Linguistic Gap Analysis
Linguistic Gap Analysis: Ezekiel (English → Bengali)
0. Purpose and Method
This analysis identifies, for the full text of Ezekiel (chapters 1–48), where the destination language (Bengali) either (a) lacks a ready term for a source concept, (b) has a term but that term is already occupied by a competing Hindu, Muslim, or generic-secular sense that must be fenced off, or (c) requires a transliteration vs. paraphrase decision because no native Bengali lexeme carries the required theological freight. It extends the Romans-baseline translation_memory.json and is the working paper behind the Ezekiel entries already fixed in 08_core_glossary.md and bible_term_registry.json. Where this document and those files might appear to differ in emphasis, the registry files are authoritative; this document records the reasoning and residual risk behind those choices, and ranks what remains unresolved.
Full-book coverage confirmation is provided in §5. The core passage (37:1-14) is the theological anchor for illustrating the single hardest gap in the book — the polysemy of רוּחַ (ruach) — but every chapter has been reviewed for vocabulary load.
1. Doctrine Vocabulary Matrix
For each of the seven curriculum doctrines: available Bengali terms, their weaknesses, and the recommended strategy (which may reuse a Romans-baseline term, coin a new fixed phrase, or require a standing translator note).
1.1 The Glory and Sovereignty of God
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| כָּבוֹד (glory) | মহিমা (baseline, reused) | Low weakness; established and safe. Ezekiel’s added burden is not lexical but narrative — মহিমা must be tracked as a moving, locatable entity across chs. 1, 8-11, 43-44, which Bengali’s language of presence/departure (চলে যাওয়া / ফিরে আসা) handles adequately if kept consistent. | Reuse মহিমা exactly; enforce identical verb-pairing for “departs”/“returns” across all four narrative beats. |
| אֲדֹנָי יְהוִה (Lord GOD) | প্রভু সদাপ্রভু (new fixed compound) | Bengali has no native compound title combining a generic lordship-title with the covenant name; the two halves (প্রভু, সদাপ্রভু) exist separately in tradition but were never obligatorily paired at ~217 occurrences. Risk of translator fatigue producing drift (some occurrences rendered just সদাপ্রভু, losing the double emphasis). | Fixed, non-negotiable compound at every occurrence; flag any deviation for automated QA (string-match check), not just doctrinal review. |
| יְהוָה (YHVH) | সদাপ্রভু | None significant; this is the settled Bengali Protestant convention. Risk is only conflation with generic প্রভু by inattentive translators. | Reuse exactly; never substitute bare প্রভু or ঈশ্বর where the Hebrew specifically has the Tetragrammaton. |
| Recognition formula (“…know that I am the LORD”) | তখন তারা জানবে যে আমিই সদাপ্রভু | No structural gap; risk is purely one of phrasal drift across 60+ occurrences with slightly varied English source phrasing (“you/they/nations shall know”). | Fix the phrase skeleton; permit only pronoun substitution (তারা/তুমি/আমরা) while keeping আমিই সদাপ্রভু invariant. |
| Messenger formula | প্রভু সদাপ্রভু এই কথা বলেন | Same drift risk as above, compounded by the sheer frequency (~120x) inviting translator paraphrase for readability. | Treat as a liturgical refrain, not ordinary prose; explicitly instruct translators/AI systems not to vary it for “natural flow." |
| "for my name’s sake” | আমার নামের জন্য | No lexical gap. Theological risk: Bengali readers may hear this as God acting for his own pride/honor in the shame-avoidance sense common to South Asian honor culture, rather than covenant-faithfulness to his own revealed character. | Reuse exactly; attach teaching note distinguishing divine self-consistency from a human honor/shame transaction, so judgment and restoration are read as flowing from the same unchanging character, not petty face-saving. |
1.2 Individual Responsibility for Sin (Ezekiel 18)
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| נֶפֶש (nephesh — the soul/person who sins and dies) | প্রাণ (recommended) vs. আত্মা (rejected as primary) | Crowded neighborhood. আত্মা in Bengali religious usage (both Hindu and folk-Islamic) trends toward an indestructible inner essence — closer to Sanskrit ātman — which contradicts Ezekiel 18’s entire argument that the whole person, nephesh, is mortal and can die for sin. | Use প্রাণ (or ব্যক্তি where a personal referent, not a philosophical one, is clearer); reserve আত্মা for the OT Spirit-of-God and NT Holy-Spirit vocabulary elsewhere, never for নেফেশ. Mandatory translator note at first Ezek 18 occurrence. |
