Hosea — Linguistic Gap Analysis (Hmong Language Package)

0. Purpose and Method

This document executes PRD Phase 1 Step 8. It does not re-derive the term-by-term work already completed in analysis/08_core_glossary.md and assets/bible_term_registry.json; instead it steps back from the item-level inventory to ask three structural questions across the whole book:

1. Doctrine vocabulary matrix — for each of the five assigned doctrines, what Hmong vocabulary is actually available, where does it strain, and what is the recommended handling strategy?
2. Missing vocabulary vs. crowded semantic neighborhoods — where does Hmong lack a term at all (a true gap, requiring periphrasis/coinage), versus where does Hmong have a term but it is already doing service for several unrelated concepts (a crowded neighborhood, requiring “fencing” — mandatory qualifiers, footnotes, or lexical partitioning so hearers don’t cross-contaminate senses)?
3. Transliteration vs. paraphrase — for the unusually dense proper-name and wordplay load of Hosea (name etymologies, the Baal pun, the Beth-Aven pun, the shuv/meshuvah pun), which items should be transliterated, which paraphrased, and which require both plus a footnote?

All findings here are binding inputs to Phase 2 generation and are cumulative with, and subordinate to, the Romans Hmong baseline documents. Nothing below authorizes a rendering that contradicts assets/translation_memory.json or the Hosea glossary/registry already produced.

