Linguistic Gap Analysis
Numbers — Linguistic Gap Analysis
English → Bengali | Doctrine Vocabulary Matrix, Semantic Collision Mapping, Transliteration Decisions, Ranked Ambiguities
Curriculum: Numbers (গণনা পুস্তক) | Core passage: Numbers 21:4-9 (the bronze serpent)
Authority base: This analysis extends, and never contradicts, the Romans-baseline translation_memory.json and bible_term_registry.json. Every baseline term reused below is marked “(baseline reuse — no deviation).”
Coverage mandate: Full-book, Numbers 1–36. Chapters contributing no new vocabulary or doctrine-relevant gap are explicitly logged in Section 5 as “reviewed, no new gap.”
1. Method and Scope
This step identifies, for each of the seven curriculum doctrines, where the Bengali language (a) simply lacks an equivalent term (a missing-vocabulary gap), (b) offers several competing terms that each carry unwanted associations from Bengal’s dominant Hindu and Muslim traditions (a crowded semantic neighborhood requiring “fencing” — i.e., a chosen term plus an explicit exclusion rule for the rejected neighbors), or (c) requires a decision between transliterating a source term versus paraphrasing it descriptively.
Bengal’s religious landscape shapes every judgment below: West Bengal is Hindu-majority with Shakta (Durga/Kali), Vaishnava (bhakti), and folk-animist strands (including living snake-deity veneration); Bangladesh is Muslim-majority with mainstream Sunni practice alongside strong Sufi/pir-shrine folk religion. Both zones share the reading audience for this curriculum, so every term must be checked against both collision fields, not just one.
2. Doctrine Vocabulary Matrix
2.1 Unbelief and the Wilderness Generation
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| অবিশ্বাস (built on baseline বিশ্বাস + negation) | None significant — clean derivational fit | Adopt. Preserves full lexical continuity with Romans’ faith doctrine; Critical risk, theologian review every occurrence. |
| সন্দেহ (“doubt”) | Too weak; reduces contemptuous covenant-rejection to intellectual uncertainty | Reject as primary rendering; may appear in commentary describing psychological symptoms, never as the doctrinal headword. |
| অগ্রাহ্য করা (“spurn/reject” for נָאַץ, 14:11) | None; correctly captures active contempt | Adopt as companion term to অবিশ্বাস — the text pairs “despised me” with “not believed,” and Bengali must keep both registers (contempt + non-trust) visible. |
| অভিযোগ করা (“complain,” for תְּלוּנָה/לוּן) | None; needs consistency across all 5 complaint episodes (ch. 11, 14, 16, 20, 21) | Adopt, enforce identical rendering at every occurrence so Bengali readers can trace the Hebrew’s own repeated-vocabulary pattern. |
Gap identified: No missing vocabulary; the gap is discipline of consistency — Bengali has adequate resources, but five distinct complaint episodes must use the same verb or the doctrine’s cumulative literary force (this is a pattern, not isolated incidents) is lost in translation.
2.2 God’s Patience and Judgment
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| ঈশ্বরের ধৈর্য (patience) | ধৈর্য alone can read as passive endurance (a virtue Bengali culture praises in the sufferer, not the sovereign) | Adopt with fencing: always pair with অবিচল প্রেম (chesed) in teaching notes so patience is read as active, purposive, covenant-grounded, not fatalistic stoicism. |
| ক্রোধ (wrath) | Risk of being read as capricious divine temper, resembling folk-Hindu deity-anger appeased by offerings | Adopt but require a standing note: this is judicial (মিশপাত-grounded), not appeasement-transactional, wrath. |
| ইচ্ছাকৃত অবাধ্যতা vs. অজ্ঞানে করা পাপ (deliberate vs. unintentional sin) | No single Bengali word carries this Mosaic legal distinction; risk of collapsing both into generic পাপ | Missing-vocabulary gap — resolved by compound phrase, not a single term. This pairing is the theological hinge explaining why Numbers 21 gets instant severe judgment while Numbers 15 provides for ordinary lapses. |
Gap identified: Bengali has no native single-word equivalent for the deliberate/inadvertent sin distinction that structures Mosaic sacrificial law (15:22-31) and explains 21:4-9’s severity — a genuine missing-vocabulary gap requiring the two-phrase compound solution above rather than a lexical shortcut.
