Executive Summary — Luke | English → Marathi Language Package

Smart Brevity format.

Why it matters

Marathi has a mature Bible translation tradition and an established Christian vocabulary for the highest-stakes terms (तारण, देहधारण, पुनरुत्थान, नीतिमत्त्व), inherited directly from the Romans-curriculum baseline. But the Gospel of Luke is a different kind of text than an epistle: it is dense, narrative, and saturated with concrete social scenes — synagogue readings, table fellowship with tax collectors, banquets, a despised Samaritan, a crucified criminal — that put doctrine into action in front of a reader shaped by two distinct Maharashtra religious currents: Warkari bhakti devotion to Vitthal, and the Ambedkarite Navayana Buddhist movement’s explicit, still-living rejection of caste-linked karma theology. Luke’s core passage (4:16-21, Jesus reading Isaiah 61 and declaring “today this Scripture has been fulfilled”) concentrates nearly every one of this curriculum’s highest-risk terms in six verses, making it the correct anchor for translator training even though full-book coverage — all 24 chapters — is mandatory and has been completed.

Key findings

• 107 total terms now govern Phase 2 translation: 41 inherited exactly from the Romans baseline (non-negotiable), plus 66 newly introduced by Luke’s narrative material (synagogue, Sabbath, anointed, Son of Man, disciple, compassion, Paradise, and more).
• 28 curriculum doctrines span the full Gospel; 14 are Critical and 12 are High risk (26 total require mandatory human theologian review), with only 2 Medium-risk doctrines routed to native speaker review and zero Low-risk doctrines in this curriculum — reflecting how much of Luke’s material is directly Christological or directly touches Maharashtra’s specific syncretism fault lines.
• The single most structurally unavoidable risk is शिष्य (disciple): Marathi has no substitute term that escapes the pervasive Hindu guru-śiṣya and Warkari sant-lineage frame, in which a disciple progresses toward the guru’s own attainment. This collision recurs across the entire Gospel and cannot be solved lexically — only through sustained, repeated translator notes.
• Son of Man (मनुष्याचा पुत्र), occurring 18+ times, is Critical and new to this curriculum: the literal Marathi reads as merely “a human’s son,” entirely erasing the Daniel 7 background that makes this Jesus’s most paradoxical and load-bearing self-designation.
• The core passage alone concentrates 8 of the curriculum’s highest-risk terms (अभिषिक्त केले, पूर्ण झाले आहे, आज, गरीब, बंदिवानांना मुक्तता, आंधळ्यांना दृष्टी, पीडित, प्रभूच्या कृपेचे वर्ष), confirming it as the correct anchor for both training data and reviewer prioritization.

Risks

• Caste-karma echo risk (acute, recurring): Luke 13:1-5’s rejection of “your suffering proves your guilt” reasoning, and the broader universal-accountability and providence doctrines, sit directly on top of the doctrine Ambedkar’s 1956 conversion explicitly and publicly rejected. Any drift toward that framing is not merely inaccurate — it is re-traumatizing for a living community.
• Savior/compassion crowding: तारणारा (Savior) and करुणा (the common word for “compassion”) both sit in crowded semantic neighborhoods already occupied by Vitthal-devotion and Buddhist Brahmavihara practice respectively; unanchored use of either risks the reader substituting a familiar devotional or meditative figure/virtue for the text’s actual claim.
• Ambiguity at doctrinally load-bearing points: Luke 17:20-21 (“kingdom within/among you”) is genuinely ambiguous in the Greek itself, and the Marathi rendering choice (तुमच्यामध्ये) carries real risk of sounding like Hindu antaryāmī or Buddhist inner-awakening theology if handled carelessly.
• Idiom misfire: Luke 14:26’s comparative “hate” of family, if translated plainly, risks readers rejecting the passage as morally repugnant rather than understanding the Semitic idiom of comparative preference it actually expresses.

Opportunities

• Luke’s universal-scope material (Jew/Gentile, Samaritan, rich/poor, tax collector, Zacchaeus, the Good Samaritan) offers an unusually direct, positive point of contact with the lived history of both Maharashtra religious communities’ concerns about hierarchy and exclusion — this can be taught as gospel content, not merely defended against misreading.
• The compassion and table-fellowship material (5:27-32, 7:36-50, 15:1-32, 19:1-10) gives this curriculum a strong, concrete, story-driven vehicle for teaching grace-as-unearned-gift to a Navayana Buddhist-background reader who has no prior category for grace at all.
• Because 41 of 107 terms are already fixed and battle-tested from the Romans package, Phase 2 translators inherit a substantial head start; effort can concentrate on the 66 genuinely new Luke terms and especially the 8 ranked highest-risk ambiguities identified in the linguistic gap analysis.

Recommended actions

1. Schedule Luke 4:16-21 (the core passage) as the first unit reviewed in Phase 2, given its unusually high concentration of Critical/High terms, while theologian review capacity is freshest.
2. Enforce mandatory, repeated (not one-time) translator notes on शिष्य, मनुष्याचा पुत्र, स्वतःला नाकारणे, आज, पूर्ण झाले आहे, अभिषिक्त केले, and तुमच्यामध्ये at every occurrence, per the ranked risk list in 06_linguistic_gap_analysis.md.
3. Route all 26 Critical/High doctrines and their associated terms to mandatory human theologian review before Phase 2 sign-off; route the 2 Medium doctrines (Kingdom Mission to All Nations, Table Fellowship as Kingdom Practice) to native speaker review.
4. Brief all theologian reviewers explicitly on the Ambedkarite 1956 conversion’s rejection of birth-karma theology before reviewing Luke 13:1-5, 15:1-32, and the providence/election doctrines, so the pastoral stakes are understood, not merely the lexical ones.
5. Version-lock assets/translation_memory.json (Luke edition) before Phase 2 begins; any new term discovered mid-translation must be added and version-incremented before further segments proceed, per the Translation Memory Load and Enforcement Instructions in 12_ai_translation_requirements.md.