| צְדָקָה / צַדִּיק (righteous conduct) | ধার্মিকতা / ধার্মিক (baseline, reused) | Doctrinal register collision, not lexical gap: Romans fixed ধার্মিকতা against a strictly forensic, faith-received sense; Ezekiel 18 uses the same root for observable, case-by-case covenantal conduct with temporal life/death consequence. No alternative Bengali word avoids this without the far worse risk of ধর্ম (already forbidden). | Reuse the identical word; attach a mandatory teaching note at every Ezek 18 occurrence framing this as covenantal-behavioral righteousness under the Old Covenant, not a rival soteriology to Romans’ forensic righteousness. |
| שׁוּב (turn/repent) | ফিরে আসা | No major gap; must stay consistent with whatever repentance vocabulary is fixed for Gospels/Acts in this curriculum family, since both trace to the same root movement (turning from sin toward God). | Reuse consistently; cross-check against Gospel/Acts glossary once produced. |
| צֹפֶה (watchman) | প্রহরী | Bengali প্রহরী defaults to a night-guard/security-guard register in modern usage (a hired, low-status role), risking under-translation of the personal, prophetic, blood-guilt weight Ezekiel assigns the office (3:17-21; 33:1-9). | Reuse প্রহরী but pair with explicit framing language (“তার রক্তের দায়” / “his blood is on his own head” idiom) at first occurrence to establish the gravity, then allow the term alone to carry it by ch. 33. |
| הָעָוֹן (guilt/iniquity as accrued liability) | পাপভার | No Bengali single word natively distinguishes “sin as act” from “guilt as accrued liability owed.” পাপভার (a compound, “sin-burden/-load”) is a constructed solution, not an inherited Bengali Bible term. | Adopt পাপভার as the fixed rendering; flag as a coined compound, not an established term, so reviewers do not assume prior Bible-translation precedent exists for it. |
1.3 The New Heart and New Spirit
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| לֵב חָדָש / רוּחַ חֲדָשָׁה (new heart/new spirit) | নতুন হৃদয় ও নতুন আত্মা | No lexical gap for হৃদয়. Gap is grammatical-mood: Bengali imperative and promissory/future forms are distinguishable, but a careless flattening (rendering both 18:31 and 36:26 as simple statements) would erase the human-responsibility-vs-divine-gift tension central to the doctrine. | Fix the noun phrase identically; require verb-mood distinction be preserved at the clause level (imperative at 18:31 vs. promissory “আমি দেব” at 36:26). |
| לֵב הָאֶבֶן / לֵב בָּשָׂר (heart of stone/flesh) | পাথরের হৃদয় / মাংসের হৃদয় | No gap; risk is abstraction drift — translators softening the vivid tactile image into “hard heart / soft heart” for smoother Bengali prose, which loses the concrete stone-vs-flesh contrast that makes the point (a heart made of the wrong substance needs replacing, not persuading). | Preserve literal tactile nouns; forbid abstraction to generic hardness/softness vocabulary. |
| רוּחַ / רוּחִי (Spirit of God, OT) | আত্মা / আমার আত্মা | Theological-continuity gap, not lexical. Bengali has no independent way to mark “the same divine Person the NT later calls পবিত্র আত্মা, but not yet under that adjective.” Using bare আত্মা risks flattening to a generic “spirit” (ghost/animating-force sense common in Bengali folk religion); using পবিত্র আত্মা risks retrojecting NT Trinitarian precision onto OT wording that does not make it explicit. | Use আমার আত্মা (first person, personal, God’s own possessed Spirit) with a standing continuity note, never identifying the two terms outright nor treating them as unrelated. |
| רוּחַ within Ezek 37 specifically (breath/wind/Spirit in one pericope) | নিঃশ্বাস (breath) / বায়ু (wind) / আমার আত্মা (Spirit) | The single largest translation gap in the book. Hebrew’s one word (רוּחַ) carries all three senses within eleven verses; Bengali has no equivalent single word spanning all three registers. Any lexical choice necessarily fragments a deliberate wordplay. | See §4 (Ranked Ambiguities) for full treatment; mandatory translator note in all ch. 37 teaching material naming the lost wordplay explicitly, with all three Bengali words glossed side by side at least once in the study notes even though the running text must use different words per verse. |
1.4 Judgment on Jerusalem and the Nations
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| תּוֹעֵבָה (abomination) | ঘৃণ্য কাজ | No structural gap; risk is register flattening across ~40 occurrences — translators softening intensity for readability, undercutting Ezekiel’s rhetorical escalation. | Fix at a uniformly strong register; forbid softer synonyms (e.g., ভুল কাজ, “wrongdoing”) anywhere in the book. |
| גִּלּוּלִים (idols, contemptuous) | প্রতিমা | Severe crowded-neighborhood risk. প্রতিমা is the ordinary, warmly-regarded Bengali word for the worshipped images at Durga Puja and other major Hindu festivals — arguably the single most emotionally loaded religious-vocabulary collision in this entire curriculum. | No euphemistic substitute exists without losing referential accuracy (Ezekiel does mean literal cultic images). Retain প্রতিমা but mandate paired teaching notes at every major cluster (chs. 6, 8, 14, 16, 20, 23) framing this as God’s own verdict on Israel’s idols, never as a general commentary on any contemporary neighbor’s religious practice. |