2.3 The Bronze Serpent as a Type of Christ
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| পিতলের সাপ (“bronze/brass snake”) | Low collision in itself, but the entire scene — snakebite, serpent image, healing — sits directly on top of Bengal’s living Manasha (মনসা) snake-goddess cult, a real folk-Hindu healing practice, not a hypothetical | Adopt with mandatory fencing: every occurrence requires a translator note distinguishing (a) obedient looking to God’s own appointed sign from (b) any independent healing power resident in the object, plus a forward cross-reference to the object’s later corruption into the idol Nehushtan (2 Kings 18:4) — proof the text itself anticipates and condemns the very misreading Bengali folk religion invites. |
| মনসার সাপ / সর্পদেবী-related vocabulary | Direct name-collision with the Hindu snake goddess | Hard reject — never use in any form, even descriptively, without extreme distancing framing. |
| জ্বলন্ত সাপ (fiery serpents, the judgment) vs. পিতলের সাপ (the remedy) | Risk of translators collapsing both into one generic “snake” term, erasing the judgment/remedy contrast | Adopt distinct terms and never substitute one for the other; the contrast is the passage’s entire theological structure. |
| দণ্ড (pole/standard, נֵס) | পতাকা (“flag/banner”) over-specifies toward a military/national symbol | Adopt দণ্ড; cross-reference John 3:14’s “lifted up” (ঊর্ধ্বে উন্নত করা) in a study note, since base-text vocabulary differs from Johannine vocabulary and must not be silently harmonized. |
| দেখা (“look/see,” רָאָה) vs. বিশ্বাস করা (“believe”) | Temptation to over-translate দেখা as বিশ্বাস করা to make the John 3 typology explicit in the base text itself | Reject the substitution. Keep দেখা in the base narrative; carry the faith-typology connection only in the commentary layer. This is the single highest-discipline rule in this doctrine’s vocabulary. |
| বাঁচবে / জীবিত থাকবে (“will live”) vs. পরিত্রাণ পাবে (“will receive salvation”) | Same collapsing risk as above, at the doctrine’s climax (21:8-9) | Reject the substitution in the base text; reserve পরিত্রাণ strictly for the commentary layer per the baseline’s established type/fulfillment discipline (cf. পুনরুত্থান vs. পুনর্জন্ম). |
Gap identified: No missing vocabulary — every needed Bengali word exists. The entire risk is a crowded semantic neighborhood with a live cultic practice sitting directly inside the narrative’s own subject matter (snakebite + serpent image + healing), the single most acute collision in the whole curriculum, ranked #1 below.
2.4 Intercession (Moses, Aaron, Phinehas)
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| মধ্যস্থতা (baseline reuse — no deviation) | Baseline treats this as Medium risk (general prayer for others); Numbers gives it three distinct narrative embodiments (verbal plea, 14:13-19/21:7; ritual atonement-action, 16:46-48; decisive zealous action, 25:6-13) that the single Bengali word must stretch to cover | Adopt, risk elevated to High for this curriculum. Require a teaching note enumerating the three modes so Bengali readers do not assume intercession is only verbal prayer. |
| ঈশ্বরের প্রতি উদ্যম (zeal) vs. bare ঈর্ষা | ঈর্ষা denotes envy/jealousy, a vice, in ordinary Bengali | Fence firmly: never use ঈর্ষা for Phinehas’s zeal (25:11); one Hebrew root (קִנְאָה) covers both senses but Bengali has no single word that does, so disambiguation must be explicit and constant. |
| প্রায়শ্চিত্ত (atonement) | Shares form/root sense with Hindu self-directed karmic penance/expiation — the deepest doctrinal collision in the curriculum, since it touches Grace/Justification territory already Critical in the baseline | Retain per baseline’s historically-stabilized-term precedent (cf. ঈশ্বর), but mandate human theologian review at every single occurrence with a standing note: God-provided/received (biblical) vs. self-generated/earned (karmic). |
Gap identified: Missing vocabulary — Bengali has no ready-made way to distinguish the three biblical modes of intercession (plea / priestly ritual act / zealous decisive act) that Hebrew covers with different verbs across these three narratives; মধ্যস্থতা must be doctrinally over-loaded and requires teaching support to carry the full range.
