01 Executive Summary — Mark 1–16 (English → Malayalam)

Why it matters

Mark is the Gospel most structured around a single theological hinge — Jesus’ identity as suffering Son of Man/Son of God — and its climactic statement of purpose, the ransom saying of Mark 10:45, is one of the two or three highest-doctrinal-density verses translated anywhere in this pipeline. Malayalam enters this task from an unusually strong position: nearly two millennia of continuous, settled Saint Thomas Christian vocabulary (documented fully in the Romans baseline) means most terms in Mark need no invention, only disciplined reuse. But Mark introduces several genuinely new, Kerala-specific collision risks the Romans baseline never had to face — a near-homophone between “Corban” and the Eucharistic “Qurbana,” a folk-exorcism collision on “demon,” a bonded-labor/caste-history collision on “slave,” a cyclical-cosmology collision on “hell,” and a structural lexical gap around the Greek word for divine “necessity” in the passion predictions. Getting Phase 2 translation right here is the difference between a curriculum that faithfully preserves the Ransom-for-Many doctrine and one that quietly drains it into generic sacrifice-language or, worse, into an unintended echo of Kerala’s own sacred liturgy.

Key findings

• 33 doctrines mapped across all 16 chapters (full-book coverage confirmed chapter-by-chapter); 8 Critical, 15 High, 7 Medium, 3 Low risk.
• 23 of 33 doctrines (70%) require mandatory human theologian review; 7 require native speaker review; only 3 are automated-review-only.
• 56 new Mark-specific terms proposed for translation memory (11 Critical, 20 High, 17 Medium, 8 Low), on top of ~45 terms inherited exactly from the Romans baseline with no rendering altered.
• The single highest-priority term in the entire curriculum is മറുവില (ransom, Mark 10:45) — vocabulary is adequate and precise, but the entire risk is AI-generated softening toward generic “sacrifice” or “gift” language that would erase Isaiah 53’s substitutionary force.
• One genuine structural lexical gap exists: Malayalam has no single word for the Greek δεῖ (“it is necessary,” the divine-determination sense behind “the Son of Man must suffer,” 8:31/9:31/10:33-34); this requires a fixed periphrastic construction and standing translator note, not a glossary term.
• Five Kerala-specific crowded-vocabulary risks were newly identified beyond anything documented in the Romans baseline: Corban/Qurbana (Critical form-collision), ഭൂതം “demon”/folk-exorcism, ദാസന്‍ “slave”/bonded-labor history, നരകം “hell”/cyclical naraka cosmology, and ഗുരു “Teacher”-address/guru-bhakti dynamic.
• The Son of Man title (മനുഷ്യപുത്രന്‍) occurs 12 times and is the highest-frequency Critical Christological term in the book — no lexical ambiguity, but strict cross-chapter consistency is essential to preserve the Messianic Secret pattern.

Risks

• Doctrinal softening of the ransom saying (10:45) is the top risk: an AI system could plausibly substitute ബലി (sacrifice) or a generic “gift,” which is fluent Malayalam but doctrinally destructive.
• Sacred-vocabulary collision: a bare transliteration of “Corban” (കൊര്‍ബാന്‍, 7:11) sits one vowel away from കുർബാന (Qurbana/Eucharist) — the single most acute form-collision risk identified anywhere in this pipeline to date.
• Fatalism drift: the necessity-of-the-cross construction and the “elect” language of chapter 13 both risk sliding toward വിധി (fate)-adjacent phrasing if not actively guarded.
• Social-history mishandling: ദാസന്‍ (“slave,” 10:44) carries live memory of Kerala’s landlord-era bonded/caste agricultural labor system; mistranslation or unflagged use risks either causing offense or being misheard as scriptural endorsement of historical servitude.
• Motif fragmentation: the Messianic Secret’s silencing-verb family and the three-fold passion-prediction formula are both scattered across many chapters; inconsistent Phase 2 segment-by-segment translation could make these deliberate authorial patterns invisible to Malayalam readers even without any single mistranslation.

Opportunities

• Malayalam’s settled vocabulary base means Phase 2 effort can concentrate almost entirely on enforcement and consistency, not on inventing new theological terminology — a comparative advantage documented already in the Romans baseline and confirmed here for Mark.
• The Gentile centurion’s confession at the cross (15:39), forming a deliberate inclusio with Mark 1:1, offers a strong pastoral teaching opportunity to address Kerala Christianity’s own internal Syrian-Christian/Dalit-convert heritage stratification using the text’s own “outsider confesses first” logic.
• Mark’s terse, urgent narrative style is well suited to Malayalam’s own strong oral/narrative Bible-teaching tradition; preserving that pace (rather than smoothing it into literary register) is an opportunity to make the translation feel vivid, not merely accurate.
• The five newly identified Kerala-specific collision risks (Corban, ഭൂതം, ദാസന്‍, നരകം, ഗുരു) can each be resolved with a single, reusable translator-note template, creating a durable asset for any future non-Pauline New Testament curriculum in this pipeline.

Recommended actions

1. Lock മറുവില and അനേകര്‍ക്കു പകരം as zero-deviation terms with mandatory theologian sign-off at every occurrence before Phase 2 begins.
2. Resolve the Corban/Qurbana collision before translation starts: obtain theologian sign-off on either the recommended descriptive paraphrase or a transliteration-plus-mandatory-note approach; do not leave this to per-segment AI judgment.
3. Fix the periphrastic necessity construction for δεῖ now, as a single approved template, and apply it identically at all three passion predictions.
4. Route Mark 8–10 and Mark 14–15 to theologian-available reviewers by default in Phase 2 batch processing, given their disproportionate concentration of Critical terms.
5. Build the five Kerala-specific translator-note templates (Corban, ഭൂതം, ദാസന്‍, നരകം, ഗുരു) into the Phase 2 tooling now, so they fire automatically rather than depending on ad hoc reviewer memory.