| בָּמָה (high place) | উচ্চস্থান | No gap; ensure it is never confused with legitimate worship-site vocabulary. | Reuse উচ্চস্থান; keep terminologically distinct from মন্দির (reserved for the legitimate temple). |
| זָנָה (harlotry/covenant unfaithfulness) | ব্যভিচার করা | No lexical gap, but tonal risk: honor-shame sensitivity in Bengali Hindu and Muslim social contexts makes this metaphor land with unusual force. | Retain formal Bible register; forbid colloquial vocabulary; teaching notes must clarify the metaphor’s covenantal target (Jerusalem’s unfaithfulness to YHWH) to prevent literalistic or voyeuristic misreading. |
| תַּנִּין (sea monster, applied to Pharaoh) | সামুদ্রিক দানব / কুমির | Crowded neighborhood via a different vector: Bengal’s living Manasa Devi folk-serpent cult makes any নাগ-adjacent vocabulary theologically dangerous, evoking a real, worshipped folk deity rather than a defeated pagan chaos-symbol. | Use সামুদ্রিক দানব or কুমির (concrete crocodile, matching the Nile referent at 29:3); explicitly forbid নাগ, সর্প-দেবতা, or any serpent-deity-adjacent vocabulary. |
| גּוֹיִם (nations under judgment) | জাতিগণ (recommended) vs. পরজাতি (baseline, rejected here) | Not a lexical gap but a cross-document contamination risk: reusing the Romans baseline’s পরজাতি (positive, gospel-inclusion sense) for Ezekiel’s judged geopolitical nations would blur two opposite theological postures toward “the nations” within the same curriculum family. | Use জাতি/জাতিগণ exclusively in Ezekiel 25-32, 35, 38-39; reserve পরজাতি strictly for NT ἔθνη contexts. |
1.5 Restoration and the Valley of Dry Bones (core passage doctrine)
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| עֲצָמוֹת יְבֵשׁוֹת (dry bones) / קִבְרוֹת (graves) | শুষ্ক অস্থি / কবর | No lexical gap. Doctrinal risk: readers may over-identify this vision with the NT doctrine of bodily resurrection (পুনরুত্থান) as though Ezekiel 37 is a resurrection text in the Pauline/Gospel sense, rather than a corporate-national restoration metaphor that later informs, but is not identical to, that doctrine. | Use পুনরুত্থান-family vocabulary or periphrasis (তোলা/উঠানো, “raise up”) that resonates with but does not assert flat identity to the NT doctrine; NEVER পুনর্জন্ম. Mandatory teaching note on the typological (not identical) relationship. |
| כָּל־בֵּית יִשְׂרָאֵל (whole house of Israel) | সমস্ত ইস্রায়েল-বংশ | No gap; risk is purely a doctrinal-frame one — the same audience addressed individually in ch. 18 is addressed corporately here, and Bengali readers must not import ch. 18’s individual-responsibility frame into ch. 37’s national-restoration frame or vice versa. | Fix rendering; add cross-reference note distinguishing corporate (ch. 37) from individual (ch. 18) address to the same people. |
| תִּקְוָה (hope, here: lost hope) | আশা | No gap in the noun; risk is under-translating the intensity of אָבְדָה (“perished/is lost”) — Bengali আশা হারিয়ে গেছে can be said of any minor disappointment. | Use strong verb pairing (আশা সম্পূর্ণ শেষ হয়ে গেছে / সব আশা হারিয়ে গেছে) to preserve the totality of the despair the vision directly answers. |
| חַיִל גָּדוֹל מְאֹד (exceedingly great army) | মহাবিশাল সৈন্যদল | No gap; risk is flattening to a generic crowd/multitude, losing the martial vigor Ezekiel intends (a mobilized, capable nation, not merely a large population). | Preserve সৈন্যদল (army/troop), not জনতা (crowd) or লোকজন (people). |
| בְּרִית שָׁלוֹם / בְּרִית עוֹלָם (covenant of peace / everlasting covenant) | শান্তির নিয়ম / চিরস্থায়ী নিয়ম | No gap; both are compounds of already-fixed baseline terms. Risk is synonym substitution by translators seeking variety (e.g., substituting অনন্তকালীন for চিরস্থায়ী inconsistently). | Fix both compounds exactly; treat as closed phrases, not free compositional Bengali. |
| עַבְדִּי דָוִיד (my servant David) | আমার দাস দায়ূদ | No lexical gap; doctrinal risk that Bengali readers (especially with limited OT background) may expect the literal historical David to be resurrected and installed as ruler. | Mandatory teaching note connecting this typological “David” to Christ’s fulfillment of the Davidic covenant, paralleling the Romans-baseline দায়ূদের বংশ doctrine. |
1.6 God’s Presence Departing and Returning
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| שָׁכַן (to dwell, permanently) | বসবাস করা | Bengali বসবাস করা is a common, low-register verb used for any person or animal taking up residence anywhere; it carries none of the covenantal, tabernacling weight the Hebrew root שָׁכַן/מִשְׁכָּן carries as the fixed technical term behind “tabernacle.” This is a genuine missing-register gap: Bengali has no verb reserved exclusively for God’s own indwelling presence, parallel to how שָׁכַן is reserved in Hebrew. | Accept বসবাস করা as the practical solution (no better single verb exists) but compensate structurally: attach it consistently to the fixed phrase আমার [সিংহাসনের] স্থানে / তাদের মাঝে, and require a teaching note at 37:26-28 and 43:7 explaining that this ordinary-sounding verb bears the full theological weight of God’s own unbroken, permanent, covenantal presence — the resolution of chs. 8-11’s departure. |