2.5 Faithfulness of God’s Promises despite Human Failure
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| অবিচল প্রেম (chesed) | No existing single Bengali word matches chesed’s specific sense of covenant-obligated, relationally-grounded loyalty-love; neighboring words each capture only part | New compound term, adopted. কৃপা (baseline: non-soteriological compassion only), অনুগ্রহ (baseline: NT justification-grace specifically), দয়া (generic mercy, no covenant-bond presupposition) — none alone suffices. অবিচল প্রেম sits deliberately between দয়া and অনুগ্রহ. |
| ঈশ্বরের পরিচালনা (providence, baseline reuse) | None significant | Adopt for the cloud/fire guidance and the land-promise’s ongoing fulfillment (ch. 21’s victories over Sihon and Og). |
| নিয়ম (covenant, baseline reuse) | None significant | Adopt for the Phinehas “covenant of perpetual priesthood” (25:12-13) and Levitical covenant (18:19). |
| উত্তরাধিকার (inheritance) | অধিকার alone loses the “passed-down/allotted-by-God” sense | Adopt উত্তরাধিকার; cross-reference the baseline’s দত্তকপুত্রতা (adoption/inheritance-rights) doctrine for teaching continuity — Numbers 27/36 (Zelophehad’s daughters) foreshadows the New Testament’s fuller inclusion of unexpected heirs. |
Gap identified: Missing vocabulary — chesed is the clearest true lexical gap in the entire book; no existing Bengali word carries its specific covenant-relational-loyalty sense, hence the new compound coinage অবিচল প্রেম rather than reuse of any existing candidate.
2.6 Census and the People of God
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| গণনা (census, the book’s own established Bengali title-term, গণনা পুস্তক) | Homonymous with everyday Bengali for astrological/fate calculation (ভাগ্য গণনা) — the same root the Romans baseline already flagged as a rejected alternative for “prophecy” | Adopt (no viable alternative exists as the book-title-term is fixed by tradition) with mandatory disambiguating note at first use (ch. 1) and deliberate thematic exploitation of the tension at ch. 22-24, where Balaam’s false গণনা (divination) is set against God’s true numbering of his people. |
| সমাজ (congregation/assembly, עֵדָה) | Must not be confused with মণ্ডলী, which the baseline reserves strictly for the NT ekklesia | Adopt সমাজ, fenced against মণ্ডলী substitution. Preserves the OT ethnic-national/NT multi-ethnic distinction directly relevant to the baseline’s already-flagged Romans 9-11 discussion. |
| সমাগম-তাঁবু (tent of meeting) | Risk of collapsing into মন্দির (temple) — explicitly forbidden by the baseline for church/temple concepts | Adopt সমাগম-তাঁবু; never মন্দির. Portability (God journeying with an imperfect, pilgrim people) is the theological point, lost if rendered as a fixed shrine-term. |
Gap identified: Crowded semantic neighborhood, not missing vocabulary — গণনা is unavoidable as the fixed book-title term but sits in direct competition with its own astrological homonym; this requires fencing by repeated disambiguation rather than a substitute word, since no substitute carries the same weight of established Bible-translation tradition.