---

1. Doctrine Vocabulary Matrix

Doctrine	Available Hmong Terms	Weaknesses	Recommended Strategy
D1 — God’s Steadfast Love despite Israel’s Unfaithfulness	txoj kev hlub ruaj khov (chesed, B1/registry#4); kev hlub (ahav, general love); txoj kev khuv leej (rachamim/mercy); kev ncaj thiab kev ruaj khov (emunah/faithfulness)	Hmong has no single native lexeme corresponding to chesed’s fusion of “loyalty owed within a binding relationship” + “love freely continued despite the other party’s default.” The nearest cultural analogue, kwv tij clan-duty loyalty, is reciprocal — it lapses when the other party defaults. Four distinct Hebrew love/loyalty/mercy terms (ahav, chesed, rachamim, emunah) risk collapsing into one generic “kev hlub” if translators are not disciplined, which would erase the very distinctions Hosea 2:19-20’s fourfold betrothal-gift list depends on.	Maintain the four-way lexical partition already fixed in the glossary (ahav / chesed+ruaj khov / rachamim / emunah+ruaj khov) as a closed set, never interchanged. Every occurrence of chesed must carry the “ruaj khov” qualifier without exception (same structural-qualifier discipline as the Romans baseline’s hwjchim→Vajtswv hwjchim rule). Teaching notes must explicitly contrast covenantal chesed with reciprocal kwv tij loyalty at first occurrence (2:19) and at the doctrine’s climax (11:8-9; 14:4).
D2 — Spiritual Adultery and Idolatry as Covenant Breaking	kev ua nkauj yi (zanah); kev deev luag / kev sib deev (naaph); cov khub deev (m’ahabim); dab mlom (idol, reused Romans baseline); tus mlom nyuj kub (golden calf); Txoj Kevcai (ntawm Vajtswv/Mauxes) (torah, reused); Vajtswv txoj kev foob thiab tsis txaus siab (riv/lawsuit)	The sexual-metaphor vocabulary (zanah/naaph) is lexically available but register-risky: Hmong formal register has no neutral, non-crude way to name “harlotry” that also carries covenant-theological weight rather than mere gossip-register shock value. There is also a scope problem: zanah (general/habitual) and naaph (specific act against an existing bond) are distinguished in Hebrew but risk being merged in casual Hmong usage, flattening ch.1-3’s escalating argument.	Fence both terms with mandatory sign-act framing language (“Hauv paj lus qhia txog…”, “this is a prophetic picture of…”) at every first occurrence in a lesson unit, per the forbidden-substitution table in 08_core_glossary.md §C. Keep zanah/naaph lexically distinct wherever the Hebrew itself distinguishes them; where the text does not force the distinction, default to zanah’s broader term rather than merging downward into naaph’s narrower one.
D3 — The Marriage Metaphor for God and His People	qhaib coj los ua poj niam mus ib txhis (betroth+forever qualifier); kuv tus txiv / kuv tus Npa-as (ishi/ba’ali contrast); Npa-as (Baal, transliterated); phau lus cog tseg (ntawm Vajtswv) (berith)	Hmong marriage vocabulary is robust (betrothal, bride-price, husband/wife roles) and transfers positively on the whole — the risk is not absence but untranslatable wordplay: the ishi/ba’ali pun (2:16-17) and the underlying Baal/master-lord ambiguity cannot be reproduced lexically in Hmong (tus tswv legitimately means both “master” and, contextually, “the LORD’s title-adjacent term” tus Tswv). This is a genuine structural gap, not a weak term.	Do not attempt a forced Hmong pun. Use the fixed disambiguated pair (kuv tus txiv vs. kuv tus Npa-as/kuv tus tswv) and carry the wordplay explanation entirely in a mandatory footnote rather than in the translated text itself. This is the clearest case in the whole book where paraphrase-plus-footnote outperforms any attempt at lexical cleverness.
D4 — Judgment and Compassion Held Together	kev npau taws ntawm Vajtswv (wrath, reused); kev khuv leej tig los hlub (nocham/relenting); kuv yog Vajtswv, tsis yog neeg (ontological ground); tub tuag teb (Sheol); kev txhiv (redemption, elevated Critical); Tus Dawb Huv (Holy One)	The core passage’s rhetorical hinge (11:8-9) requires holding two states in tension simultaneously — deserved wrath and prevailing compassion — inside a single Hmong sentence structure. Hmong narrative idiom for divine/spirit anger (npau taws) is drawn from a cultural script where anger is either vented or ritually appeased and resolved, not one where anger is voluntarily restrained by the angry party’s own character. This is a structural-narrative gap, not merely a lexical one.	Do not render 11:8-9 as a simple “God changed His mind” statement (forbidden per registry#23’s note). Require a doctrinal framing clause in every lesson touching this passage: wrath is real and deserved, and is held back by God’s own self-consistent holy compassion (Tus Dawb Huv, 11:9b) — not because the offense was minor, and not because God is unstable like the capricious spirit-anger of folk narrative.
D5 — The Call to Return and Repentance	rov qab los cuag Vajtswv (shuv, positive); kev tig rov qab ntawm Vajtswv (meshuvah, negative); hu (qara/call); kho kom zoo (rapha/heal); kev paub Vajtswv (yada); nrhiav (baqash/seek)	The shuv/meshuvah wordplay (same triliteral root deployed for both “return to God” and “turn away from God,” and even for literal “return to Egypt,” 11:5) has no equivalent root-play available in Hmong; each sense must be assigned a different Hmong phrase, meaning the Hebrew’s ironic self-echo is invisible in translation unless explicitly taught. Separately, hu (call) is a crowded-neighborhood risk (see §2 below).	Treat the shuv/meshuvah wordplay as a teaching-note item, not a lexical-recovery target: state explicitly in any lesson touching 11:1-11 or 14:1 that the Hebrew uses one word-root for both “return” and “turn away,” and that this is deliberate irony no Hmong rendering can reproduce directly. Keep the positive/negative phrase pair rigidly consistent across all six chapters where it occurs (5, 6, 7, 11, 12, 14).

---

2. Crowded Semantic Neighborhoods (Existing Term, Multiple Competing Referents)

These are cases where Hmong does have a workable word, but that word is already carrying — or is at risk of being pulled into — more than one theologically distinct referent within this curriculum alone (setting aside the Romans-baseline collisions it may already carry). Each requires an explicit fencing strategy: a structural qualifier, a closed-set discipline, or a mandatory footnote, so that Phase 2 generation never silently drifts a term across a fence line.

