Executive Summary: Matthew Javanese Language Package (Phase 1)

Why it matters

Matthew is the first curriculum built on top of the Romans-established Javanese Language Package, and it stress-tests that foundation against a much wider doctrinal surface: nine curriculum doctrines spanning the Kingdom of Heaven, Christ’s identity, the Sermon on the Mount’s ethical demands, church discipline, the end of the age, and the Great Commission. Javanese carries a live, everyday syncretic layer — kejawen mysticism, Jayabaya-rooted prophetic expectation, and keraton court-hierarchy honorifics — sitting alongside its Muslim-majority religious context. Matthew’s own vocabulary (Son of Man, Immanuel, “bind and loose,” “the end of the age,” the Beatitudes) collides with this layer far more often, and in more varied ways, than Romans did. Getting this Language Package right protects the doctrinal integrity of the single most narratively rich, most quoted, and most catechetically central Gospel in the whole New Testament curriculum set.

Key findings

• 30 doctrines analyzed, full-book coverage confirmed. Every chapter of Matthew (1–28) was mapped against the doctrine set; only chapter 14 introduces no new load-bearing vocabulary, and this was explicitly confirmed rather than silently skipped.
• 15 doctrines are Critical risk, 13 High, 2 Medium, 0 Low — a markedly higher Critical/High concentration than Romans, reflecting Matthew’s dense narrative Christology (Son of Man, Incarnation, Deity/Sonship of Christ, Resurrection) and its distinctive apocalyptic material (Judgment, Parousia, End of the Age).
• 28 of 30 doctrines require mandatory human theologian review; only 2 (Mission to the Nations, Faith) route to native speaker review, and none are automated-only — a stricter profile than the Romans baseline’s 24-of-31 theologian-review ratio.
• 57 new theological terms were introduced beyond the 37 Romans-inherited terms now carried forward into assets/translation_memory.json, with the Matthew-specific glossary alone containing 15 Critical and 22 High risk new/extended terms (37 requiring theologian review) plus 15 Medium risk terms (native speaker review).
• Four genuinely new syncretism risks not present in Romans were identified: (1) Kratoning Swarga’s collision with Hindu-Buddhist merit-earned afterlife cosmology, (2) pungkasaning jaman/Parousia’s collision with the Jayabaya Ratu Adil/Satrio Piningit cyclical-prophecy tradition, (3) “bind and loose” idioms colliding with occult spirit-binding practice (santet, pengasihan), and (4) the Beatitudes’ own vocabulary (Rahayu, andhap asor, mlarat ing roh, kang resik ing atine) each colliding with a distinct Javanese cultural counterfeit.
• The core passage, Matthew 5:1-12, is the single densest doctrinal convergence point in the curriculum, touching four separate registry doctrines (Beatitudes, Kingdom of Heaven, Adoption/Sonship of Believers, Sanctification) in twelve verses.

Risks

• Critical-tier terms cluster in eight chapters (1, 3, 16, 20, 24, 26, 27, 28), meaning theologian review capacity will be unevenly demanded across the book rather than smoothly distributed — scheduling risk for Phase 2 if not planned for explicitly.
• The “bind and loose” idiom (16:19; 18:18) and the “end of the age”/Parousia language (13:39-40; 24:3-44; 25:31; 28:20) are flagged for zero-sampling, every-occurrence theologian review — any Phase 2 workflow that samples rather than reviews exhaustively at these points would violate this Language Package’s own risk controls.
• “Perfect” (sampurna, 5:48) sits in dangerous phonetic/conceptual proximity to the already-forbidden kasampurnan (ascetic mystical perfection); this single-verse term carries outsized risk relative to its narrow textual footprint.
• The plural “sons of God” (5:9) and the singular “Son of God” title must never be allowed to collapse into one rendering; this is a structural risk specific to Matthew’s Beatitudes that Romans’ vocabulary never had to guard against.

Opportunities

• The existing Javanese Bible tradition already supplies settled, familiar renderings for the highest-frequency terms (Gusti Allah, Gusti Yesus, Roh Suci, Sang Mesias, wungu saka pati), meaning Matthew’s translation task is mainly one of extension and fencing, not invention from nothing.
• Matthew 15:11 offers the clearest single teaching text in the whole Bible for anchoring dosa (sin) to moral transgression rather than ritual impurity (cemer) — this passage should be promoted as the primary cross-reference for that doctrine across all future curricula in this language.
• A fixed, verbatim rendering strategy for the Great Commission (28:18-20) and the Beatitudes (5:1-12) — both already recommended in prior analysis — gives Phase 2 a natural quality-consistency anchor, mirroring how the Romans baseline anchored Romans 8:28 and 10:9-10.
• The kejawen-specific syncretism map produced here (Ratu Adil, manunggaling kawula gusti, santet-adjacent binding magic, svarga cosmology) is directly reusable for any future Javanese curriculum touching eschatology, Christology, or church polity, reducing analysis cost for subsequent books.

Recommended actions

1. Load assets/translation_memory.json, assets/bible_term_registry.json, and assets/doctrine_risk_registry.json (Matthew) alongside the unmodified Romans baseline files before any Phase 2 segment translation begins.
2. Schedule theologian review capacity with explicit weighting toward chapters 1, 3, 16, 20, 24, 26, 27, and 28, rather than even distribution across all 28 chapters.
3. Enforce zero-sampling, full-occurrence theologian review for “bind and loose” (16:19; 18:18) and all end-of-the-age/Parousia language, per the escalation rules in 12_ai_translation_requirements.md.
4. Adopt fixed, verbatim renderings for Matthew 5:1-12 and 28:18-20 across all curriculum materials before Phase 2 batch processing begins, to prevent drift across parallel workers.
5. Carry the kejawen-specific comparative theology and gap analysis produced for this curriculum forward as a reusable reference for future Javanese-language Bible curricula.