2.7 Balaam and God’s Sovereign Blessing
| Available Bengali terms | Weakness | Recommended strategy |
|---|---|---|
| ভাববাদী (prophet, baseline reuse) | Baseline treats as Low risk (Romans register); applying it to Balaam — a pagan diviner through whom God nonetheless speaks true prophecy — is doctrinally complex | Adopt with risk elevated to High. Require a translator note distinguishing Moses’ legitimate prophetic office (“face to face,” 12:8) from Balaam’s compelled, non-legitimate pagan mouthpiece role, who nonetheless received genuine ভাববাণী. |
| ভাগ্যগণনা (divination, קֶסֶם) | Built on the same root as census-গণনা; also risks confusion with generic astrology | Adopt with fencing, deliberately retaining the shared root with গণনা to preserve, via disambiguating modifiers, the text’s own ironic contrast between Balaam’s manipulative omen-counting and God’s true covenant-numbering. High risk given Bengal’s living folk-astrological and tantric-divination culture (a real, not hypothetical, collision — clients regularly consult astrologers/tantriks in both West Bengal and rural Bangladesh). |
| আশীর্বাদ / অভিশাপ (bless/curse) | আশীর্বাদ risks being read as a Hindu-devotional “boon” (বর) granted by a deity in transactional exchange for devotion — directly analogous to the baseline’s existing caution about বর for spiritual gifts | Adopt with fencing: reinforce in every occurrence that these are sovereign, effective pronouncements entirely under সদাপ্রভু’s own control (23:8), never negotiable boons exchanged for devotion or payment (Balak’s whole scheme, which the text itself mocks). |
Gap identified: No missing vocabulary; the entire risk is crowded-neighborhood collision between (a) legitimate biblical prophecy/blessing and (b) Bengal’s extremely active living traditions of paid astrological consultation, fortune-telling, and boon-seeking — arguably the most culturally live (not merely historically-rooted) collision zone outside the bronze serpent.
3. Missing Vocabulary vs. Crowded Semantic Neighborhoods — Summary Table
| Type | Term/Concept | Nature of the problem | Fencing/resolution strategy |
|---|---|---|---|
| Missing vocabulary | chesed (steadfast covenant love) | No single existing Bengali word covers it | New compound: অবিচল প্রেম |
| Missing vocabulary | deliberate vs. unintentional sin distinction | No single-word Mosaic legal category in Bengali | Two-phrase compound solution: ইচ্ছাকৃত অবাধ্যতা / অজ্ঞানে করা পাপ |
| Missing vocabulary | three distinct modes of intercession (plea/ritual act/zealous act) | Bengali মধ্যস্থতা must be stretched to cover all three | Single term retained, but doctrine-note enumerates the three modes explicitly |
| Crowded neighborhood | bronze serpent / snakebite-healing scene | Collides with living Manasha (মনসা) snake-goddess folk healing cult | পিতলের সাপ retained; mandatory per-occurrence translator note + Nehushtan cross-reference |
| Crowded neighborhood | atonement (প্রায়শ্চিত্ত) | Collides with Hindu self-directed karmic penance/expiation | Retained per baseline’s stabilized-term precedent; mandatory theologian review every occurrence |
| Crowded neighborhood | census/divination root (গণনা / ভাগ্যগণনা) | Both share a root with everyday fate-calculation/astrology vocabulary | Retained as fixed terms with disambiguating notes; root-sharing exploited deliberately for the Balaam narrative’s own irony |
| Crowded neighborhood | grain offering (নৈবেদ্য) / burnt offering (হোমবলি) | Structurally near-identical to Hindu puja food-offering and Vedic fire-sacrifice practice | Retained as established Bible terms per baseline precedent (cf. ঈশ্বর); mandatory standing note that the offering is to সদাপ্রভু alone |
| Crowded neighborhood | vow (প্রতিজ্ঞা) vs. ব্রত/মানত | ব্রত = Hindu merit-earning ritual vow; মানত = transactional shrine-vow common to both Hindu and Muslim folk practice | প্রতিজ্ঞা adopted; both rejected alternatives explicitly excluded with grounded reasons |
| Crowded neighborhood | idol (মূর্তি) vs. প্রতিমা | প্রতিমা specifically evokes Durga/Kali/Saraswati festival images | মূর্তি adopted as the generic, less festival-loaded idol-term |
| Crowded neighborhood | zeal (উদ্যম) vs. ঈর্ষা | Same Hebrew root covers sinful envy and righteous zeal; Bengali has no equivalent single-root ambiguity | উদ্যম adopted; ঈর্ষা hard-excluded |
