Jeremiah — Linguistic Gap Analysis (Hmong Language Package, Baseline v1.0)

0. Purpose and Scope

This document executes PRD Phase 1 Step 8 for the Hmong Jeremiah Language Package. It does three things the glossary (08_core_glossary.md) and term registry (assets/bible_term_registry.json) do not do on their own:

1. Organizes available Hmong vocabulary by doctrine rather than by lexical item, so translators can see at a glance what resources and weaknesses exist for each of the seven assigned doctrines, anchored to the core passage Yelemis 31:31-34.
2. Names the crowded semantic neighborhoods — clusters of Hmong root-words (siab, dab, fij, kevcai, cog lus, cawm/dim, hwjchim, quaj ntsuag) that Jeremiah’s vocabulary repeatedly draws on, and specifies the fencing (disambiguating strategy) required to keep doctrinally distinct concepts from bleeding into one another — as distinct from genuine vocabulary gaps, where no adequate Hmong term exists at all and a transliteration/paraphrase/footnote decision is required.
3. Produces a ranked list of the highest-risk ambiguities in the letter, so Phase 2 generation and human review can be prioritized.

This document does not introduce new permanent renderings; all renderings referenced here are already fixed in assets/bible_term_registry.json (Jeremiah) or inherited from the Romans TM, and any item still marked [TRANSLATOR REVIEW NEEDED] there remains unresolved here as well, by design.

---

1. Doctrine Vocabulary Matrix

For each doctrine: the destination-language (Hmong) resources currently available, their specific weaknesses, and the recommended handling strategy for Phase 2 generation.

1.1 The New Covenant (Critical — core passage doctrine)

Available Hmong Terms	Weakness	Recommended Strategy
Kev Cog Lus Tshiab (new covenant), Kev Cog Lus Qub (old covenant), kev cog lus (covenant, general)	Hmong cog lus customarily denotes a bilateral, revocable pact (betrothal, clan-alliance oath); nothing in ordinary usage prepares a hearer for a covenant unilaterally sustained by God alone after the human party’s breach.	Every occurrence of Kev Cog Lus Tshiab in Yelemis 31 must be taught alongside explicit language that God alone guarantees its continuation (31:35-37’s creation-order guarantee); never let the term stand without this anchor.
Txoj Kevcai (ntawm Vajtswv / Mauxes) (Law)	Mandatory qualifier risk: precisely because 31:33 teaches continuity (the same Law, now internalized), any accidental drop of the qualifier reads as “a new, different law,” inverting the doctrine.	Enforce the qualifier with zero exceptions, including inside quotations of 31:33 itself.
lub siab (heart) + “raug sau rau hauv” (written upon)	siab is already used for will/conscience/emotion across many unrelated contexts (see §3.1 below); a bare mention of “God writing on the siab” risks being heard as a poetic flourish rather than a decisive, once-for-all divine act.	Pair with explicit contrastive teaching: this writing is something only God does (cf. Yelemis 17:9’s despairing verdict on the natural heart), never a human self-improvement achievement — this also protects against re-collapsing into the justification-as-moralism risk already flagged Critical in the Romans TM (justify_reckon_as_righteous).
paub Vajtswv (know the LORD)	Epistemic collapse risk: Hmong paub covers simple factual knowing (“I know that place”) as readily as relational intimacy; 31:34’s point is that everyone will have direct relational knowledge without needing a mediating teacher.	Always render with an explicit relational/intimacy gloss on first occurrence in a lesson; never let paub Vajtswv stand as a bare factual-knowledge claim.
zam txim (forgive iniquity), yuav tsis nco qab ... ntxiv li lawm (remember sin no more)	zam txim already exists in general Hmong Christian usage but must be anchored here to God’s exclusive, sovereign, once-for-all prerogative under this covenant, not ordinary interpersonal pardon; tsis nco qab (literally “does not remember”) risks sounding like divine forgetfulness rather than a sovereign covenantal choice.	Teach tsis nco qab explicitly as legal non-prosecution, not memory failure; cross-reference the Romans TM propitiation_atoning_sacrifice (Critical) so learners connect Jeremiah’s promise to its Christological fulfillment.
Kuv yuav ua nej tus Vajtswv, thiab nej yuav ua kuv haiv neeg (covenant formula)	No inherent weakness in the Hmong wording itself; the risk is inconsistency if this fixed formula is paraphrased differently across lessons (it recurs at Ex 6:7; Yelemis 24:7; 30:22; 31:1,33; 32:38; Ezek 36:28; 37:27; Rev 21:3).	Lock the string verbatim across every cross-reference; treat any variation as a defect.

