01 Executive Summary — James 1–5 Translation Requirements (English → Bengali)

Why it matters

James is the New Testament book most likely to be misread as contradicting Romans on the ground of salvation — its core passage (James 2:14-26) says a person “is justified by works,” using the same Greek verb family, and quoting the same Old Testament text (Genesis 15:6), that Romans 3–4 uses to say the opposite. In Bengali, this risk is compounded, not reduced: the destination culture supplies ready-made merit-accounting frameworks (Hindu karma, Islamic weighed deeds) that a careless translation could accidentally confirm rather than correct. Getting this letter right requires the same rigor as Romans, applied to a shorter, blunter, proverb-dense text with its own distinct hazards — a hell/rebirth collision (Gehenna vs. Naraka), a “wisdom” term with a liberation-theology cognate, and a single Greek verb (σῴζω) that means both “heal” and “save eternally” within eleven verses of chapter 5.

Key findings

• The core passage is the single highest-risk unit in the curriculum. James 2:14-26 requires the identical Bengali phrase for “justified” that Romans uses (ধার্মিক বলে গণ্য হওয়া), but with a mandatory, theologian-approved disambiguating note at every occurrence (2:21, 24, 25) — using a different phrase would break cross-document consistency; using the same phrase silently would suggest Scripture teaches contradictory doctrines of salvation.
• “Works” (ἔργα) must never be কর্ম. Bengali কর্ম is saturated with Hindu karma-merit theology — the theological inverse of James’s actual point that works are faith’s fruit, not its cause. This single forbidden-term rule is the most consequential lexical decision in the whole letter.
• σῴζω’s healing/salvation polysemy is a genuine intra-book trap. The same Greek verb spans physical healing (5:15, রক্ষা করা/সুস্থ করা) and eternal salvation (2:14, 5:20, পরিত্রাণ) — defaulting to পরিত্রাণ throughout chapter 5 would wrongly equate a granted healing with a guaranteed eternal outcome, a distortion this curriculum’s target audience (spanning folk-healing-saturated Hindu, Muslim, and prosperity-adjacent charismatic Christian backgrounds) is unusually primed to make.
• Bengal-specific religious collision terms recur across all five chapters, not just the core passage: নরক/Naraka (hell, ch.3), সংসার/samsara-renunciation (world, ch.1 & 4), অভিষেক/Durga Puja anointing (ch.5), সিদ্ধ/yogic attainment (perfect/mature, ch.1-3), জ্ঞান/jñāna-mārga (wisdom, ch.1 & 3), and জিন/ambiguous jinn morality (demons, ch.2-3).
• One deliberate, counter-intuitive exception exists to the baseline’s own rule: ধর্ম, forbidden everywhere for “righteousness,” is the correct rendering for “religion” in James 1:26-27 (θρησκεία) — and this exception itself is a documented risk if not clearly flagged, since an under-briefed reviewer might “correct” it back.
• Full-book coverage is confirmed. All five chapters and every verse-range have been traced to a doctrine, a term, or an explicit “reviewed, folded into adjacent doctrine” note — no chapter was silently skipped (see analysis/11_doctrine_analysis.md §2).

Risks

• 19 Critical-risk and 34 High-risk terms (53 total) are recorded in assets/bible_term_registry.json (James) requiring human theologian sign-off before first use — nearly every one carries a specific, named Bengali religious-vocabulary collision (Hindu, Islamic, or both), not a generic caution.
• 5 Critical and 8 High doctrines (13 of 15 total doctrines) require human theologian review per assets/doctrine_risk_registry.json; only 2 doctrines (Truthful Speech and Oaths; Effective Prayer of the Righteous) are downgraded to native-speaker-only review, and zero doctrines are automated-only — a materially higher theologian-review burden proportionally than the Romans baseline (which had 3 automated-only doctrines).
• James 4:5 is textually and interpretively disputed even within mainstream Bengali print editions (Holy Spirit vs. human spirit as subject) and must never be rendered with unflagged confidence.
• Gehenna/নরক has no clean resolution — unlike every other forbidden-term case in this curriculum, no safe alternative Bengali word exists at all; mitigation is a mandatory per-occurrence note, not a better word choice.
• Fringe-rendering exposure is real and documented, not theoretical: Muslim-idiom contextualized translations, Jehovah’s Witnesses’ subordinationist NWT, Word-of-Faith prosperity paraphrase, and Hindu-inculturation devotional literature each have specific, identified failure patterns against James 2:1, 2:14-26, 5:14-16, and the wisdom passages (analysis/05_translation_landscape.md §3-4).

Opportunities

• James 1:27 and chapter 2’s poverty ethics land unusually well across both Bengali audiences. Care for orphans/widows and rejection of status-based favoritism find strong independent warrant in both Islamic Qur’anic ethics and Hindu charitable-giving traditions (dana), while directly and constructively confronting residual caste-consciousness and communal/class hierarchy — a rare point of near-universal cross-tradition reception.
• The “faith and works” tension, properly footnoted, is a strong teaching opportunity, not just a risk — Bengali readers from both Hindu (guru-testing-disciple) and Muslim (Ibrahim’s testing) backgrounds already carry a schema for “a test that proves genuine devotion,” which the mandatory disambiguation note can redirect toward James’s actual point rather than a merit-earning misreading.
• “If the Lord wills” (4:15) and “submit to God” (4:7) are genuine, theologically compatible points of contact with Islamic devotional idiom (ইনশাআল্লাহ; the root sense of islam/submission) — flagged for native-speaker framing rather than avoidance, these can serve as authentic bridges in gospel conversation rather than liabilities to be designed around.
• The Genesis 15:6 verbatim cross-reference (James 2:23 ↔ Romans 4:3) is a ready-made, textually anchored harmonizing tool — featuring it prominently in study notes turns the curriculum’s single hardest doctrinal puzzle into a demonstrable proof that Scripture is self-consistent, rather than leaving the tension for critics to exploit.

Recommended actions

1. Lock the “justified” disambiguation note and the σῴζω consolidated note as reusable Phase 2 templates before any batch translation begins — do not allow these to be re-authored ad hoc per lesson, given how many separate passages (2:21/24/25 and 2:14/5:15/5:20 respectively) each one touches.
2. Route James 2:14-26 in its entirety, plus James 3:6 (Gehenna) and James 5:14-16 (healing/anointing/confession), to the deepest-expertise theologian reviewers first in Phase 2 Step 16 batch scheduling — these three zones concentrate the letter’s highest-stakes risk.
3. Flag James 4:5 for mandatory dual-reading treatment (record both candidate subject readings as alternatives_considered) rather than letting any single worker silently resolve the textual ambiguity.
4. Audit every Phase 2 draft against the ten Bengal-specific forbidden substitutions (কর্ম, সিদ্ধ, জ্ঞান, জিন/অপদেবতা, প্রতিমা/প্রতিমূর্তি, অভিষেক-as-primary, সংসার, bare কামনা, দ্বীন-outside-scope, শক্তি-root compounds) using the validation table in analysis/12_ai_translation_requirements.md before any segment is marked approved.
5. Confirm the YouVersion version ID (provisionally 1207) against the live catalog before publishing any hyperlink, per analysis/05_translation_landscape.md §5, and propagate the confirmed ID to every downstream James document.
6. Brief every Phase 2 reviewer explicitly on the single sanctioned ধর্ম exception (James 1:26-27 only) so it is not “corrected” back to an alternative rendering by a reviewer over-applying the Romans baseline’s blanket prohibition.