitimacy frames.

Key findings

• 44 Haggai-specific terms documented, 38 newly introduced by this book and 20 inherited exactly from the baseline Romans package (some terms overlap both lists structurally, e.g. election/glory/peace, which are reused baseline entries with Haggai-specific context notes).
• 7 Critical-risk terms and 14 High-risk terms — 21 of 44 terms (48%) require human theologian review at every occurrence, more than double the density of Critical/High terms in the baseline Romans registry proportionally, driven by Haggai’s short length concentrating high-stakes vocabulary (耶和华, 万军之耶和华, 殿, 我的灵, 万国所羡慕的, 赐福 are all Critical).
• 4 Critical-risk doctrines (Priorities: God’s House vs. Self-Interest; Obedience and Blessing; the Presence and Spirit of God with His People; the Coming Glory of the Temple) and 2 High-risk doctrines (Davidic Covenant and Messianic Hope; Divine Sovereignty over Nations and Creation) — 6 of 9 total doctrines (67%) require human theologian review, with the remaining 3 requiring native speaker review. Zero doctrines are automated-review-only, unlike the baseline Romans registry’s 4 automated-only doctrines.
• The single highest-stakes translation decision in the book is 赐福 (Haggai 2:19), which shares its root character with the already-flagged baseline gospel collision (福音/祈福) but now anchors the curriculum’s own doctrinal climax rather than merely sitting adjacent to it.
• A genuine, unresolved textual ambiguity (2:7, “desire of all nations” / 万国所羡慕的) cannot be silently settled by Phase 2 automation and must be flagged for theologian review at every occurrence.

Risks

• 福-root collision (Critical): 赐福/祝福 in 2:19 risks being read as luck, a lucky omen, or feng-shui-style auspiciousness rather than YHWH’s personal, covenantal response to a specific act of obedience — the same risk class as baseline 福音 but now load-bearing for an entire doctrine rather than incidental.
• YHWH/神 collapse (Critical): Nearly every divine reference in Haggai is the personal covenant name (耶和华), not the generic referent (神); collapsing the two erases the personal-ownership force behind “my house” (1:9) and the book’s entire rhetorical structure.
• 殿 defaulting to Buddhist/imperial temple-hall associative field (Critical): Readers’ default mental category for “temple” is drawn from traditions the baseline church entry already excludes (庙/寺/道观); 殿 must never stand unanchored.
• 我的灵 vs. 激动…的心 conflation (Critical/High): Two distinct Hebrew “spirit” actions (2:5 vs. 1:14) land on two different Mandarin renderings; without disciplined dual-footnoting, readers with no Trinitarian catechesis are likely to merge them or read 我的灵 through the ancestral-spirit (祖灵) associative field.
• 印 + 拣选 double Mandate-of-Heaven exposure (High): Signet ring and election language occur in the same climactic verse (2:23); fencing one term without the other leaves China’s imperial-seal (玉玺) / Mandate-of-Heaven (天命) legitimacy framing half-open.
• 万国所羡慕的 silent resolution (Critical): A genuine treasure-reading/personal-messianic-reading fork exists beneath the CUV’s settled rendering; automated translation could resolve it silently, losing an exegetically significant ambiguity.

Opportunities

• Mandarin is well-positioned for this curriculum: the CUV tradition already supplies settled, community-recognized renderings for nearly every Haggai term (耶和华, 万军之耶和华, 大祭司, 我的灵, 万国所羡慕的), meaning the translation task is defensive curation rather than invention.
• Haggai’s compact size (four dated oracles, 38 verses) makes full-book, verse-level theologian review operationally feasible within Phase 2 — every Critical/High term can realistically receive individual review rather than sampling.
• The book’s tight my-house/his-house rhetorical structure (1:9) and repeated refrain (万军之耶和华, 14+ occurrences) create natural teaching anchors: once the fencing notes for these two terms are locked, they carry consistent protective value across the whole book with minimal marginal effort per additional occurrence.
• Jewish yirat Hashem and Islamic taqwa, and Jewish Shekhinah theology, offer legitimate, respectful bridge-teaching cross-references for “Reverent Fear of the LORD” and “the Presence and Spirit of God” that do not require importing risky Chinese-tradition analogs.

Recommended actions

1. Load assets/translation_memory.json (Haggai) alongside the baseline Romans translation memory for every Phase 2 segment; never allow a Haggai segment to be processed against the baseline file alone.
2. Route all 7 Critical-risk terms and both proximate Critical/High Spirit-texts (2:5/1:14) to mandatory human theologian review with no automated-approval path, per assets/doctrine_risk_registry.json.
3. Lock the 1:9 possessive-parallel clause structure and the 2:19 temporal-causal blessing anchor as fixed sentence-level templates in translation memory, not merely as lexical entries, before batch Phase 2 processing begins.
4. Require the standing dual-footnote discipline (我的灵 vs. 激动…的心) and the joint 印/拣选 fencing note to be applied at every occurrence, not just first occurrence, given the short book’s high repeat-exposure density.
5. Flag 万国所羡慕的 for theologian review at every occurrence without exception, recording both readings, rather than treating the CUV’s settled rendering as closing the question.