Executive Summary: Song of Solomon (English → Spanish)

Why it matters

Song of Solomon puts one of the highest concentrations of Critical-tier translation risk in this entire Language Package family into just eight short chapters. Unlike Romans, whose risks cluster around abstract doctrinal vocabulary (justification, grace, saints), this book’s risks are concrete and immediate: an ancient Hebrew love poem whose central passage (8:6-7) doubles as this curriculum’s core text and its single most theologically loaded unit. A poor translation choice here doesn’t just blur a doctrine — it can make Scripture appear to endorse incest (“hermana”), praise toxic jealousy (“celos”), or erase the only likely explicit divine-name reference in the book (shalhevetyah). Getting this right requires holding two things together that Spanish-speaking Christian culture, both Catholic and Evangelical, tends to pull apart: the goodness of embodied marital desire, and the historic Christ-and-Church allegorical reading. Neither may cancel the other.

Key findings

• 6 Critical and 8 High doctrine-level risks span all eight chapters (doctrine_risk_registry.json), with zero Low-risk doctrines in the whole book — every chapter carries at least High-tier content.
• 56 total glossary terms are now locked in assets/translation_memory.json: 8 inherited exactly from the baseline Romans package (Señor, Jesús, Cristo, pacto, David, Israel, iglesia, paz) and 48 new terms unique to this book’s Hebrew poetic vocabulary.
• The core passage (8:6-7) alone carries five overlapping Critical/High doctrines simultaneously — the seal image, righteous covenant jealousy, the possible embedded divine name, the sealed-garden cross-reference to 4:12, and faithfulness/longing imagery — making it the single highest-density unit in the curriculum.
• One decision has no baseline precedent anywhere in this Language Package: whether to render shalhevetyah as “llama del SEÑOR” (recovering the embedded divine name) or follow RV1960/RVA2015’s “fuerte llama” (which obscures it). This requires explicit theologian sign-off before any Phase 2 work begins.
• Destination-culture collision points are unusually severe and specific: “celos” defaults negative in Hispanic pastoral counseling; “hermana” applied romantically reads as incest with no ANE-idiom equivalent; “Rosa de Sarón” is already a popular devotional title for Christ, inverting the source text’s speaker; “terrible” is a false friend for the archaic “awe-inspiring.”

Risks

• 14 doctrines require mandatory human theologian review (all Critical + all High tier); only 3 require native-speaker-level review, and none are automated-only — a materially higher review burden per chapter than the Romans baseline.
• Simultaneous, opposite-direction failure modes on the allegorical doctrine: over-spiritualizing erases the literal goodness of marital intimacy; under-allegorizing severs the book from Ephesians 5 and Revelation 19/21. Both are live in the same Spanish-speaking audience.
• Silent, undetectable loss risk on teshuqah (7:10): a translator working only from an English gloss of “desire” would have no lexical signal that anything unusual is happening, since ordinary Spanish “deseo” carries no built-in marker of the Genesis 3:16/4:7 echo.
• Colorist-connotation risk on “morena” (1:5-6) is specific to this book and has no equivalent flagged risk elsewhere in the baseline Romans package.

Opportunities

• This curriculum can become the Language Package’s clearest teaching case for holding literal and typological reading together without collapsing either — directly modeling the “dual register” discipline this whole pipeline exists to enforce.
• Mandatory end-user-facing teaching notes (not just internal translator flags) at 9 specific trigger points give this curriculum an unusually strong, built-in correction mechanism against the most common popular-devotional misreadings (Rosa de Sarón, sister-bride, jealousy).
• Locking the seal-vocabulary family (sello/sellada/cerrado) across 4:12 and 8:6, and the vineyard inclusio across 1:6 and 8:12, lets Phase 2 translation surface the book’s own deliberate literary structure for Spanish readers — a payoff unavailable to translations that treat each chapter in isolation.

Recommended actions

1. Gate Phase 2 on theologian sign-off for the four framework-defining Critical decisions before any segment translation begins: shalhevetyah/“llama del SEÑOR” (8:6), qin’ah/“celo” (8:6), achoti kallah/“hermana mía, esposa mía” (4:9-5:1), and gan na’ul…/“jardín cerrado… fuente sellada” (4:12).
2. Process chapters 4 and 8 first, since downstream vocabulary decisions (the seal/sealed word family) are set there and every other chapter’s Critical/High content depends on consistency with those two chapters.
3. Attach all nine mandatory end-user-facing teaching notes as shipped curriculum content, not internal-only flags, per analysis/12_ai_translation_requirements.md.
4. Enforce the three locked refrains (adjuration: 2:7/3:5/8:4; mutual-belonging: 2:16/6:3; gazelle epithet: 2:9/2:17/8:14) with automated consistency checks at validation, since these are mechanically verifiable and any drift is a structural, not just lexical, failure.
5. Route the executive summary’s risk profile to stakeholders before scheduling Phase 2 timelines: with 14 of 17 doctrines requiring theologian review and zero automated-only doctrines, this curriculum’s review burden per verse is substantially higher than the Romans baseline and should be resourced accordingly.