01 Executive Summary — 2 Corinthians — English → Bengali Translation Requirements

Why it matters

2 Corinthians is Paul’s most personal, theologically dense, and rhetorically exposed letter — and its anchor passage, 5:11-21, compresses reconciliation, new creation, and substitutionary atonement into eleven verses that carry the highest doctrinal risk concentration of any passage analyzed in this Language Package to date, Romans included. Getting Bengali right here is not a stylistic exercise: a single mistranslated word (মিলন read as ritual appeasement, নতুন সৃষ্টি read as reincarnation, “became sin” read as Christ’s moral corruption) can quietly relocate the gospel into a Hindu or Muslim theological framework without a single sentence appearing overtly wrong to a lay reader.

Key findings

• 86 terms are now tracked in this book’s translation memory: 24 inherited exactly from the Romans baseline, 62 newly coined or newly fenced for this letter.
• Of those 86 terms, 17 are Critical-risk and 38 are High-risk — 55 terms (64% of the tracked lexicon) require mandatory human theologian review under the escalation rules.
• Of the 29 doctrines tracked in doctrine_risk_registry.json, 7 are Critical and 16 are High — 23 of 29 doctrines (79%) route to mandatory human theologian review; only 5 route to native-speaker review and 1 (Thanksgiving) is automated-only.
• The core passage (5:11-21) alone contains 6 of the 7 Critical-risk doctrines in the entire book: Reconciliation with God, New Creation in Christ, Substitutionary Atonement, Deity of Christ (via 5:19), Christian Identity in Christ, and the Trinitarian benediction pattern echoed later at 13:14.
• Two entirely new forbidden-substitution terms join the baseline’s list: ক্লেশ (forbidden for “affliction” — collides with the Hindu/Buddhist yogic pañca-kleśa doctrine) and প্রতিমা (forbidden for “image/likeness” — collides with Hindu devotional idol-statue imagery); পর্দা and বিমা are added as forbidden primary renderings for “veil” and “judgment seat” respectively.
• Four terms require a mandatory contextual qualifier rather than outright prohibition: গর্ব (boasting — must always specify “in the Lord” vs. “in the flesh”), ঈর্ষা (jealousy — must always appear as “ঈশ্বরসম্মত ঈর্ষা”), দর্শন (visions — must always be qualified against Hindu temple-image viewing), and দান (giving — must always be anchored to “অনুগ্রহে,” grace).

Risks

• Highest single risk: 5:21’s “he made him to be sin for us” — a Christological tightrope between implying Christ’s moral corruption and softening the verse into a vague “took our sins.” No existing Bengali Bible translation fully resolves this tension; every occurrence requires theologian sign-off.
• Highest cultural-collision risk: “New creation” (5:17) sitting one semantic step from পুনর্জন্ম (reincarnation) for Hindu-background readers, especially in excerpted/devotional-card quotation stripped of surrounding argument.
• Highest naming-collision risk: “New covenant”/“old covenant” (3:6,14) sharing identical Bengali phrasing with the established Bengali Bible names for the New and Old Testament canons — a structural ambiguity, not a mistranslation, that requires a disambiguating note at every occurrence.
• Highest pastoral-sensitivity risk: the “false apostles”/“angel of light” polemic (11:13-15) in a religious landscape dense with gurus, pirs, and charismatic teachers — mistranslation could either blunt Paul’s specific warning or inflame it into a blanket condemnation of other Bengali religious traditions.
• Systemic risk across chs. 8-9: the same word (অনুগ্রহ) carrying both soteriological grace and grace-motivated giving, adjacent to দান’s live Hindu dāna merit-generating connotation — two independent collision vectors stacked in the same two chapters.

Opportunities

• Bengali culture’s lived, everyday experience of discerning authentic from fraudulent gurus gives the “Genuine versus False Apostleship” doctrine (chs. 10-12) unusually strong real-world resonance, if framed without naming other traditions condemningly.
• The familiar Bengali commercial practice of বায়না (advance payment) gives “guarantee/pledge” (ἀρραβών, 1:22; 5:5) a vivid, immediately intelligible cultural anchor with zero syncretism risk.
• “Power in Weakness” (chs. 4, 12) offers a genuine apologetic contrast point against guru-culture siddhi (supernatural-attainment) expectations, turning a translation risk into a teaching opportunity when the paradox is preserved rather than softened.
• The clay-pot “treasure in jars of clay” image (4:7) and the agrarian “sowing and reaping” metaphor (9:6) are both culturally native to rural Bengali life and require no cultural adaptation, only doctrinal guardrails already documented.

Recommended actions

1. Route every segment inside 5:11-21 to mandatory human theologian review by default, regardless of which specific term triggers the flag, per the core-passage override rule now written into 12_ai_translation_requirements.md.
2. Enforce the two new forbidden substitutions (ক্লেশ, প্রতিমা) and two new forbidden-as-primary terms (পর্দা, বিমা) at the automated validation layer before any segment reaches human review.
3. Require the ten mandatory translator notes listed in 12_ai_translation_requirements.md (new creation, reconciliation, became sin, covenant naming collision x2, veil, temple metaphor, grace-in-giving, দান-anchoring, false apostles scope, visions, power-in-weakness) to be generated automatically at first occurrence rather than left to translator discretion.
4. Brief all native-speaker reviewers specifically on the গর্ব/ঈর্ষা/দর্শন/দান contextual-qualifier rules, since these are validation failures that a purely doctrinal (non-native-speaker) reviewer is unlikely to catch.
5. Confirm the YouVersion numeric version ID for the Bengali Common Language Bible before Phase 2 production begins, per the open action item in 05_translation_landscape.md Section 5.1.