1.2 Judgment for Covenant Unfaithfulness and Idolatry

Available Hmong Terms	Weakness	Recommended Strategy
dab mlom (idol), kev ua nkauj rau luag dab (spiritual harlotry)	dab mlom is the actual term for spirit-images in Hmong traditional religion — doctrinally correct but carries strong lived-experience weight; kev ua nkauj (harlotry) risks being heard as a private sexual-ethics issue rather than covenant betrayal if the rau luag dab (toward another’s spirits) qualifier is ever dropped.	Never separate the harlotry phrase from its dab-object qualifier; teach explicitly as covenant-fidelity language, cross-referencing marriage as the controlling metaphor (chs. 2-3).
kev npau taws ntawm Vajtswv (wrath of God), lub khob cawv ntawm Vajtswv txoj kev npau taws (cup of wrath)	Same collision already flagged High in the Romans TM: unqualified npau taws reads as capricious ancestor/spirit anger, remediable by ritual appeasement — a risk that recurs at far higher density in Jeremiah’s sustained judgment-oracle genre than anywhere in Romans.	Every occurrence must carry the “righteous, judicial, not appeasable by ritual” teaching note; do not assume one contrast-teaching instance early in the curriculum suffices for the whole book — repeat at each major judgment oracle (chs. 4-6, 21, 25, 36, 42-44).
kev muab menyuam fij rau dab (child sacrifice at Topheth)	Sits inside the crowded fij (sacrifice/offering) neighborhood alongside living sacrifice and Christ’s atonement (see §3.3).	Never allow fij to appear here without the menyuam ... rau dab object phrase; this is condemned atrocity, never assimilated to any positive offering vocabulary.
kev tig rov qab tso Vajtswv tseg (backsliding)	Descriptive compound; risk is under-weighting repeated/habitual apostasy as a single lapse.	Reinforce “habitual, repeated” sense explicitly in teaching notes at each occurrence (chs. 2-3, 8).
Txoj Kevcai (ntawm Vajtswv / Mauxes) reused as the standard against which unfaithfulness is measured	Shares the crowded kevcai root (see §3.2).	Maintain the mandatory qualifier at every occurrence without exception, even in judgment-oracle contexts where space is tight.

1.3 The Weeping Prophet and the Cost of Faithful Ministry

Available Hmong Terms	Weakness	Recommended Strategy
kev quaj ntsuag (weeping/lament)	Strong positive cultural bridge to the Hmong sung-lament (kwv txhiaj) tradition — but this same tradition is entangled with funerary ancestor-soul-guidance ritual, risking a reduction of Jeremiah’s covenant-grief to generic funerary emotion.	Explicit teaching note distinguishing prophetic covenant-lament (grief over sin and coming judgment) from funerary soul-guidance ritual; use at every occurrence (chs. 8, 9, 13, 14, 16, 20, 31), not only on first introduction.
txwv tsis pub quaj ntsuag raws li kevcai (forbidden mourning sign-act, 16:5-7)	Inverse case: withholding the positive lament custom is itself the prophetic sign; risk of the withholding reading as contradicting Jeremiah’s characteristic grief elsewhere.	Frame explicitly as an exceptional sign-act with its own explanatory note, cross-referenced back to the general lament entry.
zoo li muaj ib qho hluavtaws kub nyob hauv kuv pob txha (fire in my bones, 20:9)	No traditional-religion collision; purely comprehension-level vivid imagery.	Use as a positive teaching bridge contrasting involuntary, Spirit-compelled true prophecy against the self-generated speech of false prophets (§1.4).
kev tuavtes ntxeev siab (conspiracy)	Descriptive, no collision.	Use straightforwardly to reinforce the persecution/cost theme (chs. 11, 26, 37-38).
No dedicated term yet for Jeremiah’s personal complaint against God (the “confessions,” e.g. 15:10-18, 20:7-8’s “You have deceived me”)	Gap: existing vocabulary (quaj ntsuag) captures grief but not the confrontational, almost accusatory tone of Jeremiah’s complaint prayers.	Recommend paraphrase rather than a single fixed term: render descriptively each time (e.g., “Yelemis tau hais nws tej lus tu siab thiab nug Vajtswv” — “Jeremiah spoke his grieving words and questioned God”), logged for future TM-revision review rather than a forced permanent single-word coinage.