Crowded Term	Competing Referents Pulling on It	Fencing Strategy
hu (call)	(a) God’s covenantal historical summons of Israel as son (11:1, D5); (b) the animistic hu plig ritual of calling a wandering soul back into the body; (c) generic everyday “to call someone’s name.”	Reserve hu for (a) and (c) only, and require every (a)-sense occurrence to carry an explicit relational/historical anchor clause (“…raws li kev cog lus, tsis yog ib txoj kev ua neeb hu plig…”). This exact fence is already flagged [TRANSLATOR REVIEW NEEDED] in registry#31; this document elevates it to a standing structural rule for all of chapter 11’s Phase 2 content, not a one-off flag.
dawb huv (holy/pure)	(a) Tus Dawb Huv as a divine title (11:9, D4); (b) the Romans baseline’s kev raug ua kom dawb huv (progressive sanctification); (c) folk ritual purity from spirit-contamination (the baseline’s own documented collision).	Keep the title usage (Tus Dawb Huv) and the process usage (kev raug ua kom dawb huv) visually and grammatically distinct — one is a fixed nominal title, the other a verbal-noun process phrase — and never blend them into a single sentence without a clarifying clause. Both must carry the folk-ritual-purity contrast note at first occurrence in any given lesson module, per the Romans baseline’s existing discipline, now extended to the title-usage as well.
fij (sacrifice/offer) family	(a) illegitimate sacrifice to Baal (2:8, 11:2); (b) legitimate-but-misused ritual observance critiqued in 6:6 and 8:11-13; (c) the Romans baseline’s Christ’s-atonement referent (not textually present in Hosea but must not be contradicted by any Hosea-curriculum usage); (d) “bulls of our lips” as reinterpreted verbal praise (14:2).	Maintain the three/four-way disambiguation discipline already established in registry#36: every occurrence must be textually marked as to which of these senses is intended; (d) in particular must never be rendered literally (see §3 below — this is also a paraphrase decision). No Hosea-curriculum sentence may use bare, unmarked fij for any of these senses.
kev npau taws (anger/wrath)	(a) God’s own righteous judicial wrath (D4, reused Romans baseline sense); (b) the folk-religious script of capricious ancestor/spirit anger, provoked by ritual error and resolved by appeasement — the very script Hosea 11:8-9 deliberately does NOT follow.	Never let 11:8-9 stand without the explicit “God’s wrath is real and just, but He restrains it because of who He is (Tus Dawb Huv), not because of any ritual appeasement performed by Israel” framing clause. This is the single highest-density fencing requirement in the core passage.
kev hlub family (ahav / chesed / rachamim / emunah)	See D1 matrix row above — four Hebrew terms, one crowded Hmong semantic field of “love/warmth/loyalty/compassion.”	Treat as a closed four-item set (§1 above): ahav = kev hlub (bare); chesed = txoj kev hlub ruaj khov (never bare); rachamim = txoj kev khuv leej; emunah = kev ncaj thiab kev ruaj khov. No item in this set may substitute for another, even in space-constrained contexts (titles, quiz options), mirroring the Romans baseline’s “qualifier is structural, not optional” rule.
tswv-root vocabulary (tus Tswv / tus tswv / Npa-as)	(a) tus Tswv — the fixed divine title (reused Romans baseline, = YHWH/kyrios); (b) tus tswv lowercase common-noun “master/husband” (the ba’ali sense); (c) Npa-as — the rival deity’s proper name.	Capitalization/nominal fixing is not phonetically distinguishable in spoken Hmong, so this fence must be carried by context sentence structure and footnoting, never by relying on the written case distinction alone (since RPA lowercase/uppercase conventions may not survive all delivery formats, e.g. audio curricula). Every Baal-adjacent occurrence of “tswv” vocabulary must be footnoted at first occurrence per chapter.
kevcai-root vocabulary (Txoj Kevcai (ntawm Vajtswv/Mauxes) / kev cai raus dej [Romans baseline, not in Hosea itself but must not be contradicted] / qhaib coj los ua poj niam betrothal customs)	Multiple distinct “customs/rites” senses sharing the root that also underlies kev cai dab qhuas (folk religion).	Continue the Romans baseline’s mandatory-qualifier discipline for Torah (ntawm Vajtswv/Mauxes); extend the same logic to the betrothal phrase, which must always carry the mus ib txhis (forever) qualifier so it is not read as an ordinary bride-price kevcai transaction.
shuv-root pair (rov qab los cuag Vajtswv / kev tig rov qab ntawm Vajtswv)	Positive covenant-return vs. negative apostatic turning-away vs. literal historical-geographic return (to Egypt, 11:5) — three senses from one Hebrew root.	Assign and rigidly maintain three distinct Hmong phrasings (return-to-God / turning-away-from-God / literal-return-to-a-place), and require a teaching note on the underlying wordplay wherever two of the three senses occur within the same chapter (as they do in ch.11 and ch.7).

