Executive Summary — Genesis → Bengali Translation Requirements

Why it matters: Genesis is the theological foundation for all eight curriculum doctrines — Creation, the Image of God, the Fall, the Abrahamic Covenant, Providence, Common Grace, Covenant Faithfulness, and Joseph’s story — and it lands in Bengali directly on top of the two dominant religious vocabularies of the region: Hindu devotional/cosmogonic language (West Bengal) and Islamic prophetic/covenantal language (Bangladesh). More than Romans, Genesis is concrete, ritual, and kinship-dense — exactly the terrain where those two vocabularies compete hardest with the intended biblical meaning. Getting this Language Package right determines whether Bengali learners receive the doctrine of a personal, sovereign Creator or unconsciously absorb an emanationist, karmic, or prophet-successionist substitute.

Key findings:

• প্রতিমূর্তি (Image of God) is the single highest syncretism risk in the entire curriculum: it shares its root with মূর্তি, the ordinary word for a crafted Hindu idol, and inverts the doctrine’s actual direction (God→humanity, not human→god) unless fenced with a mandatory note at every occurrence (Gen 1:26-27; 5:1,3; 9:6).
• সৃষ্টি (create ex nihilo) is the only viable Bengali rendering of bara but is also the standard Hindu term for cyclical, emanationist cosmic re-manifestation — a collision that hits readers in the curriculum’s very first verse.
• Genesis 15:6 is the exact Old Testament verse Paul quotes in Romans 4:3 for imputed righteousness; this package mandates identical Bengali rendering across both curricula, making Genesis-Romans consistency a hard requirement, not a stylistic preference.
• Three Critical-tier terms — আশীর্বাদ (blessing), ত্বকচ্ছেদ (circumcision), and ঈশ্বর তা ভালোর জন্য ব্যবহার করলেন (Providence, Gen 50:20) — each collide with a specific, well-documented Bengali religious practice (Hindu ritual blessing-transfer, Islamic Sunnah-linked circumcision, and Hindu Mangal-Kavya auspiciousness literature, respectively) and each requires an absolute forbidden-substitution rule, not just a preference.
• The Genesis doctrine risk registry carries 10 Critical and 13 High risk doctrines (23 total requiring human theologian review), 6 Medium (native speaker review), and 2 Low (automated) — a heavier theologian-review load than the Romans baseline, reflecting Genesis’s higher concentration of cosmogonic and ritual vocabulary.
• The Genesis translation memory adds ~99 new terms to the ~27 terms inherited exactly from the Romans package; of the new terms, 11 are Critical risk and 27 are High risk — combined with inherited Critical/High terms, 57 terms across both files require Critical/High-tier enforcement priority in Phase 2.

Risks:

• Silent doctrinal drift if সৃষ্টি, প্রতিমূর্তি, আশীর্বাদ, or জাতি are translated without their mandatory fencing notes — each has a “natural,” fluent-sounding Bengali rendering that is also a well-worn path into a different religion’s theology.
• Genesis-Romans inconsistency at the doctrinal seams (15:6 vs. Romans 4:3; 50:20 vs. Romans 8:28) would sever the cross-curriculum bridges this project depends on for teaching continuity.
• The Isaac/Ishmael covenant-line divergence and the Shiloh/messianic-scepter cruces are pastorally live for Bengali Muslim-background readers and must never be silently resolved or softened.

Opportunities:

• Bengali’s family/lineage vocabulary (বংশ, গোষ্ঠী) offers a natural teaching bridge for Covenant Faithfulness across Generations, once fenced against an automatic-inheritance misreading.
• The Joseph narrative’s close parallel to the well-known and beloved Surah Yusuf offers a genuine, low-risk bridge point for Bengali Muslim-background audiences, provided Qur’anic phrasing and prophetology are not imported into the Genesis-specific providence framing.
• El Shaddai’s fixed creedal rendering (সর্বশক্তিমান ঈশ্বর) demonstrates that established liturgical entrenchment can coexist safely with a bare-noun prohibition (শক্তি) elsewhere — a documented, reusable pattern for future books facing similar title-vs.-bare-noun tension.

Recommended actions:

1. Load assets/translation_memory.json (Genesis) and assets/bible_term_registry.json (Genesis) alongside the unmodified Romans-package files before any Phase 2 segment translation begins — never Genesis-only.
2. Enforce the seven Genesis-specific mandatory (every-occurrence, not first-occurrence-only) footnote triggers specified in 12_ai_translation_requirements.md v1.28: creation, image of God, blessing, nation, circumcision, burnt offering (at Gen 22), and Providence’s “good” (Gen 50:20).
3. Route all 23 theologian-review doctrines and every Critical/High term occurrence through human theologian review before Phase 3 sign-off; route the 6 Medium-tier doctrines (dreams, Table of Nations, marriage complementarity, Angel of the LORD, intercessory prayer, Sabbath/creation order) through native speaker review.
4. Treat Genesis 15:6 and Genesis 50:20 as hard consistency checkpoints against the existing Romans 4:3 and Romans 8:28 renderings in every quality-assurance pass.
5. Carry forward the El Shaddai/সর্বশক্তিমান exception-handling pattern explicitly into Phase 2 automated validation so the শক্তি-prohibition does not generate false-positive flags on this fixed title.