| קִנְאָה (covenant jealousy) | ঈর্ষা | Bengali ঈর্ষা is near-uniformly negative in ordinary usage — petty envy, a vice condemned elsewhere in Scripture itself (Galatians-family vocabulary, if reused across this curriculum family). Using it unqualified for God’s own jealousy risks readers hearing God accused of a sin. | Reuse ঈর্ষা (matching existing Bengali Bible convention for Exodus 20’s “jealous God”) but require a standing translator note at every occurrence (chs. 8, 16, 23, 36, 38) distinguishing righteous covenant-exclusivity jealousy from sinful personal envy. |
| מְקוֹם כִּסְאִי (the place of my throne) | আমার সিংহাসনের স্থান | No gap; risk is flattening to generic “holy place” for smoother prose, which would lose the explicit royal/throne-room claim. | Preserve সিংহাসন (throne) explicitly; forbid substitution with পবিত্রস্থান alone. |
| יְהוָה שָׁמָּה (YHVH Shammah) | সদাপ্রভু-শাম্মা | Proper name with embedded meaning; Bengali has no native mechanism for a name that is simultaneously a sentence. | Transliterate as a proper name (সদাপ্রভু-শাম্মা) and gloss immediately in-text or in a footnote (“সদাপ্রভু সেখানে আছেন”); treat with Romans 1:16-17-level fixed-consistency discipline, since this closes the book’s central plot (Presence departs → Presence returns → the city itself is renamed for that presence). |
1.7 The Vision of the New Temple
| Concept | Available Bengali term(s) | Weakness | Recommended strategy |
|---|---|---|---|
| בַּיִת / מִקְדָּשׁ (temple, literal OT building) | মন্দির (with mandatory divine-name qualifier: সদাপ্রভুর মন্দির) | Apparent but not actual collision with the Romans baseline’s rejection of মন্দির for “church.” The referents are entirely different (a literal building vs. the NT ekklēsia-as-community), so this is a false-alarm gap that nonetheless requires explicit documentation so downstream translators and reviewers do not mistakenly “correct” মন্দির here to মণ্ডলী or vice versa. | Retain মন্দির for the literal OT temple, always divine-name-qualified; document explicitly (as here) that this does not reopen or contradict the baseline’s church-vocabulary decision. |
| טָמֵא / טָהוֹר (unclean/clean, ritual) | অশুচি / শুচি | No gap; this is the one context in the entire curriculum where শুচি-family vocabulary is textually correct, since Ezekiel’s own text is describing literal ceremonial purity categories the priests must teach (44:23) — contrast with পবিত্র (holy), which must never be rendered with শুচি elsewhere. | Use অশুচি/শুচি only within chs. 40-46 ritual-purity material; reinforce via teaching note that this is the sole authorized exception to the general পবিত্র ≠ শুচি rule. |
| עוֹלָה / חַטָּאת / מִנְחָה (burnt/sin/grain offering) | হোমবলি / পাপার্থক বলি / ভক্ষ্য নৈবেদ্য | Minor collision: নৈবেদ্য is also standard vocabulary for a Hindu puja food-offering. Established Bengali Bible tradition already uses it for OT offerings, so this is a low-residual, accepted risk rather than a genuine gap. | Retain established sacrificial vocabulary; brief teaching note sufficient, no substitution needed. |
| מְקוֹם כִּסְאִי, river/trees (ch. 47) | মন্দিরের চৌকাঠের নিচ থেকে প্রবাহিত জল; ফলদায়ক গাছ ও আরোগ্যদায়ক পাতা | No lexical gap; risk is typological — losing the Eden→Ezekiel 47→Revelation 22 chain if translated as an isolated nature description. | Mandatory teaching note tracing the chain at first occurrence. |
| נָשִׂיא (prince, deliberately not king) | রাজপ্রধান | Bengali has রাজা (king) and একাধিক words for “prince” (রাজপুত্র, যুবরাজ), all of which default to hereditary royal son — the wrong sense, since Ezekiel’s נָשִׂיא is a deliberately demoted, subordinate ruling title distinguishing this figure from מֶלֶך. This is a genuine missing-precision gap: no existing Bengali word carries “ruler subordinate to a greater king” without either overclaiming royal lineage (রাজপুত্র) or sounding like a bureaucratic administrator with no royal connotation at all (অধ্যক্ষ). | Adopt রাজপ্রধান as a semi-coined compound (রাজ- + প্রধান, “chief/head” without “-son” lineage claim); flag as a constructed solution requiring a first-occurrence teaching note (ch. 34) establishing the deliberate demotion from “king.” |
2. Missing Vocabulary Inventory
Terms with no inherited Bengali Bible-translation precedent, requiring either a newly coined compound or a fixed periphrasis. These differ from “crowded neighborhood” terms (§3) in that the problem is absence, not competing occupancy.