---

3. True Missing-Vocabulary Gaps (No Existing Hmong Term; Periphrasis/Coinage Required)

These differ from §2 in kind: there is no pre-existing Hmong word doing any relevant work, so a periphrastic phrase must be constructed and then held fixed as if it were a single lexical item.

Gap	Why No Native Term Exists	Constructed Hmong Phrase (per glossary/registry)	Coinage Risk
Chesed as a unitary concept	Hmong kinship ethics distinguish duty owed from affection felt, but has no single term fusing “binding loyalty” + “freely continued despite betrayal” + “warmth.”	txoj kev hlub ruaj khov	Medium — a compound built from two existing words (hlub + ruaj khov) risks being heard as two separable qualities rather than one unified attribute; teaching must present it as a single indivisible concept, not “love that happens to also be steady.”
Berith as two-party binding bond (vs. one-directional promise)	Existing baseline lus cog tseg (promise) is unidirectional-pledge shaped; Hmong customary law has covenant-like plaub ntug dispute-settlements but no single term for an ongoing two-party sacred bond.	phau lus cog tseg (ntawm Vajtswv)	Medium — risk of default collapse back into the baseline promise sense if the qualifier is dropped or abbreviated in space-constrained contexts (headers, quiz stems).
Sheol as a defeated hostile power (vs. venerated ancestral realm)	Hmong cosmology’s land-of-the-dead is a place of ongoing relationship with ancestors via veneration ritual; it is not conceived as an enemy to be militarily/legally defeated.	tub tuag teb (descriptive: “land of the dead”)	High — this is a genuine worldview gap, not solvable by lexical choice alone; requires sustained teaching-note reinforcement every time it appears (only 13:14, but a doctrinally climactic verse cross-referenced to 1 Cor 15:55).
Pequddah — personal, appointed, non-fatalistic “day of reckoning”	Nearest available register is either fatalistic hmoov (forbidden) or a bare generic “someday” — neither carries “appointed by a personal, sovereign, morally-motivated agent.”	descriptive clause “hnub uas Vajtswv yuav txiav txim”	Medium — periphrasis is functional but verbose; must not be shortened in ways that reintroduce ambiguity with ordinary future-tense judgment language elsewhere in the book.
”Bulls of our lips” as praise-replacing-sacrifice idiom (14:2)	No Hmong idiom equates verbal praise with animal sacrifice; a literal rendering is not merely weak but actively misleading (implies real cattle spoken about, not offered instead of cattle).	tej lus qhuas ntawm peb daim tawv ncauj	Medium-High — periphrasis sacrifices the metaphor’s original cattle-imagery entirely in favor of clarity; a supplementary footnote restoring the original “bulls” image for teaching purposes is recommended so learners are not cut off from the wider sacrificial-substitution argument of the book.
Meshuvah as the ironic noun-form of shuv	As above (§2 shuv-root row) — but specifically the noun form naming a settled disposition/character trait of apostasy, for which Hmong has no single-word equivalent to “backslidden-ness” as a named condition rather than an event.	kev tig rov qab ntawm Vajtswv	Low-Medium — periphrasis works but reads as an event-description (“turning away”) rather than a named condition/trait; teaching note should clarify this names Ephraim’s characteristic disposition, not a single lapse.
Yada as covenantal/relational (vs. factual) knowing	Ordinary paub covers factual knowledge only; Hmong has relational-intimacy vocabulary for spouses/kin but it is not conventionally paired with “knowing God.”	kev paub Vajtswv	Medium — risk of silent collapse to the factual sense (paub alone) in careless generation; requires the fixed compound at every occurrence, never abbreviated to bare paub when the covenantal sense is intended.

---

4. Transliteration vs. Paraphrase Decisions

Hosea’s proper-name density (child-names with theological meanings, place-name wordplay, foreign nation names) requires a consistent, documented policy rather than ad hoc choices at each occurrence.

