Executive Summary: Numbers → Arabic Language Package

Why it matters

Numbers is the first Old Testament / Torah curriculum built on this Arabic Language Package, and it introduces a fundamentally different risk profile than the Romans baseline: instead of epistolary argument about grace and justification, the destination audience encounters wilderness narrative, sacrificial law, and a healing serpent-image that sits directly on top of the single most sensitive folk-religious collision surface in the region — talismanic object-veneration. The core passage, Numbers 21:4-9, is doctrinally dense (unbelief, intercession, and a Christ-typology converge in seven verses) and must be translated with the same rigor the baseline applies to Romans 1:16-17 and 10:9. Because this is Torah, several doctrines (corporate generational judgment, God-initiated atonement, messianic prophecy spoken through a pagan diviner) have no clean existing Arabic vocabulary and required new compositional terms, not just new glossary entries.

Key findings

• 81 total glossary terms now govern Numbers translation: 24 terms inherited verbatim from the Romans baseline (faith, grace, salvation, law, messiah, lord, etc.) and 57 new terms surfaced by full-book analysis (chs. 1–36), including 7 entirely new compositional terms Arabic did not previously have fixed renderings for (unbelief, divine judgment, God’s faithfulness, typology, wilderness generation, red heifer, Nechushtan).
• 44 terms carry Critical or High risk (12 Critical + 32 High) and require human theologian review at every occurrence, more than double the term-level review burden of a typical Romans document segment.
• 14 doctrines are tracked for Numbers (7 named in the curriculum brief + 7 surfaced by full-book coverage); 10 of the 14 require human theologian review (3 Critical + 7 High), 3 require native-speaker review, and only 1 (Leadership Succession) is automated-review-only.
• The single highest-priority correction from this analysis phase is not new vocabulary but a revision: the Nazirite term (ch. 6) was found to collide, in its bare form (نذير), with the far more common Quranic homograph for “warner/prophet” — the entry has been revised to a disambiguated compound (الناذر لِلرَّبّ) and escalated to High risk.
• All 36 chapters of Numbers were reviewed; 14 chapters carry no independently new doctrine and are explicitly logged as reviewed-with-no-new-content, satisfying the full-book coverage mandate.

Risks

• Talismanic misreading of the bronze serpent (Critical). Regional folk practice — amulets and protective objects, formally condemned within Islamic theology itself as shirk-adjacent but culturally persistent — is the most likely default reading of Numbers 21:8-9 if the command-obedience-trust chain is not scaffolded in every occurrence.
• Atonement collapsing into self-performed kaffarah (Critical). Aaron’s and Phinehas’s mediating acts (16:46-48; 25:13) risk being read as the Islamic legal category of kaffarah — a fixed, self-performed penitential act — rather than God-initiated, mediator-accomplished atonement halting judgment already underway.
• Messianic prophecy diluted or read astrologically (Critical). Balaam’s “star… scepter” (24:17) is the book’s highest-density single-verse risk: it must receive the same active correction as every occurrence of المسيح, while also resisting an astrological or divination-endorsing misreading given its pagan-diviner source.
• Corporate judgment individualized away (High). Islamic fitrah theology has no category for a whole generation collectively excluded from covenant blessing; translators may unconsciously soften Numbers’ corporate accountability into isolated personal failures.
• Political over-reading of إسرائيل (High). The Balaam narrative’s triumphant, irreversible blessing language (23:19-24:9) is a strong flashpoint for contemporary political misreading if not anchored explicitly to redemptive history.
• Legal/cultural over-identification (High). Inheritance law (chs. 27, 32, 34, 36), cities of refuge (ch. 35), and cherem (chs. 21, 31) each sit unusually close to live Arab legal/customary or Islamic-legal categories (fiqh al-mirath, wali al-dam/qisas-diya, haram) and risk being flattened into those familiar frameworks rather than taught as their distinct covenantal counterparts.

Opportunities

• Genuine, positive bridge vocabulary exists and should be leveraged deliberately, not just defended against: manna (shared with Quranic al-mann wa-l-salwa), qorban/offering (shared with Eid al-Adha qurban), divination’s shared condemnation (kihanah), and rizq/tawakkul-adjacent providence language all give the translator authentic common ground to build from before pivoting to the text’s distinct claims.
• The bronze serpent’s own internal safeguard (2 Kings 18:4) is a ready-made apologetic tool: teaching that this very object was later destroyed as an idol pre-empts the talismanic misreading using the Bible’s own historical record rather than an imported argument.
• Rabbinic tradition’s own anti-magic reading of the bronze serpent (Mishnah Rosh Hashanah 3:8) and its own caution against generalizing Phinehas’s zeal (b. Sanhedrin 82a) offer respected, non-Christian precedent for exactly the boundaries this Language Package needs to teach — useful for cross-tradition credibility.
• Establishing this package’s seven new compositional terms now (unbelief, judgment, faithfulness, typology, wilderness generation, red heifer, Nechushtan) closes a vocabulary gap that will otherwise resurface in every future Old Testament narrative curriculum processed for this language.

Recommended actions

1. Load assets/translation_memory.json and assets/bible_term_registry.json (Numbers) alongside the Romans baseline files before any Phase 2 segment translation begins — never substitute Numbers-local improvisation for a baseline-shared term.
2. Route the core passage (21:4-9) and Balaam’s star/scepter oracle (24:17) to priority, first-pass human theologian review; these two loci carry the highest doctrinal density in the book.
3. Apply the mandatory first-use glosses (الراية, الجماعة, الناذر لِلرَّبّ, and the seven new compositional terms) in every Phase 2 document — repeated at sufficient intervals per document, not stated once and assumed retained.
4. Enforce the revised Nazirite rendering (الناذر لِلرَّبّ) from the first Phase 2 batch onward; this correction was not caught in the initial glossary pass and must not propagate the bare نذير form into any translated output.
5. Confirm the John 3:14-15 cross-reference lock for “look… and live” before Numbers 21:8-9 is finalized in any learner-facing material, per the baseline’s cross-document consistency rules.

---

Full detail: analysis/07_semantic_analysis.md, analysis/08_core_glossary.md, analysis/11_doctrine_analysis.md, analysis/04_comparative_theology.md, analysis/06_linguistic_gap_analysis.md, assets/doctrine_risk_registry.json, assets/bible_term_registry.json, assets/translation_memory.json.