1.4 False Prophets versus True Prophecy

Available Hmong Terms	Weakness	Recommended Strategy
tus Cev Vajtswv Lus (prophet) vs tus Cev Vajtswv Lus Cuav (false prophet)	The distinguishing element is entirely the Cuav (false/counterfeit) qualifier — a single dropped word erases the entire doctrinal contrast.	Structural enforcement identical to the Romans TM’s Vaj-prefix rule for the Holy Spirit: Cuav is never optional, never abbreviated, in every occurrence (chs. 6, 8, 14, 23, 27-29).
kev saib yaig / kev ua noaj (divination), npau suav (dream, as discernment-test)	Both name real, currently-practiced Hmong shamanic/divinatory categories (txiv neeb consultation) rather than abstractions — strong lived referent but must never be presented as a neutral alternative revelation channel.	Always paired with explicit condemnation-framing; never presented descriptively without a value judgment attached, unlike (e.g.) the neutral historical description used for dab mlom in narrative-only contexts.
kev thaj yeeb used ironically in the “peace, peace” formula (6:14; 8:11)	Same lexical item is Low-risk and positive everywhere else in the Language Package (Romans peace entry, and ordinary Jeremiah restoration contexts, e.g. ch. 29). A single term carrying opposite valences in different passages.	Structural fencing: the ironic instances must be rendered with explicit surrounding negation (“muaj kev thaj yeeb, muaj kev thaj yeeb,” tiamsis tsis muaj kev thaj yeeb li) so the irony cannot be missed, and a footnote must clarify that kev thaj yeeb itself remains a stable, positive term elsewhere.
No permanent Hmong rendering for the massa (“burden”/“oracle”) pun (23:33-40)	Genuine gap: no Hmong homograph carries both “a heavy load” and “a divine utterance” senses simultaneously.	Do not force a coined pun. Use footnote-only handling: translate contextually (as “load/burden” or as “oracle” depending on the specific clause) and add a translator’s note explaining the wordplay exists in Hebrew but cannot be reproduced. This is a permanent design decision, not a pending gap to be closed later.
Yaj Saub — the Hmong culture-hero/seer legend	Forbidden substitution, not an available resource: bare Yaj Saub for nabi is PERMANENTLY FORBIDDEN (per bible_term_registry.json id 5) because it imports an entire alternate folk-messianic narrative that could redirect covenant hope from Vajtswv/Yexus Khetos onto a nativist deliverer-figure.	The descriptive phrase tus Cev Vajtswv Lus must always be used instead; this restriction should be treated with the same permanence as the Romans TM’s koob hmoov-for-grace and hmoov-for-election prohibitions.

1.5 God’s Sovereign Plans for His People

Available Hmong Terms	Weakness	Recommended Strategy
tus Kws Puab Av thiab cov Av Nplaum (potter/clay, ch. 18)	Directly underlies the Election/Sovereignty doctrine already Critical in the Romans TM for its hmoov collision risk; Jeremiah 18:7-10’s explicitly conditional language (the potter changes His plan in response to the clay’s repentance or rebellion) must not be lost.	Preserve the conditional structure explicitly in every retelling — do not compress 18:1-10 into a bare “God shapes people like clay” statement without the response-conditioned reversal, or divine sovereignty will silently swallow human responsibility.
Vajtswv tej Homphiaj (plans/purposes of God, 29:11)	The single most tempting site in the whole book for hmoov (impersonal fate) substitution, since the English “plans for you” idiom sounds naturally fortune-adjacent.	hmoov remains PERMANENTLY FORBIDDEN here, extending the identical Romans-baseline prohibition already fixed for election_choose.
hwjchim (power), always Vajtswv hwjchim	Identical shamanic-power collision already Critical/High in the Romans TM; recurs at even higher density across the oracles-against-the-nations block (chs. 46-51) than anywhere in Romans.	Qualifier is structurally non-negotiable at every single occurrence, including inside long oracle-poetry blocks where the temptation to drop it for meter/brevity will be highest.
tus uas Vajtswv siv ua cuab yeej [TRANSLATOR REVIEW NEEDED] (instrument of judgment — pagan king as God’s “servant”)	Unresolved gap, not merely a weakness: no candidate rendering has been ratified. Reusing qhev (bondservant) bare, as in Romans, would badly mismatch an unbelieving foreign king used merely as a disciplinary tool, since qhev in the Romans baseline is anchored to willing, honored bondservice to Christ.	Do not resolve unilaterally. Retain the flag; escalate to human theologian + native speaker per 12_ai_translation_requirements.md §9 escalation table before any Phase 2 content uses this referent. See ranked list §4, item 4.
rho tawm thiab rhuav tshem, ua kom puas tsuaj thiab ua kom vau, tsim thiab cog (sovereignty formula, 1:10)	Descriptive six-verb formula; risk is purely one of framing (sovereign historical action vs. cosmic fatalism).	Teach explicitly as God’s sovereignty exercised through the prophetic word acting in history, not an impersonal cosmic force — reinforces rather than duplicates the potter/clay fencing above.

