Executive Summary — Malachi (English → Arabic)

Why it matters

Malachi closes the Old Testament with the promise that anchors this curriculum: a coming Messenger who prepares the way for the Lord himself, come to purify his people and renew a broken covenant (Malachi 3:1-4). For Arabic-speaking readers — the large majority from a Muslim-majority cultural context — this book sits closer to live legal and theological categories than almost any other in Scripture: it names a messenger (a word Arabic readers default to “angel,” not “human forerunner”), a tithe (a word that is also Islam’s codified agricultural tax), a divorce (a word that names a fully proceduralized fiqh category), and a prophetic oracle (a word that collides head-on with Qur’anic wahy) — all inside its first three chapters. Getting Phase 2 translation right here is not a stylistic exercise; it determines whether readers hear Malachi’s own argument or an Islamic legal/eschatological framework wearing Malachi’s words.

Key findings

• 63 tracked terms now govern Phase 2: 9 copied exactly from the baseline Romans package (لord, covenant, righteousness, law, gentiles, election, father, imputed_righteousness, grace) and 54 new Malachi-specific terms.
• 14 terms are Critical risk (messenger, messenger of the covenant, burden/oracle, tithe, divorce, day of the LORD, sun of righteousness, law of Moses, offerings in righteousness, plus 5 inherited Critical terms — lord, righteousness, law, father, imputed_righteousness) — every occurrence requires mandatory human theologian review.
• 18 terms are High risk (temple, refiner’s fire, refine-and-purify, “I have loved you,” “Esau I hated,” fear of the LORD, priest, dealt treacherously, godly offspring, God unchanging, return-to-me, those-who-fear-the-LORD, consuming fire, Elijah the prophet, profaned-the-covenant, stumbled-in-the-law, plus 2 inherited High terms — election, grace).
• 32 of 63 terms (Critical + High) require mandatory theologian sign-off before any segment is approved — roughly half the entire glossary, a materially higher proportion than the Romans baseline due to Malachi’s density of concrete, legally-codified collision points.
• 13 doctrines are tracked (doctrine_risk_registry.json): 4 Critical, 5 High, 3 Medium, 1 Low — 9 of 13 doctrines require human theologian review, 3 require native speaker review, 1 is automated-only.
• The single highest-density risk cluster in the book is The Coming Messenger and the Day of the Lord (Malachi 3:1-5; 4:1-6): it stacks an angelic-being misreading, a collision with Muhammad’s own defining title (الرسول), an implicit incarnation/deity claim, and a competing Islamic eschatology (Yawm al-Qiyamah) into five verses.

Risks

• Messenger (מַלְאָכִי, 3:1 and throughout): defaults to “angel” in Arabic, erasing the human-forerunner referent (John the Baptist) and, if corrected via رسول-root vocabulary instead, collides with Muhammad’s title. No perfect word exists — only disambiguation notes close this gap, and every occurrence must carry one.
• Tithe (العشور) and Divorce (الطلاق): both are unavoidable, exact-word collisions with fully codified Islamic legal categories (ushr tax; talaq procedure). Absent explicit disambiguation, readers will import an entire foreign jurisprudential framework — rates, exemptions, waiting periods — onto Malachi’s covenant argument.
• Burden/Oracle (الوحي, Malachi 1:1): the very first verse of the book collides with the Qur’an’s own defining category of revelation, setting the interpretive frame for everything that follows before any corrective teaching can occur.
• Election language (Malachi 1:2-3, “Jacob I loved, Esau I hated”): directly quoted in Romans 9:13 — any wording drift between this Malachi curriculum and the existing Romans Language Package rendering creates a visible, citable cross-curriculum inconsistency.
• The Lord / master context-split (אדון at 3:1 vs. 1:6): a single Hebrew root deliberately renders two different Arabic words (الرب vs. السيد) carrying opposite doctrinal weight; a well-intentioned automated “consistency” pass could silently destroy this distinction in Phase 2.
• Priest (الكاهن): carries a pre-Islamic pagan-soothsayer (kahin) association explicitly condemned in hadith tradition, risking a negative first impression of the entire priestly office before its typological fulfillment in Christ can be taught.

Opportunities

• Genuine bridge vocabulary exists and should be used deliberately: God’s justice (إله العدل / al-‘Adl), God’s immutability (لا أتغيّر / kalam divine-attribute discussions), and repentance-as-return (التوبة / teshuvah-adjacent) all have real conceptual overlap with Islamic and Jewish theology that this curriculum can name explicitly as a starting point for teaching the biblical distinctive.
• The two-fires contrast (3:2-3 purifying vs. 4:1 consuming) and the curse-escalation structure (general curse → cherem) give lesson-writers vivid, textually-grounded teaching devices that require no theological argument to land — only faithful preservation of the book’s own internal distinctions.
• Malachi’s dialogical, question-and-answer rhetorical style (God’s charge, the people’s skeptical rebuttal, God’s response) is well-suited to Arabic didactic and homiletic tradition and can be preserved rather than flattened, strengthening both fidelity and audience engagement.
• Elijah/John-the-Baptist typology (4:5-6) offers a concrete, teachable bridge to the New Testament curriculum already governed by this Language Package, reinforcing whole-Bible coherence for Arabic-speaking learners moving between Testaments.

Recommended actions

1. Load assets/translation_memory.json (Malachi) alongside the baseline Romans file for every Phase 2 segment; treat the 14 Critical and 18 High terms as non-negotiable enforcement points per analysis/12_ai_translation_requirements.md.
2. Route all 9 theologian-required doctrines (assets/doctrine_risk_registry.json) through mandatory human review before Phase 2 output is marked approved; do not allow automated tooling to downgrade any tier.
3. Before finalizing Malachi 1:2-3, verify wording against the Romans curriculum’s existing rendering of Romans 9:13.
4. Verify the two documented context-split exceptions (אדון at 1:6/3:1; torah at 2:6/2:8-4:4) are encoded as explicit non-normalization rules in any automated consistency-checking tool used in Phase 2.
5. Ensure every Critical-risk term list in this summary carries its mandatory translator disambiguation note in Phase 2 output, per the escalation rules in analysis/12_ai_translation_requirements.md.