Executive Summary: Zephaniah → Mandarin Translation Requirements

Curriculum: Zephaniah 1–3 | Core passage: Zephaniah 3:14–17 | Destination: Mandarin Chinese (Simplified, CUV 神版 register)

Why it matters

Zephaniah is a short, three-chapter prophetic book, but it is doctrinally dense and sits directly on top of the highest-risk semantic territory the Mandarin Bible-translation tradition has already identified: the folk-religious vocabulary of fortune and misfortune (福/祸), the Confucian virtue vocabulary (义), and the folk-devotional model of a resident local deity. This package extends — without contradicting — the baseline Romans Language Package to cover Zephaniah’s full text, ensuring the curriculum’s four core doctrines (the Day of the LORD, judgment on idolatry and complacency, the remnant of Israel, and God’s joyful love for his people) are taught with theological precision rather than lost to cultural-collision drift. Because the core passage (3:14–17) climaxes with God himself singing over his people — a relational direction with no ready folk-religious parallel — getting this package right is a prerequisite for the curriculum’s central pastoral payoff to land as intended.

Key findings

• Full-book coverage confirmed. All three chapters were analyzed end-to-end; no section was silently omitted. Chapters 1–2 and 3:1–13/18–20 supply the necessary judgment, idolatry, complacency, and remnant vocabulary that the core passage (3:14–17) presupposes and resolves.
• 16 distinct doctrines identified in the Zephaniah-specific risk registry: 7 Critical, 5 High, 4 Medium, 0 Low — a markedly higher concentration of severe risk than a typical NT epistle, driven by the book’s dense concentration of named rival deities, a book-defining covenant name (YHWH) requiring strict distinction from the generic term for God, and the single highest cultural-collision line identified anywhere in this translation program (Zephaniah 1:12’s 福/祸 creed).
• 12 of 16 doctrines require mandatory human theologian review; the remaining 4 require native-speaker review. Zero doctrines are automated-review-only — a notably stricter routing profile than the Romans baseline, reflecting the book’s OT covenant-lawsuit genre and its unusually acute folk-religion collision surface.
• ~85 terms are now enforced in the Zephaniah translation memory: 20 inherited exactly from the Romans baseline (god, grace, faith, salvation, righteousness, holy, law, sin, gentiles, glory, power_of_god, messiah, prophet, prophecy, covenant, election, providence, kingdom_of_god, adoption, israel) and ~65 new terms introduced by Zephaniah’s Hebrew prophetic vocabulary. Combined, approximately 13 terms carry Critical risk and 37 carry High risk — roughly 50 terms requiring absolute glossary enforcement with no permitted deviation.

Risks

• Highest risk (Critical, book-defining): Zephaniah 1:12’s “the LORD will do neither good nor harm” is rendered with 福/祸, the exact vocabulary of Chinese folk divination (求福避祸). If excerpted without its condemned-creed framing, this line could be mistaken for devotional wisdom rather than the false, refuted creed the prophet is diagnosing.
• Critical, structural: The covenant name 耶和华 (YHWH) must remain strictly distinct from the generic 神 (Elohim) throughout — a distinction the Romans baseline did not need to enforce (it uses only 神/主) but which is load-bearing for Zephaniah’s entire judgment-to-restoration argument.
• Critical, doctrinal payoff: 余民 (“remnant”) must never drift toward the generic Mandarin word for “survivor” (幸存者), which would strip the covenant-election dimension from the book’s central positive category — the very addressee of the core passage’s joy.
• Critical, core-passage-specific: In 3:17, God must remain the grammatical subject and his people the object of joy (“因你…”); Mandarin’s unmarked default direction of religious joy (worshipper toward deity) risks inverting the passage’s central, most cross-culturally distinctive claim.
• High, pervasive: Named rival deities (Baal, Milcom) risk being read as acceptable participants in a normal multi-deity devotional landscape rather than covenant-breaking rivals, given Chinese folk religion’s non-exclusive pantheon norms.

Opportunities

• Zephaniah’s short length and tight three-doctrine structure make it an efficient, high-impact module for reinforcing (not just introducing) the baseline’s hardest-won Mandarin theological vocabulary — 义, 圣洁, 律法, 恩典-adjacent concepts — in a fresh narrative context, deepening retention across the wider curriculum.
• The core passage’s picture of God singing over his people is a distinctive, memorable teaching hook precisely because it has no folk-religious analogue; framed well, this can become a positive point of cultural contrast rather than a liability.
• Zephaniah 3:9’s “pure speech” and “call on the name of the LORD” give a natural, ready-made bridge to Romans 10:13, letting the curriculum demonstrate Old Testament–New Testament continuity using vocabulary the learner has already encountered.

Recommended actions

1. Route every occurrence of the 1:12 福/祸 creed, the YHWH/神 distinction, 余民, and 3:14–17 in its entirety through mandatory human theologian review, per the escalation rules in analysis/12_ai_translation_requirements.md.
2. Enforce the Zephaniah assets/translation_memory.json and assets/doctrine_risk_registry.json alongside — never instead of — the Romans baseline files in every Phase 2 session.
3. Require explicit teaching-note cross-references at three structural points: (a) ch.1 wrath ↔ 3:14-17 joy as one Day of the LORD; (b) 3:11 pride ↔ 3:14 exultation as a deliberate root-echo reversal; (c) 2:7/2:9 ↔ 3:12-13 as one continuous remnant term.
4. Prioritize native-speaker + theologian joint review capacity for Phase 2, given zero doctrines in this book are automated-review-eligible.

Prepared under TRI Phase 1, Step 11, consistent with assets/doctrine_risk_registry.json (v1, 2026-07-19) and analysis/11_doctrine_analysis.md.