| Gap | Why it is missing | Solution adopted | Residual risk |
|---|---|---|---|
| אֲדֹנָי יְהוִה as an obligatory ~217x compound | Bengali tradition has both halves (প্রভু, সদাপ্রভু) separately but never fixed them as an invariant compound title at this frequency. | প্রভু সদাপ্রভু, fixed. | Translator fatigue across 217 occurrences; requires automated string-match QA, not just spot review. |
| Recognition/messenger/covenant formulas | English and Bengali Bible tradition both have prophetic-formula vocabulary in general, but Ezekiel’s near-liturgical repetition at this density has no prior fixed-phrase precedent in Bengali translation history to draw on directly. | Three fixed phrases established in §1.1/§1.5 and the registry. | Phrase drift under paraphrase pressure; treat as non-negotiable refrains. |
| עָוֹן as distinct from פֶּשַׁע/חַטָּאת (accrued guilt-liability vs. the act of sin) | Bengali পাপ covers the general moral-transgression field but has no separate lexeme for “guilt owed/accrued over time,” the specific sense needed for ch. 4’s day-count symbolism. | পাপভার (coined compound: “sin-load/-burden”). | Newly constructed term; must be introduced with a clear first-occurrence gloss so it reads as intentional, not as an error or unfamiliar neologism. |
| נָשִׂיא (subordinate “prince,” not king) | See §1.7. | রাজপ্রধান (coined compound). | Same fatigue/consistency risk as any coined term; must not silently revert to রাজা in later chapters (40-48) where the temptation to normalize increases. |
| שָׁכַן’s covenantal register (vs. ordinary “reside”) | See §1.6. | বসবাস করা + structural compensation via fixed phrasing and teaching notes. | The single largest “quality register” gap in the doctrine matrix — a common verb is being asked to carry uncommon theological weight; every occurrence needs a support note, not just the first. |
| יְהוָה שָׁמָּה as a meaning-bearing proper name | Bengali (like most languages) does not have a native convention for treating a full clause as a place-name. | Transliterate + gloss (সদাপ্রভু-শাম্মা / “সদাপ্রভু সেখানে আছেন”). | If glossed only once (e.g., only at first mention in study notes but not restated at 48:35 itself), the thesis-closing force of the book’s final verse could be lost on readers encountering that chapter in isolation. |
3. Crowded Semantic Neighborhoods Requiring Fencing
Terms where a serviceable, even necessary, Bengali word already carries strong competing associations from Bengali Hindu tradition, Bengali Muslim tradition, or modern secular usage, such that the term must be used but explicitly fenced with teaching notes rather than replaced (replacement being unavailable or worse).