Item	Decision	Rationale
Symbolic child-names (Jezreel, Lo-Ruhamah, Lo-Ammi)	Transliterate + mandatory parenthetical meaning-gloss, every occurrence, identical pairing each time (e.g., Yexeliyees (Vajtswv Tso Tawg / Vajtswv Tseb)).	The names function theologically only through their meanings (judgment reversed to blessing, 1:4-5→2:22-23; 1:6→2:23; 1:9→2:23 and Rom 9:25-26). Transliteration alone would strip the entire rhetorical point; meaning-gloss alone would lose the proper-name identity of Hosea’s actual children. Both together, held identical across both occurrences of each name, preserve the reversal argument.
Baal / Beth-Aven wordplays	Transliterate the proper name; paraphrase the pun’s meaning in a footnote; never attempt a Hmong pun.	As established in §1 (D3) and registry#8-9, #52: Hmong has no lexical mechanism to reproduce a Hebrew consonantal pun. Forcing one would either fail silently (readers miss the wordplay entirely) or produce a strained, distracting invented pun that risks sounding like folk etymology rather than Scripture. Footnoted paraphrase is the only reliable path.
Ordinary proper names (Ephraim, Israel, Jacob, Gomer, Diblaim, Assyria, Egypt, Admah, Zeboiim)	Transliterate only, using existing/established Hmong OT transliteration precedent where confirmable, flagged for confirmation where not.	These names carry no in-text wordplay payload (Jacob’s typological echo in ch.12 is thematic/narrative, not lexical-pun-based, and is handled by teaching note, not by translation choice). Over-glossing every proper name would clutter the text and dilute the reader’s attention from the names that do bear theological freight.
Sheol	Transliterate-adjacent descriptive rendering (tub tuag teb), not a bare loanword.	A bare transliterated loanword (“Seyaus” or similar) would be opaque and would forfeit the chance to immediately signal “realm of the dead” to the hearer; but a literal existing Hmong term risked collision with ancestral-realm veneration cosmology (§3 above), so the descriptive compound is preferred over both a loanword and an unqualified native equivalent.
”Bulls of our lips” (14:2)	Paraphrase the meaning in the main text; transliterate/restore the literal image only in a footnote.	See §3 above — this is the clearest case in Hosea where literal rendering actively misleads rather than merely under-communicating.
Dew / lily / Lebanon-cedar restoration imagery (14:5-7)	Paraphrase with descriptive equivalents; footnote the exotic referents (Lebanon cedars) rather than substituting a “dynamic equivalent” local plant.	Substituting a locally-familiar tree/flower would sever the text’s deliberate echo of Solomonic/royal-garden imagery elsewhere in the Hebrew canon; footnoting preserves both comprehension and canonical intertextual linkage, consistent with the Romans baseline’s handling of the unfamiliar olive tree (11:17-24; note reused analogously here).
Ishi / Ba’ali (2:16-17)	Retain the effect through paraphrase of the two Hmong phrases (kuv tus txiv / kuv tus Npa-as) with a footnote carrying the transliterated Hebrew terms ishi/ba'ali for reference.	Mirrors the Romans baseline’s handling of Abba, Leej Txiv — an Aramaic/Hebrew loanword retained in a footnote for teaching purposes even though the main running text uses the native equivalent.

---

5. Ranked List of Highest-Risk Ambiguities

Ranked by combined severity (doctrinal centrality × collision specificity × likelihood of silent drift in unsupervised generation). All items at rank 1-6 require mandatory human theologian review per the routing rules already established in 08_core_glossary.md §D; items 7-10 require at minimum native-speaker review escalating to theologian review if doctrinal ambiguity is suspected.

1. ruach zenunim (“spirit of whoredom,” 4:12; 5:4) rendered with any ntsuj plig/Vaj Ntsuj Plig vocabulary. Catastrophic-severity if it occurs: would either sound like literal spirit-possession (animistic collision) or, worse, confuse the divine Holy Spirit’s fixed collocation with a metaphor for compulsive idolatry — a direct violation of the single most tightly-guarded rule inherited from the Romans baseline. Permanently forbidden; mandatory descriptive phrasing only.

2. Chesed (txoj kev hlub ruaj khov) losing its qualifier and collapsing into generic kev hlub or reciprocal kwv tij-loyalty logic. This would invert Doctrine 1 at its root — precisely the way koob hmoov inverts grace in the Romans baseline — by making God’s love read as conditional on Israel’s continued faithfulness, the opposite of 11:8-9 and 14:4’s argument.