1.6 Individual Responsibility

Available Hmong Terms	Weakness	Recommended Strategy
txhua tus yuav ris nws tus kheej lub txim (each will bear his own guilt, 31:29-30)	Directly collides with strong Hmong clan-based corporate/intergenerational liability beliefs (ancestral curses, unresolved ritual debts, unperformed funeral rites bringing trouble on descendants). Two opposite mis-readings are both live risks: (a) hearing this as denying any real corporate solidarity at all (over-correcting the clan worldview), or (b) assimilating it back into fatalistic ancestral-debt logic it is designed to correct.	Requires a dedicated two-sided contrast-teaching note every time this doctrine is taught, not a single passing gloss: affirm real corporate consequences of sin in history (which Jeremiah elsewhere clearly teaches, e.g. the nation suffers exile corporately) while insisting that final covenantal guilt/forgiveness under the New Covenant is decided person-by-person before God.
cov Lekhaj (Rechabites, ch. 35)	Positive point of contact with genuine Hmong clan ancestral-vow-keeping custom — but risks a subtle misreading that ancestor-honor-style vow-fidelity is equivalent to covenant faith toward God.	Teach the Rechabites strictly as an analogy for steadfastness, with an explicit note that the object of fidelity (God’s covenant vs. a clan patriarch’s rule) is not interchangeable.
vam khom neeg / vam khom Vajtswv (trust in man vs. trust in the LORD, 17:5-8)	Clear contrastive idiom; minimal risk.	Use as reinforcing, not primary, vocabulary for this doctrine.
lub siab (uas dag ntxias tshaj plaws) (the deceitful heart, 17:9)	Risk of leaving hearers in unresolved anthropological despair if not paired forward with 31:33.	Never teach 17:9 in isolation from 31:31-34’s promise of divine heart-transformation; the two passages are a matched pair in this Language Package.

1.7 Hope of Restoration after Exile

Available Hmong Terms	Weakness	Recommended Strategy
kev coj rov qab los / kev pub kom vam meej dua (restore the fortunes)	Same hmoov-adjacent temptation as Vajtswv tej Homphiaj above — “restoring fortunes” sounds like reversing bad luck.	hmoov vocabulary PERMANENTLY FORBIDDEN here as well; must be taught as a deliberate, sovereign, covenant-keeping act of God recurring at 30:3,18; 31:23; 32:44; 33:7,11,26 — teach the formula’s recurrence explicitly so learners recognize it as a fixed refrain, not incidental phrasing.
tus Ceg Ncaj Ncees (Righteous Branch), Tus Tswv Yog Peb Txoj Kev Ncaj Ncees (the LORD is our righteousness)	Both must reuse the Romans-fixed ncaj ncees root exactly; any drift severs the intended cross-Testament bridge into the Romans Justification-by-Faith doctrine.	Lock both phrases verbatim; treat as the clearest single lexical bridge in the entire Jeremiah package into the Romans baseline’s Critical-tier justification vocabulary.
cov txiv ncuav qoob zoo / cov txiv ncuav qoob phem (good figs/bad figs, ch. 24)	No collision; comprehension-level fruit imagery.	Use to complicate any simplistic “exile = worst outcome” reading — exile itself is not the judgment’s full measure; disposition toward God after exile is.
cov seem (remnant)	Same term serves two related-but-distinct referents: Jeremiah’s post-judgment survivors who disobediently flee to Egypt (chs. 42-44) vs. Romans 11’s remnant of God’s preserved elect.	Explicit teaching note required distinguishing “having survived judgment” from “being saved” — surviving is not itself salvific (bridges directly into Individual Responsibility, §1.6).
kev txhiv (redemption, kinsman-redeemer, 31:11; 32:6-15)	Strong positive continuity with the Hmong debt-redemption custom already documented as an asset in the Romans TM.	Use Jeremiah’s concrete field-purchase narrative (ch. 32) as the primary illustrative anchor for this term across the whole curriculum, since it gives redemption a vivid narrative referent Romans lacks.