| Bengali term | Doctrinal use in Ezekiel | Competing association | Fencing strategy |
|---|---|---|---|
| প্রতিমা | idols (গִּלּוּלִים), chs. 6, 8, 14, 16, 20, 22, 23, 36, 37, 44 | The central devotional image of Durga Puja and other major Bengali Hindu festivals — the single highest-intensity cultural collision term in this curriculum. | No substitution possible without inaccuracy. Fence with mandatory teaching notes at every major cluster framing the referent precisely as Israel’s own ancient cultic objects under God’s specific verdict, never as commentary on any living neighbor’s practice. Especially critical at ch. 14’s “idols in the heart,” where the internalized sense could otherwise read as a blanket accusation against Hindu-background readers or neighbors. |
| আত্মা | Multiple competing uses across the curriculum: (a) spirit_of_god_ot/“my Spirit” (High), (b) risk of bleeding into নেফেশ/soul (fenced off, §1.2), (c) NT পবিত্র আত্মা continuity reference | Vedantic ātman (indestructible soul/self); also generic “ghost/spirit” in Bengali folk usage. | Fence by strict contextual reservation: আত্মা only for God’s own Spirit (with আমার for possession) and NT continuity references; never for নেফেশ (use প্রাণ); never for ordinary “breath” or “wind” senses within Ezek 37 (use নিঃশ্বাস/বায়ু). |
| ঈর্ষা | covenant jealousy (קִנְאָה), chs. 8, 16, 23, 36, 38 | Ordinary Bengali envy/spite, a listed vice. | Fence with standing translator note at every occurrence; consider a modifier phrase (গভীর, প্রবল, “deep/intense”) to visually mark the elevated sense even within running text. |
| ধার্মিকতা | Ezekiel 18’s conduct-based righteousness | Romans baseline’s forensic, faith-received righteousness (already fixed Critical-risk term) | Fence via mandatory teaching note contrasting covenantal-behavioral (Ezekiel) vs. forensic (Romans) senses at every Ezek 18 occurrence, so the two doctrines inform rather than contradict one another in a reader’s mind. |
| জাতিগণ vs. পরজাতি | Ezekiel’s judged geopolitical nations vs. Romans’ gospel-included Gentiles | Same root concept (non-Israelite peoples), opposite theological posture (judgment target vs. mission field) | Fence by document-level segregation: জাতিগণ exclusively within Ezekiel 25-32/35/38-39; পরজাতি exclusively within NT curriculum materials. Cross-reference note recommended at first occurrence of each explaining the deliberate split. |
| মন্দির | Literal OT temple (Ezekiel) vs. rejected for “church” (Romans baseline) | Modern Bengali default sense: any Hindu temple | Fence by mandatory divine-name qualifier (সদাপ্রভুর মন্দির), never left bare; documented in §1.7 as a false-alarm collision (different referents), but the fencing discipline (qualifier) still must be enforced to avoid casual-reader confusion. |
| সামুদ্রিক দানব / (explicitly not নাগ) | Chaos-monster imagery for Pharaoh (תַּנִּין) | Bengal’s living Manasa Devi folk-serpent cult | Fence by exclusion: forbid the entire নাগ/সর্প-দেবতা word-family outright rather than attempting a positive teaching note, since any serpent-deity-adjacent vocabulary here would actively evoke a still-practiced folk religion, not merely an ancient one. |
| নৈবেদ্য | Grain/food offering (מִנְחָה), chs. 45-46 | Standard Hindu puja food-offering vocabulary | Lower-intensity fencing (brief note only); already accepted in established Bengali Bible tradition and lower pastoral stakes than প্রতিমা. |
4. Transliteration vs. Paraphrase Decisions
| Term | Decision | Rationale |
|---|---|---|
| Gog and Magog (גּוֹג וּמָגוֹג) | Transliterate (গোগ ও মাগোগ) | Proper names with no recoverable Bengali meaning-equivalent; any attempt at semantic translation would fabricate false etymological content. Requires a teaching note distinguishing Ezekiel’s near-historical referent from Revelation 20:8’s later reuse of the same names. |
| YHVH Shammah (יְהוָה שָׁמָּה) | Transliterate + gloss (সদাপ্রভু-শাম্মা, glossed as “সদাপ্রভু সেখানে আছেন”) | A meaning-bearing proper name; transliteration preserves its status as a name (the city’s new identity), while the obligatory gloss preserves the theological payload for readers who would otherwise miss that this is a full sentence functioning as a place-name. |
| Son of man address (בֶּן־אָדָם) | Paraphrase, not transliteration nor reuse of NT title (হে মানব, not মনুষ্যপুত্র) | No proper name is involved; the risk here runs the opposite direction from most entries in this table — the danger is a false-positive borrowing from existing Bengali NT vocabulary (মনুষ্যপুত্র, fixed for Christ’s Danielic title) rather than any transliteration temptation. Deliberately fenced apart with its own paraphrase. |
| Cherubim (כְּרוּב) | Transliterate (established: করূব) | Already an established Bengali OT technical term; no paraphrase (e.g., generic দূত/“angel”) is acceptable since it would erase the specific throne-bearing imagery. |
| Abba-class Aramaic/Hebrew terms of intimacy | Not applicable in Ezekiel (no Aramaic terms of this kind occur); noted for completeness against the Romans baseline’s আব্বা precedent, which does not extend to this book. | — |
| tav mark (תָּו) | Paraphrase (চিহ্ন, “mark/sign”) | No Bengali transliteration convention exists for this single Hebrew letter-as-symbol; a generic but accurate paraphrase avoids inventing a false technical term, while a teaching note handles the typological echo to Revelation 7 without asserting identity. |
| berith shalom / berith olam (compounds) | Paraphrase using fixed baseline components (শান্তির নিয়ম / চিরস্থায়ী নিয়ম) | These are ordinary Hebrew nouns in compound, not proper names; paraphrase using already-fixed baseline vocabulary (নিয়ম, শান্তি) is both accurate and consistency-preserving across the curriculum. |
| nasi (prince, subordinate ruler) | Paraphrase via coined compound (রাজপ্রধান) | Not a proper name; the gap is semantic precision, not name-preservation, so paraphrase (with the coining flagged in §2) is correct rather than transliteration. |
| Sons of Zadok (בְּנֵי צָדוֹק) | Transliterate the proper name, paraphrase the relation (সাদোকের সন্তানগণ) | Standard proper-name handling; the doctrinal point (gradated priestly access) rides on context, not on the name itself. |
General rule derived: transliterate where (a) the Hebrew term is a proper name (Gog, Magog, Zadok, Shammah as part of YHVH Shammah, cherubim as an established technical loan) or (b) no accurate Bengali semantic equivalent exists at all. Paraphrase — using either existing fixed baseline vocabulary or, where necessary, a clearly flagged coined compound — everywhere else. Never transliterate a common noun merely because the underlying theology is difficult (e.g., נֶפֶש, רוּחַ, נָשִׂיא are all paraphrased, never transliterated, despite being doctrinally loaded).
