01 Executive Summary — Jude (English → Malayalam)

Why it matters

Malayalam is the best-positioned language in this pipeline for AI-assisted Bible translation — nearly two millennia of settled, ecumenically shared Christian vocabulary means most of Jude’s theological terms already have safe, established renderings inherited directly from the Romans baseline. But Jude compresses more Critical-tier doctrine per verse than any Romans chapter, and several of its most vivid, rhetorically loaded terms (desire, master, worldly, angel, blameless, last time) sit directly on top of Kerala’s own religious and social vocabulary neighborhoods — Hindu bhakti and Kāma associations, Jyothisham astrology, Samkhya philosophy, caste-purity history, and Kali Yuga cyclical time. Getting this short letter wrong is not a vocabulary-shortage problem; it is a vigilance problem.

Key findings

• 60 total glossary terms govern Jude in this Language Package: 18 inherited exactly from the Romans baseline (several escalated in risk for this book) and 42 newly introduced by Jude’s distinctive vocabulary.
• 32 of those 60 terms (53%) are Critical or High risk — 10 Critical, 22 High — requiring mandatory human theologian review at every occurrence per the escalation rules.
• 22 doctrines span the full letter (salutation through doxology); 17 of 22 (77%) require human theologian review, 5 require native speaker review, and zero are automated-only — a materially higher theologian-review burden than typical Romans chapters, reflecting the letter’s density rather than any deficiency in Malayalam’s Christian vocabulary tradition.
• Jude introduces one genuinely new structural translation-discipline requirement absent from Romans: the τηρέω (“kept”) word-family recurs at vv. 1, 6, 13, 21, and 24 in both protective and punitive senses, and requires the identical Malayalam verb root at every occurrence to preserve the letter’s deliberate wordplay — a consistency requirement, not a lexical gap.
• The single most theologically delicate point in the letter is Jude’s citation of material paralleling the non-canonical text 1 Enoch (v.14) and his allusion to extra-biblical tradition about Moses’ body (v.9); every occurrence requires theologian review with a mandatory canon-clarification note.

Risks

• Highest-severity lexical risk: ἐπιθυμία (“lustful desire,” vv. 16, 18) rendered with കാമം instead of the required മോഹം would import the entire mythological frame of Kāma, the Hindu deity of desire, into a passage about sin.
• Highest-severity structural risk: silent drift in the τηρέω/“kept” verb root across vv. 1, 6, 13, 21, 24 would destroy Jude’s central literary and theological structure (God’s keeping vs. judicial custody) without producing any single-verse error an isolated review would catch.
• Highest-severity doctrinal risk: mishandling the Enoch citation (v.14) either implies 1 Enoch is canonical Scripture or, in overcorrection, undermines confidence in Jude’s own inspired citation — both are serious failures.
• Kerala-specific collision risks requiring mandatory fencing notes: ഭക്തിഹീനര്‍ (“ungodly,” negating ഭക്തി), നാഥന്‍ vs. rejected യജമാനന്‍ (“master,” feudal/caste-tenant risk), ദൂതന്‍ vs. rejected മലക്ക് (“angel,” Islamic-register risk), ജഡികര്‍ vs. rejected പ്രാകൃതര്‍ (“worldly,” Samkhya-prakriti risk), നിത്യാഗ്നി vs. purgatory conflation (“eternal fire”), അലഞ്ഞുനടക്കുന്ന നക്ഷത്രങ്ങള്‍ (“wandering stars,” Jyothisham-astrology risk), അന്ത്യകാലം/എന്നേക്കും vs. Kali Yuga cyclical-time conflation (“last time”/“forever”), and കളങ്കമില്ലാത്തവരായി vs. caste-purity conflation (“blameless”).
• Escalated-from-baseline risk: grace (v.4) is escalated from the baseline’s High to Critical for this curriculum specifically because Jude’s charge is grace being deliberately perverted into license, a sharper and more severe claim than the baseline’s general grace-vs.-merit caution.

Opportunities

• Malayalam’s established vocabulary handles nearly all of Jude’s OT-typology proper names (Cain, Balaam, Korah, Sodom and Gomorrah, Moses, Michael, Enoch) and core Christological/soteriological terms without inventing new renderings — translation work here is disciplined enforcement, not invention.
• The Saint Thomas Christian community’s own apostolic-era self-understanding gives Jude 1:3’s “faith once for all delivered” strong, ready-made local resonance rather than requiring the term to be built from nothing.
• Jude’s compact length (25 verses) means the full theologian-review burden, while proportionally high, is absolutely small — full-book coverage is achievable with focused, high-attention review rather than large-scale review capacity.
• The fully populated translation_memory.json for this book (60 terms) and its accompanying 12_ai_translation_requirements.md extension give Phase 2 translators and AI systems a complete, load-and-enforce-ready package requiring no further term research before segment translation begins.

Recommended actions

1. Load the Jude translation_memory.json together with the baseline Romans translation_memory.json before any Phase 2 segment translation; never translate Jude terms from either file in isolation.
2. Enforce the τηρέω/“kept” verb-root consistency rule as a Priority-0 structural check, ahead of ordinary Critical/High term enforcement, in every automated validation pass.
3. Route all segments touching vv. 3, 4, 6–7, 8–10, 9 & 14–15, 17–19, 20–21, 22–23, and 24–25 to mandatory human theologian review per the escalation rules in 12_ai_translation_requirements.md; do not allow automated-only approval for any of these spans.
4. Attach the mandatory explanatory teacher/translator notes (Enoch/canon, grace-abuse, blameless/caste-purity, wandering-stars/astrology, last-time/Kali-Yuga) to every lesson touching the relevant verse, not only to the first lesson that introduces the term.
5. Carry the eight absolute forbidden-substitution terms specific to Jude (കാമം, യജമാനന്‍, മലക്ക്, പ്രാകൃതര്‍, plus the four inherited baseline forbidden terms still relevant here) into every automated validation pass alongside the baseline’s existing forbidden list.