---

2. Missing Vocabulary (Genuine Gaps)

These are contexts where no existing Hmong lexical item, even with qualification, adequately carries the source concept, requiring either footnote-only handling, deliberate paraphrase, or an acknowledged unresolved flag — as distinct from §3’s crowded neighborhoods, where the words exist but need fencing.

Gap	Nature of the Gap	Current Handling	Status
massa pun (“burden”/“oracle,” 23:33-40)	No Hmong homograph spans both senses; a coined pun would be artificial and confusing.	Footnote-only; contextual rendering per clause, no fixed term.	Permanent design decision — will never be resolved into a single term.
Baal/“husband” (ba'al) wordplay (31:32; cf. the deity name at chs. 2, 7, 11, 19, 23, 32)	Hebrew exploits one root for both “husband/master” and the rival god’s name/title; Hmong tus Tswv/tus txiv and the transliterated proper noun Npa-as share no root.	Render YHWH’s self-description plainly as tus Tswv/tus txiv with a translator’s footnote explaining the underlying Hebrew wordplay; the transliteration Npa-as must never be echoed in that clause.	Permanent design decision — see ranked list §4, item 5, for the specific risk of getting this wrong.
Instrument-of-judgment title for Nebuchadnezzar as God’s “servant” (25:9; 27:6)	No Hmong term captures “unbelieving foreign king used as a disciplinary tool” without either (a) dignifying him with qhev’s honored-bondservice connotation, or (b) some other unintended implication.	[TRANSLATOR REVIEW NEEDED] retained, unresolved.	Open gap — escalate before Phase 2 use; see ranked list §4, item 4.
Jeremiah’s confrontational “confession” register (15:10-18; 20:7-18)	Existing lament vocabulary (kev quaj ntsuag) captures grief but not the accusatory, almost legal-complaint tone directed at God Himself.	Paraphrase per-passage rather than a fixed term; logged for future TM-revision consideration.	Open gap, low urgency (comprehension-level, no collision).
”Cutting” a covenant (idiom underlying karath berith, e.g. 31:31; 34:18)	The literal “cutting” ratification idiom (rooted in an animal-sacrifice rite) has no natural Hmong equivalent and risks sounding like unrelated violence if translated literally.	Rendered non-literally as kev cog lus uas kuv tau cog (“the covenant that I made/covenanted”), with the sacrificial-ratification background explained only in a teaching note, never as a literal verb.	Resolved via paraphrase; stable.
A neutral, non-shamanic Hmong term for “true, Spirit-given prophetic insight” as distinct from both tus Cev Vajtswv Lus (the office) and npau suav (mere dream)	No single word; the concept must always be expressed relationally (a message from God) rather than as an inherent personal faculty.	Handled adequately by the existing periphrasis system (§1.4); not a true gap, but worth flagging since translators may be tempted to coin a shortcut term — this should be resisted.	Resolved via existing periphrasis; monitor for drift.

---

3. Crowded Semantic Neighborhoods Requiring Fencing

Unlike §2, these are cases where adequate Hmong vocabulary exists, but a single root or word family is asked to carry multiple, doctrinally distinct senses across the book. Each requires an explicit fencing strategy — a mandatory qualifier, a fixed periphrasis, or a structural separation rule — so the senses do not bleed into one another.

3.1 The siab (liver/heart) neighborhood

lub siab is the seat of will, conscience, emotion, and moral character across an unusually wide range of doctrinally distinct uses in this Language Package:

• lub siab = the heart God writes His law upon (31:33, New Covenant)
• lub siab (uas dag ntxias tshaj plaws) = the deceitful heart (17:9, universal corruption)
• kev txiav lub siab = circumcise the heart (4:4, moral-figurative rite)
• ntxuav lub siab kom tseg qhov phem = wash the heart from evil (4:14)
• Vajtswv lub siab nyiam = will of God (inherited from Romans)
• tej kev paub zoo paub phem hauv lub siab = conscience (inherited from Romans)

Fencing strategy: Each compound must always retain its full fixed periphrasis (never abbreviated to bare siab with the doctrinal sense left to context alone), and lessons introducing the New Covenant’s heart-transformation promise (31:33) must explicitly cross-reference 17:9 and 4:4/4:14 as the problem this promise solves, rather than treating them as unrelated occurrences of the same word.