3. 11:8-9’s wrath/compassion tension flattened into a simple “God changed His mind” statement, or into a folk-religious appeasement-and-resolution narrative. Both failure modes are equally available to unsupervised generation and both neutralize Doctrine 4’s central theological claim (that God’s compassion is self-consistent with His holiness, not a mood swing, and not a response bought by ritual).

4. hu (call, 11:1-2, 7) drifting toward hu plig ritual soul-calling connotation without the mandatory relational/historical anchor clause. High likelihood of silent occurrence because hu is otherwise a completely ordinary, low-alarm verb; the risk is precisely that it will not trigger a flag in casual generation.

5. Baal/Lord title collision (Npa-as vs. tus tswv vs. tus Tswv) blurring in running text without capitalization surviving delivery format (audio, plain-text export, quiz UI). A single dropped distinction here could produce sentences that appear to identify YHWH with Baal or vice versa — an unusually severe failure mode for a book whose central argument is that Israel wrongly did exactly that.

6. Zanah/naaph (harlotry/adultery, chs.1-5) rendered without sign-act/metaphor framing, read as sensational biographical content rather than covenant theology. Severity is doctrinal-register rather than lexical: the words themselves are correctly chosen, but absent the framing clause, Phase 2 content could drift into a tone inconsistent with the required formal/literary register and could be experienced as inappropriate rather than as prophetic theology.

7. “All Israel”-adjacent covenant-restoration language (2:23; Lo-Ammi reversal) generated without the Romans 9:25-26 cross-reference consistency check. Not a Hosea-internal doctrinal risk per se, but a cross-Language-Package consistency failure that would surface only when a future Romans-curriculum module quotes Hosea — must be pre-empted now.

8. 13:14’s redemption-from-Sheol/Death clause generated without the 1 Corinthians 15:55 cross-reference and without distinguishing Sheol from ancestral-realm veneration cosmology. Same cross-curriculum consistency risk as #7, compounded by the §3 worldview gap around Sheol.

9. fij-family terms (worship/sacrifice) used across chs. 2, 4, 6, 8, 14 without the three/four-way disambiguation marker, especially at 6:6 (mercy over sacrifice) and 14:2 (lips as bulls). Risk is comprehension/register rather than doctrinal inversion, but repeated unmarked use across five chapters compounds the risk of habituated, unflagged drift.

10. Betrothal language (2:19-20) losing its mus ib txhis (forever) qualifier and reading as an ordinary bride-price transaction, blurring into ch.3’s deliberately different literal-purchase image. Would flatten two distinct theological images (everlasting covenant pledge vs. costly redemptive purchase) into one, weakening both Doctrine 3 and the Doctrine-1/4 redemption argument that depends on ch.3’s purchase remaining a distinct image.

---

6. Summary Table: Fencing Priorities for Phase 2 Generation

Priority Tier	Items	Required Action Before Publication
Tier 1 — Absolute prohibition	ruach zenunim + any spirit-vocabulary (#1 above)	Automatic rejection of generated content; no automated-review bypass.
Tier 2 — Mandatory structural qualifier, zero exceptions	chesed (#2), wrath/compassion tension framing (#3), Baal/Lord distinction (#5), betrothal forever-qualifier (#10)	Enforced exactly as the Romans baseline enforces hwjchim→Vajtswv hwjchim: qualifier is atomic with the term, never dropped even in space-constrained contexts.
Tier 3 — Mandatory teaching-note/footnote at first occurrence per lesson module	hu (#4), zanah/naaph framing (#6), shuv/meshuvah wordplay, ishi/ba’ali wordplay, Beth-Aven wordplay, fij disambiguation (#9)	Footnote or explicit framing clause required; absence is a publication-blocking defect, not a stylistic preference.
Tier 4 — Cross-Language-Package consistency check	Lo-Ammi/Rom 9:25-26 (#7), Sheol-redemption/1 Cor 15:55 (#8)	Verify exact string match against any existing or future Romans-curriculum treatment before publication.

This tiering is the operative output of Step 8 and should be consulted directly by Phase 2 prompt-construction alongside 12_ai_translation_requirements.md’s validation checklist once that document is extended for Hosea.