5. Full-Book Coverage Confirmation
Every chapter of Ezekiel has been reviewed for vocabulary load. Chapters not separately itemized in §1-4 above contributed no terms beyond those already fixed and are recorded here explicitly per the full-coverage mandate.
| Chapters | Contribution | Status |
|---|---|---|
| 1 | Inaugural theophany: glory (מהিমা), cherubim (করূব), wind/storm sense of רוּחַ (বায়ু), hand of the LORD (সদাপ্রভুর হাত) | Reviewed — terms fixed above |
| 2-3 | Commissioning: son of man address begins, house of rebellion, watchman role begins, scroll-eating sign-act | Reviewed — terms fixed above |
| 4-5 | Siege sign-acts: iniquity/guilt-liability (পাপভার) established via day-count symbolism, abomination begins | Reviewed — terms fixed above |
| 6 | Idols, high places, abomination — first major cluster | Reviewed — terms fixed above |
| 7 | Day of the LORD/wrath | Reviewed — terms fixed above |
| 8-11 | Glory departs; idols in the temple itself; covenant jealousy trigger; tav mark; oth sign vocabulary | Reviewed — terms fixed above |
| 12-13 | Continued sign-acts; false peace (ironic use of শান্তি) | Reviewed — terms fixed above |
| 14 | Idols internalized “in the heart”; individual-responsibility theme anticipated | Reviewed — terms fixed above |
| 15-17 | Parabolic material (useless vine, two eagles); covenant-unfaithfulness metaphor (זָנָה) introduced at 16 | Reviewed — terms fixed above |
| 18 | Individual Responsibility for Sin — core doctrinal chapter: righteousness, nephesh, turn/repent | Reviewed — terms fixed above; highest divergence-note density in the book |
| 19 | Lament over Israel’s princes (allegory of lion cubs/vine) | Reviewed — no new terms; reuses established vocabulary only |
| 20 | Historical recital; “for my name’s sake” established; high places recur | Reviewed — terms fixed above |
| 21-24 | Sword oracles; bloodguilt (22); covenant-unfaithfulness metaphor extended (23); wife’s death as sign-act, oth vocabulary reinforced (24) | Reviewed — terms fixed above |
| 25-27 | Judgment on Ammon, Moab, Edom, Philistia, Tyre — nations vocabulary (জাতিগণ) established | Reviewed — terms fixed above |
| 28 | Tyre’s pride; Eden/garden-of-God used ironically | Reviewed — terms fixed above |
| 29, 32 | Egypt/Pharaoh oracles; sea-monster/tannin imagery | Reviewed — terms fixed above |
| 30-31 | Egypt continued; day-of-the-LORD language reapplied | Reviewed — no new terms beyond §1.4 |
| 33 | Watchman role reprised; individual-responsibility recap | Reviewed — reuses ch. 3/18 vocabulary |
| 34 | Failed shepherds; servant David; covenant of peace; prince (נָשִׂיא) introduced | Reviewed — terms fixed above |
| 35 | Judgment on Edom (Mount Seir) | Reviewed — no new terms; reuses §1.4 nations vocabulary |
| 36 | New Heart and New Spirit promissory form; heart of stone/flesh; Eden restored; covenant formula; spirit_of_god_ot | Reviewed — terms fixed above |
| 37 | Core passage. Dry bones; whole house of Israel; lost hope; great army; ruach polysemy (breath/wind/Spirit); covenant formula; servant David; covenant of peace; everlasting covenant; shakan (dwell); prophesy as causally effective speech | Reviewed — full treatment throughout this document and §6 below |
| 38-39 | Gog and Magog; latter days; recognition formula at climactic density; “for my name’s sake” | Reviewed — terms fixed above |
| 40-42 | Temple measurements and architecture | Reviewed — largely technical measurement vocabulary (cubits, chambers, gates); no doctrinal-risk terms beyond মন্দির/সদাপ্রভুর মন্দির already fixed |
| 43 | Glory returns; place of my throne; shakan (dwell) | Reviewed — terms fixed above |
| 44 | Sons of Zadok; clean/unclean ritual vocabulary (the one authorized শুচি-family context) | Reviewed — terms fixed above |
| 45-46 | Sacrificial offerings; prince’s cultic duties | Reviewed — terms fixed above |
| 47 | River of living water; trees for healing; sojourner land-inheritance | Reviewed — terms fixed above |
| 48 | Tribal land allocation; YHVH Shammah — book’s closing thesis statement | Reviewed — terms fixed above |
6. Ranked List of Highest-Risk Ambiguities
Ranked by a composite of (a) doctrinal risk tier, (b) frequency/centrality, (c) severity of cultural-collision exposure, and (d) depth of residual ambiguity even after the recommended strategy is applied. The core passage (37:1-14) supplies the single highest-ranked item.