3.2 The kevcai (custom/rite) neighborhood

Shared root across:

• Txoj Kevcai (ntawm Vajtswv / Mauxes) = the Mosaic Law (Critical, mandatory qualifier)
• kev txiav (raws li Txoj Kevcai Mauxes) = circumcision
• kev cai raus dej = baptism (Romans-inherited, contrast reference)
• kev cai dab qhuas = “religion” generically, literally “customs of spirit-worship” (the collision term named explicitly in the Romans baseline)
• txwv tsis pub quaj ntsuag raws li kevcai = the forbidden-mourning sign-act (16:5-7)

Fencing strategy: The mandatory (ntawm Vajtswv / Mauxes) qualifier on the Law entry is the single load-bearing fence separating divinely-given covenantal Law from generic ancestral-custom-keeping; it must never be dropped, including in citations of 31:33 where the doctrinal point is precisely the Law’s continuity. The forbidden-mourning sign-act’s use of kevcai (customary rite) must be clearly marked as referring to ordinary funerary custom being withheld, not to the Mosaic Law.

3.3 The fij (sacrifice/offering) neighborhood

Three sharply distinct referents share this verb/domain:

• kev muab menyuam fij rau dab = child sacrifice at Topheth — condemned atrocity
• khoom fij uas muaj txoj sia nyob = living sacrifice — ethical devotion (Romans-inherited)
• qhov theej txhoj kev txhaum = Christ’s propitiation — unique, sufficient atonement (Romans-inherited; note this rendering itself avoids the bare fij root, which is part of why it was chosen in Romans)

Fencing strategy: fij alone, unqualified, must never appear without one of its three disambiguating object-phrases. Teaching content addressing Judgment-for-Idolatry (which cites child sacrifice) must explicitly state that this is condemned, not merely differentiated in vocabulary, to prevent any implicit assimilation of the cross into “one more sacrifice among the ones people already offer.”

3.4 The dab (spirit) neighborhood

• dab mlom = idol (High risk, correct-as-is)
• kev ua nkauj rau luag dab = spiritual harlotry / idolatry-as-adultery
• kev muab menyuam fij rau dab = child sacrifice
• Vaj Ntsuj Plig (must retain Vaj prefix) vs. ntsuj plig (bare, human/animistic-soul sense) — inherited Critical/High distinction from Romans, still fully in force throughout Jeremiah wherever the Spirit or human spirit/soul is mentioned

Fencing strategy: No new fencing is required beyond what the Romans TM already establishes for Vaj Ntsuj Plig vs. bare ntsuj plig; this neighborhood is flagged here only to confirm that Jeremiah introduces no new referent requiring the Vaj prefix to be extended or modified, and that all dab-family idolatry vocabulary in Jeremiah remains outside that prefix system entirely (idols/spirits are never referred to with Vaj).

3.5 The cawm / dim (save/escape) and hwjchim (power) neighborhood

Both terms are inherited directly from the Romans TM with their existing High-risk shamanic-collision flags (kev cawm dim parallels ua neeb soul-rescue; bare hwjchim parallels a shaman’s ritual power). Jeremiah does not introduce new senses, but does introduce new high-density contexts: the oracles against the nations (chs. 46-51) use hwjchim repeatedly in rapid succession, and restoration-hope passages (chs. 29-33) repeatedly promise deliverance/rescue.

Fencing strategy: No relaxation of the existing qualifier discipline is permitted merely because of repetition-fatigue risk in long oracle blocks; Vajtswv hwjchim and Christ/God-anchored kev cawm dim phrasing must be maintained at the same density as the source text repeats the underlying concept.

3.6 The cog lus (pledge/covenant) vs. lus cog tseg (promise) neighborhood

• kev cog lus = covenant (Critical, new to Jeremiah) — a formally binding pact
• lus cog tseg = promise (Low, Romans-inherited) — a general solemn pledge

Fencing strategy: These must never be used interchangeably. kev cog lus carries the doctrinal weight of covenant theology (New Covenant, broken covenant, covenant formula) and must be reserved exclusively for that; lus cog tseg remains available for lighter promissory statements (e.g., the promise of return, 29:10) that do not carry full covenant-ratification weight. A drift toward using lus cog tseg as a “simpler” substitute for kev cog lus in simplified teaching materials must be treated as a §3 fencing violation, not a stylistic choice.

3.7 The tig rov qab (return/turn) neighborhood

• tig rov qab los cuag Vajtswv = return/repent (shuv) — a decisive turning act
• hloov siab tshiab = renewal of the mind (Romans-inherited) — ongoing Spirit-wrought transformation
• kev coj rov qab los = restore the fortunes (shuv shevut) — God’s sovereign historical restoration of the exiles

Fencing strategy: All three share the “turning/returning” concept-field but operate at different levels (personal decisive act; ongoing sanctification; God’s sovereign historical action). Teaching content must keep them terminologically distinct even when discussing them together (e.g., a lesson on Yelemis 31 restoration should not blur “the exiles returning to the land” with “a sinner turning back to God” without an explicit bridge-note connecting the two as related but distinct senses).

