Executive Summary — 2 Corinthians (English → Dutch)

Why it matters: 2 Corinthians puts more Critical-tier Dutch theological collisions into a single 11-verse core passage (5:11-21) than any comparable stretch of the Romans baseline. Reconciliation, non-imputation of sin, and “Christ made sin” all sit directly on top of Dutch-specific fault lines — a heavily secularized conflict-mediation sense of “verzoening,” a self-help culture primed to hear “nieuwe schepping” as a lifestyle rebrand, and a colloquial drift of “zonde” toward “a pity/waste.” Get this passage wrong in Dutch and the letter’s thesis — God, not humanity, reconciled the world through Christ — collapses into a horizontal, human-negotiated peace process.

Key findings:

• Full 13-chapter sweep confirms 27 distinct doctrines map onto 2 Corinthians (vs. 8 named curriculum doctrines), none of which is contradicted by or contradicts the Romans baseline package.
• Doctrine-level risk (doctrine_risk_registry.json): 4 Critical, 16 High, 6 Medium, 1 Low — 20 of 27 doctrines require mandatory human theologian review, versus 6 native-speaker-only and 1 automated-only.
• Term-level risk (translation_memory.json, this curriculum): of 92 total entries (24 inherited verbatim from the Romans baseline, 68 newly introduced by 2 Corinthians), 6 are Critical and 27 are High — a combined 33 terms requiring absolute or near-absolute enforcement, concentrated overwhelmingly in the core passage (5:11-21) and in chapters 3, 10-12.
• Four genuinely new Dutch-specific collision risks not present anywhere in the Romans baseline: (1) “verzoening’s” secular truth-and-reconciliation-commission register, (2) “nieuwe schepping” vs. Dutch wellness/self-help “nieuwe start” culture, (3) “bekering” (repentance) — entirely absent from the Romans glossary — competing with everyday loose usage, and (4) the Geest/geest capitalization decision in 2 Corinthians 3, which the Jehovah’s Witnesses’ Nieuwe-Wereldvertaling exploits to teach an impersonal Spirit.
• Every mainstream Dutch Bible tradition surveyed (SV, HSV, NBV21, WV95) agrees on the Critical-tier renderings this curriculum requires (capitalized Geest at 3:6/3:17, “toerekenen” not “aanrekenen”/“vergeven” at 5:19, “Mijn genade is u genoeg” at 12:9) — giving this curriculum unusually strong cross-tradition backing, except at exactly the points where Het Boek’s paraphrase and the NWT diverge.

Risks:

• Reconciliation with God, Non-Imputation of Sin, Substitutionary Atonement (Christ Made Sin), Imputed Righteousness — all 4 Critical doctrines cluster in 5:18-21 alone; a single mistranslation here (e.g., “vergeven” for “niet toerekenen,” or an active “verzoen u met God” instead of the required passive) undoes the letter’s central argument.
• “Ambassadeur” vs. “gezant” (5:20) and “andere apostelen” vs. “valse apostelen” (11:13) are single-word substitution risks that each trivialize an entire High-risk doctrine (Apostolic Authority; Genuine vs. False Apostleship) if the wrong cognate is chosen by default.
• “Roem/roemen” (boasting) and “zwakheid” (weakness) both require context-sensitive valence-flipping across chs. 1-12 — the highest-frequency, widest-scope risk pattern in the book, exceeding even the Romans baseline’s “called/roeping” flag in scope.
• Trinitarian benediction (13:14) and “Mijn genade is u genoeg” (12:9) are fixed liturgical/pastoral phrases; any document-to-document drift will be immediately noticeable to churched Dutch readers.

Opportunities:

• Dutch Reformed forensic vocabulary (“toerekenen,” “toegerekende gerechtigheid”) already precisely matches 2 Corinthians 5:19’s non-imputation logic — a genuine asset inherited directly from the Romans baseline requiring no new coinage.
• Dutch’s built-in distinction between “voorbede” (intercession) and “voorspraak” (saintly/Marian advocacy), already noted in the baseline, extends cleanly to this book’s comfort/suffering passages.
• Near-universal cross-tradition agreement on capitalized “Geest” (3:6, 3:17) and on “Mijn genade is u genoeg” (12:9) means this curriculum can adopt firm fixed renderings with high confidence and minimal reviewer disagreement risk.

Recommended actions:

1. Route all Critical-tier segments in 5:18-21 through mandatory human theologian review with zero exceptions, per the escalation rules in 12_ai_translation_requirements.md.
2. Enforce the two single-point forbidden substitutions (“gezant” not “ambassadeur”; “valse apostelen” not “andere apostelen”/“concurrerende leraars”) as hard validation-rule blockers, not soft flags.
3. Require explicit contextual anchoring (“in de Heer”/“in Christus”) for every legitimate-sense occurrence of “roem/roemen,” and explicit power/grace-sufficiency co-location for every occurrence of “zwakheid.”
4. Lock the three fixed phrases (5:19-21’s imputation clauses, 12:9, 13:14) into translation memory as verbatim, cross-document-consistent strings before any Phase 2 segment translation begins.
5. Confirm reviewer sign-off on the Geest/geest capitalization convention for all of 2 Corinthians 3 before Phase 2 begins, since this single decision propagates across every occurrence in the chapter and carries cult-collision risk (NWT) if defaulted incorrectly.