-
רוּחַ polysemy within Ezekiel 37:1-14 itself (breath / wind / Spirit). No other ambiguity in the book is simultaneously this central to the core passage, this theologically loaded (New Heart and New Spirit doctrine; God’s Presence Returning), and this structurally unsolvable in Bengali — three distinct target words (নিঃশ্বাস, বায়ু, আমার আত্মা) are unavoidable, and the deliberate Hebrew wordplay is simply lost in the running text no matter which choices are made. Mitigation is entirely at the teaching-note level, not the lexical level. Highest priority for Phase 2 QA flagging.
-
גִּלּוּלִים / প্রতিমা — idols vs. Durga Puja devotional imagery. Highest cultural-collision severity of any term in the curriculum; occurs ~40+ times across judgment material and the internalized-idolatry material of ch. 14 and 20. Misapplied teaching (or its absence) risks either causing real offense to Hindu-background readers/neighbors or, conversely, under-communicating the seriousness of Ezekiel’s charge against Israel.
-
נֶפֶש vs. আত্মা (nephesh vs. soul-as-ātman). Central to ch. 18’s entire argument (the soul who sins shall die); a wrong lexical choice here does not just weaken a nuance but actively reverses the doctrine (mortal whole-person accountability vs. an indestructible eternal self that cannot really die).
-
ধার্মিকতা’s dual register (Ezekiel 18 covenantal-behavioral vs. Romans forensic). Because this word is fixed as Critical-risk in the Romans baseline already, any Ezekiel 18 occurrence lacking its mandatory divergence note risks a reader building an entire (wrong) doctrine of self-earned salvation directly out of curriculum-approved vocabulary.
-
בֶּן־אָדָם vs. NT “Son of Man.” 93 occurrences; a single unmarked lapse into মনুষ্যপুত্র anywhere in the book would import Christological weight into God’s address to a mere prophet, a direct doctrinal inversion (creaturely lowliness vs. veiled divine authority).
-
קִנְאָה / ঈর্ষা — covenant jealousy vs. sinful envy. Every one of its five occurrences (chs. 8, 16, 23, 36, 38) is load-bearing for the God’s-Presence-Departing doctrine; without the fencing note, God’s own motive for the Presence’s departure could be misheard as a divine character flaw.
-
תַּנִּין and the Manasa Devi collision risk. Lower frequency (two clustered occurrences, chs. 29 and 32) but a uniquely severe collision vector, since Bengal’s serpent-cult is a still-living folk-religious practice, not a defunct ancient one; a single careless নাগ-adjacent word choice would be far more damaging than its low frequency might suggest.
-
שָׁכַן / বসবাস করা — the under-weighted “dwell” verb. Structurally the weakest single lexical choice in the doctrine matrix (§1.6): an ordinary verb bearing the entire weight of the covenant-presence resolution at 37:26-28 and 43:7. Risk is not mistranslation but under-communication — readers simply not registering the theological climax the verb is meant to carry.
-
אֲדֹנָי יְהוִה consistency at ~217 occurrences. Not a conceptual ambiguity but an execution-risk ambiguity: at this frequency, partial compliance (some instances collapsing to bare সদাপ্রভু) is statistically likely without automated, not just human, QA enforcement.
-
גּוֹיִם / জাতিগণ vs. পরজাতি cross-document contamination. Risk is not internal to Ezekiel but emerges only in the context of the wider multi-book curriculum family sharing one translation memory; a translator or AI system trained heavily on Romans-style পরজাতি usage may default to it reflexively inside Ezekiel’s judgment oracles.
-
מקדש/מ্যাটেরিয়াল, מן্দির for OT temple vs. baseline’s church-vocabulary rejection. Documented in §1.7/§3 as a false-alarm collision, but ranked here because the appearance of contradiction with the Romans baseline is likely to trigger unnecessary reviewer escalation unless this document’s clarification is actively consulted.
-
נָשִׂיא / রাজপ্রধান — coined-term normalization drift. Lowest-ranked but still notable: across chs. 34-48, the temptation to relax the deliberately demoted “prince” title back into রাজা (king) for smoother prose would quietly erase the subordination-to-YHWH’s-kingship signal Ezekiel’s own vocabulary choice was designed to send.