---

4. Transliteration vs. Paraphrase Decision Log

Item	Decision	Rationale
YHWH (divine name)	Paraphrase/substitution: tus Tswv (never a transliterated or substituted divine name)	Inherited directly from the Romans baseline’s forbidden-substitution rule against NWT-style divine-name transliteration (Yehauvas); extended without exception to all Jeremiah occurrences.
nabi (prophet)	Paraphrase: tus Cev Vajtswv Lus	Transliteration is unnecessary (no distinct loanword tradition exists) and the native culture-hero term Yaj Saub is PERMANENTLY FORBIDDEN due to syncretic-messianism risk (see §1.4).
berith (covenant)	Paraphrase: kev cog lus	A loanword would be doctrinally inert (would not carry any of the relational/legal weight needed) and would evade rather than resolve the bilateral-vs-unilateral comprehension gap; paraphrase allows the needed contrast-teaching to attach directly to native vocabulary.
qesem (divination)	Paraphrase: kev saib yaig / kev ua noaj	Naming the real existing Hmong practice directly (rather than a vague loanword) is necessary for the condemnation to land; a loanword would sound like an imported foreign category rather than the practice hearers actually know.
shabbat (Sabbath)	Established loanword: hnub Xanpataus	Already standard in existing Hmong Christian usage; no comprehension or collision issue, unlike berith or nabi.
Ba'al (deity proper name)	Transliteration: Npa-as	As a proper name (not a common theological term), transliteration is the correct default; risk is confined to the specific wordplay context at 31:32 (husband/Ba’al pun), which is handled separately by NOT echoing this transliteration there (see §2).
Nevukhadretsar (Nebuchadnezzar)	Transliteration, standard existing Hmong Bible rendering (to be confirmed against the existing Hmong OT tradition’s historical-book transliteration conventions)	Proper historical name; no theological content to paraphrase. Logged for confirmation against 05_translation_landscape.md-equivalent source-tradition documentation for Jeremiah, not yet separately produced for this book.
Beth ha-Rekhavim (Rechabites)	Transliteration: cov Lekhaj	Proper clan name; Low risk as a name, though the underlying custom carries Medium risk requiring a teaching note (see §1.6).
massa pun (“burden/oracle”)	Neither transliteration nor forced paraphrase-pun: footnote-only, contextual per-clause rendering	No permanent single rendering is possible without fabricating an artificial Hmong pun; documented as a permanent design decision in §2.
ba'al / “husband” wordplay (31:32)	Paraphrase without pun-reproduction: tus Tswv / tus txiv, with footnote	Reproducing the pun would require echoing the Npa-as transliteration in YHWH’s self-description, directly associating the true God’s covenant-headship with the condemned idol’s name — an unacceptable risk; footnote-only handling preserves the theological point without the wordplay.
karath berith (“cut a covenant” idiom)	Paraphrase, non-literal: kev cog lus uas kuv tau cog	A literal “cutting” verb would suggest unrelated violence with no ratification association in Hmong; paraphrase preserves meaning, sacrificial-ratification background reserved for a teaching note.
avdi Nevukhadretsar (“my servant Nebuchadnezzar”)	Undecided — neither transliteration nor a ratified paraphrase yet exists	See §2 and ranked list below; explicitly retained as [TRANSLATOR REVIEW NEEDED] pending theologian/native-speaker review, per the escalation rule in 12_ai_translation_requirements.md §9.

---

5. Ranked List of Highest-Risk Ambiguities

Ranked by combined severity (doctrinal centrality × collision specificity × likelihood of occurring without a safeguard in ordinary generation).

1. kev cog lus (covenant) heard as an ordinary revocable bilateral pact. Because Hmong customary covenants are mutually binding and voidable by either party’s breach, any Phase 2 content that narrates Israel’s covenant-breaking (11:1-13; 31:32) without equally emphasizing God’s unilateral sustaining of the relationship will implicitly teach that God, too, is now free of obligation — precisely backwards from the New Covenant’s point that God alone bridges the breach. Highest risk because it undermines the core passage (31:31-34) directly.

2. Yaj Saub bleed-through. Even though Yaj Saub is permanently forbidden as a direct rendering, the underlying folk-messianic narrative is culturally salient enough that generated content describing “a coming prophet/deliverer” (e.g., messianic Branch language, 23:5, 33:15) could unintentionally activate the Yaj Saub narrative frame in hearers’ minds even while using correct vocabulary. Requires explicit teaching contrast, not merely correct lexical choice.

