Executive Summary: Acts (Urdu Language Package)

Why it matters

Acts is the theological hinge connecting the Gospels to the epistles this pipeline has already translated for Urdu — and it puts the baseline’s highest-stakes doctrines on repeat, in the mouths of multiple independent apostolic witnesses, rather than in a single epistolary argument. Where Romans states the crucifixion once as settled fact, Acts has Peter, then Peter again, then Peter again, then Paul assert the same “you crucified him… God raised him” formula across five separate sermons (2:23,36; 4:10; 5:30; 10:39; 13:28-29) — precisely the event Qur’an 4:157 denies happened at all. Acts also introduces a risk category the baseline never had to handle: contemporary, live blasphemy law. Stephen’s trial (ch. 6-7) uses vocabulary (کفر/گستاخی) that carries real, present-day legal exposure under Pakistani and Indian criminal statutes, not merely historical-doctrinal friction. Getting Acts right in Urdu is not a bigger version of the Romans task — it is a distinct task requiring its own review protocols.

Key findings

• Doctrine-level risk: 36 doctrines identified across all 28 chapters — 13 Critical, 16 High, 5 Medium, 2 Low. 29 of 36 (81%) require mandatory human theologian review; only 5 route to native speaker and 2 to automated review.
• Term-level risk: 76 glossary terms tracked for Acts — 19 Critical, 25 High (44 total, 58%, requiring theologian-tier enforcement), 20 Medium, 12 Low.
• The core passage (Acts 2:1-41) is the single densest concentration of Critical risk in the whole curriculum: it stacks Holy Spirit personhood, the crucifixion accusation, Messiahship, Lordship, Sonship-adjacent exaltation language, and the repentance/baptism/forgiveness/Spirit-gift formula (2:38) into 41 verses.
• One genuinely new risk category, absent from Romans/Galatians entirely: blasphemy (Acts 6-7, 26:11) collides with contemporary Pakistan Penal Code Section 295-C and related statutes — a live legal-safety issue, not a historical-theological one, requiring combined theologian + native-speaker legal-context review, a review pairing found nowhere else in this Language Package family.
• توبہ (repentance) joins نجات as the second major “shared word, two soteriological mechanisms” risk term: established Christian vocabulary that is simultaneously Islam’s central tawbah concept operating within the mizan (deeds-weighed-at-judgment) framework, requiring contrastive teaching at every load-bearing occurrence (2:38 above all).
• شہید (martyr) and یہ راہ (the Way) are genuine bridge-and-collision terms: correct, established renderings that also carry strong pre-existing associations (jihad-martyrdom paradigm; al-sirat al-mustaqim) requiring explicit reframing, not avoidance.
• Full-book coverage confirms no chapter was silently skipped: chapters 25 and 27 introduce no new doctrine and are explicitly marked “reviewed, no new registry doctrine” per the full-coverage mandate.

Risks

1. Highest-priority escalation item in the entire curriculum: the repeated crucifixion-accusation formula. Any softening, passivization, or allegorization across its five occurrences would both misrepresent the text and obscure its own rhetorical strategy (independent, consistent apostolic testimony).
2. Legal/physical safety risk, not just translation risk: mishandling blasphemy vocabulary or omitting the mandatory contemporary-safety note in Stephen’s narrative could expose readers, teachers, or the material’s distributors to real-world consequences under blasphemy statutes.
3. False-agreement risk: several Acts passages (10:34 impartiality; 17:23 unknown God) sound like immediate common ground with Islamic theology and risk being taught as settling doctrinal agreement rather than as a “yes, and also” bridge toward Acts’ specific, exclusive content.
4. Consistency-drift risk in parallel processing: Paul’s threefold conversion narrative (ch. 9, 22, 26) and the five-fold crucifixion-accusation formula both depend on cross-chapter lexical consistency that standard per-file parallel batch processing could silently erode.
5. Circumcision and clean/unclean (ch. 10-11, 15): both name currently, widely practiced Islamic categories (khatna; halal/haram-adjacent dietary law); careless anchoring could be read as direct commentary on contemporary Muslim practice rather than the text’s own narrow, historical point.

Opportunities

• توبہ, نذر, رویا/خواب, خدا کسی کا طرفدار نہیں, and دیگر زبانیں are genuine, positive bridge points documented in 04_comparative_theology.md — usable pastorally to build initial rapport before introducing Acts’ distinctive content, provided the “yes, and also” contrastive move is always completed, never left implicit.
• Acts’ own narrative (Peter’s vision, the Jerusalem Council, Paul’s inclusion argument) already models the “teach on its own terms without appearing to concede or rebut” approach this Language Package recommends for Qur’anic-parallel material — the source text itself is a resource for pastoral framing, not just a translation challenge.
• The book’s closing bracket (1:8 → 28:31, “unhindered”) offers a non-triumphalist, hopeful frame for discussing real persecution risk with contemporary readers without minimizing the cost.

Recommended actions

1. Load assets/translation_memory.json (Acts) and analysis/12_ai_translation_requirements.md (Acts) alongside the baseline files for every Phase 2 session — never substitute one for the other.
2. Route all Acts 6:11, 6:13, and 26:11 segments to combined theologian + native-speaker legal-context review before any teaching material is finalized; treat this as a hard gate, not a recommendation.
3. Batch-process the five crucifixion-accusation occurrences and the three Paul-conversion narrations together (same worker/session) to protect cross-chapter consistency per the performance notes in 12_ai_translation_requirements.md.
4. Prioritize theologian review capacity toward the 29 Critical/High doctrines before Medium/Low material; do not let native-speaker-routed content (Great Commission fulfillment, church leadership, tongues-as-known-languages) consume theologian review bandwidth needed elsewhere.
5. Brief all reviewers on the “shared word, two mechanisms” pattern (نجات, توبہ) and the “bridge-and-collision” pattern (شہید, یہ راہ) as named categories before review begins, so contrastive-teaching gaps are caught systematically rather than ad hoc.