3. The Baal/husband wordplay mishandled (31:32). If a future generation pass ever attempts a “more vivid” or “more literal” rendering that reintroduces the Npa-as root into YHWH’s self-description as husband, the result directly associates the true God’s covenant-headship title with the condemned idol — a severe, specific, avoidable doctrinal error. Flagged as high-priority precisely because it is an easy mistake for a well-meaning “improve the poetry” pass to make.

4. avdi Nevukhadretsar left unresolved and used improvisationally. Because no ratified rendering exists, any Phase 2 generation touching chs. 25 or 27 is at risk of an ad hoc translator or AI system defaulting to bare qhev (bondservant) for the pagan king — which would wrongly dignify Nebuchadnezzar with the Romans baseline’s honored, willing bondservice connotation. Must be escalated before any generation touches these chapters.

5. Individual Responsibility (31:29-30) colliding with clan corporate-liability belief. This is a doctrinally central term (bears directly on the core passage’s immediate context) with a specific, named competing framework (ancestral curse/ritual-debt transmission) and no existing single-sentence fix — it requires sustained, repeated two-sided teaching every time it appears, making it structurally harder to safeguard than a simple qualifier-word rule.

6. hmoov (fate) contamination across three separate doctrines simultaneously. Election (potter/clay, ch. 18), Sovereign Plans (Vajtswv tej Homphiaj, 29:11), and Hope of Restoration (kev coj rov qab los, chs. 29-33) all sit adjacent to the same forbidden fatalistic vocabulary. Because this prohibition must be independently remembered and enforced across three different doctrine-teaching contexts rather than one, the aggregate risk of a slip somewhere in the curriculum is higher than any single instance suggests.

7. kev quaj ntsuag (lament) collapsing into funerary ancestor-soul-guidance ritual. The positive cultural bridge that makes this term valuable is the same feature that creates the risk: hearers’ existing associative network for sung lament is saturated with funerary-ritual function, and a single missed teaching note anywhere in the curriculum’s seven weeping/lament-touching chapters (8, 9, 13, 14, 16, 20, 31) could quietly re-import that ritual function.

8. The fij (sacrifice) three-way crowding (child sacrifice / living sacrifice / atonement). High combined risk because all three share a single verb root and appear across doctrines that are otherwise unrelated (Judgment for Idolatry vs. Christian Ethics [contrast reference] vs. Christ’s Atonement [contrast reference]); a lesson that discusses more than one of these in the same unit (plausible, since Topheth condemnation naturally invites a contrast with true sacrifice) is at elevated risk of blurring them if the mandatory object-phrases are not independently verified for each mention.

9. “Peace, peace” ironic reuse (6:14; 8:11) read as literal. Because kev thaj yeeb is Low-risk and positive in the vast majority of its occurrences (including elsewhere in Jeremiah, e.g. ch. 29’s restoration promise), the ironic instances are structurally the least self-signaling high-risk items in the book — nothing about the individual word warns a generation system to add the negation-and-footnote treatment unless the specific verse context is recognized.

10. cov seem (remnant) conflating “survived judgment” with “saved.” Jeremiah’s own narrative undercuts a naive reading of this term (the post-587 BC remnant disobediently flees to Egypt, chs. 42-44, and is judged again), but a generation pass focused only on the positive Romans 11 remnant sense could import an assumption of automatic future blessing onto every Jeremiah remnant reference, contradicting the book’s own narrative arc.

---

6. Cross-References and Governing Note

This analysis presupposes and must be read together with:

• analysis/08_core_glossary.md (Jeremiah core glossary, all row numbers referenced above)
• assets/bible_term_registry.json (Jeremiah term registry, all id numbers referenced above)
• assets/translation_memory.json and assets/bible_term_registry.json (Romans baseline, source of all “Inherited” terms and their pre-existing fencing rules)
• analysis/12_ai_translation_requirements.md (Romans baseline Phase 2 enforcement rules, whose qualifier-enforcement, forbidden-substitution, and escalation mechanisms extend automatically to every item flagged in this document)

No rendering, fencing strategy, or transliteration/paraphrase decision recorded here may be treated as finalized cross-curriculum-permanent TM policy until it passes the same formal TM-revision review process specified for “New (Jeremiah)” glossary rows. Items explicitly marked unresolved in this document (massa pun handling excepted, which is a permanent design decision) must not be silently resolved by any downstream Phase 2/3 generation process.