| Crowded neighborhood | ceremonial purity (শুচি/অশুচি) vs. holiness (পবিত্র) | Both could plausibly render category near “purity/impurity,” but the baseline reserves পবিত্র for moral holiness only | শুচি/অশুচি adopted specifically for ch. 19’s ritual/corpse-defilement category; standing side-by-side note required wherever both vocabularies co-occur |
| Crowded neighborhood | prophet/prophecy applied to Balaam | ভাববাদী/ভাববাণী otherwise clean terms, but application to a compelled pagan diviner is doctrinally unusual | Retained with risk elevated to High and a mandatory legitimacy-distinguishing note |
4. Transliteration vs. Paraphrase Decisions
| Term | Decision | Rationale |
|---|---|---|
| YHWH (יהוה) | Established-convention substitution, not transliteration: সদাপ্রভু | Bengali Bible tradition does not transliterate the Tetragrammaton; it uses the fixed compound সদাপ্রভু (“the eternal/true Lord”). This is not a new decision but an extension of settled OT translation convention, distinct from the baseline’s প্রভু (reserved for Adonai/NT kyrios). |
| Abba-style intimacy terms | Not applicable in Numbers (no Aramaic terms of this kind occur); noted here only to confirm no transliteration decision is needed for this book beyond proper names. | — |
| Nazirite (נָזִיר) | Paraphrase into a compound descriptive term, not transliteration: নাসরীয় প্রতিজ্ঞা (“Nazirite vow”) | A bare transliteration (e.g., নাযীর) would be opaque to a reader with no Hebrew background; the compound preserves the proper-name flavor (নাসরীয়) while making the vow-content recoverable from the phrase itself. |
| Cherem (חֵרֶם) | Paraphrase, not transliteration: সম্পূর্ণ বিনাশের জন্য উৎসর্গ করা (“dedicated for complete destruction”) | No established Bengali transliteration exists in Bible tradition; a bare transliterated loanword would carry zero meaning and risk being read as a technical military term without its dedicatory-to-God dimension. |
| Manna, Sukkot-type festival names | Transliterate/retain established Bengali Bible forms: (feast of tabernacles →) কুটিরবাস পর্ব (descriptive, established) | Where Bengali Bible tradition already has a stable descriptive rendering, retain it rather than introduce a Hebrew loanword unfamiliar to readers. |
| Proper names (Balaam, Balak, Phinehas, Zelophehad, Sihon, Og, Midian, Korah) | Transliterate per established/derived Bengali Bible transliteration conventions (see Section 3 of 08_core_glossary.md) | Standard Bible-translation practice; low doctrinal risk, but consistency across all 36 chapters is required for narrative trackability. |
| Chesed | Paraphrase into a new compound, not transliteration: অবিচল প্রেম | Transliterating (হেসেদ) would be meaningless to a Bengali reader with no Hebrew background and would not connect to the existing দয়া/অনুগ্রহ semantic field the doctrine needs to interact with. |
| Amen/Hallelujah-class liturgical transliterations | Not present as new decisions in Numbers; baseline’s আমেন/হাল্লিলূয়া convention carried forward unchanged if such terms appear in doxological commentary. | — |
General rule derived: Bengali Bible tradition transliterates almost nothing from Hebrew narrative vocabulary except a small closed set of already-fixed liturgical/name terms; theological concepts (chesed, cherem, nazir-vow) are consistently paraphrased into descriptive Bengali compounds rather than imported as loanwords. This curriculum follows that established pattern rather than introducing new transliterations.
5. Full-Book Chapter Coverage Log
| Ch. | New vocabulary / gap contribution | Status |
|---|---|---|
| 1 | census_numbering (গণনা), congregation_assembly (সমাজ) | New terms flagged |
| 2 | Camp arrangement around the tent of meeting | Reviewed — no new gap (uses সমাগম-তাঁবু, already flagged at ch. 9) |
| 3 | redemption_of_firstborn, sanctification (Levite consecration) | New terms flagged |
| 4 | service_ministry (Levitical duties) | New term flagged |
| 5 | ceremonial_purity, guilt_restitution_offering (jealousy/sotah law) | New terms flagged |
| 6 | vow_general, nazirite_vow, bless_curse (priestly blessing) | New terms flagged |
| 7 | burnt_offering, grain_offering, peace_fellowship_offering | New terms flagged |
| 8 | sanctification (Levite ordination) | Reviewed — reuses ch. 3 terms |
| 9 | passover, cloud_and_fire, providence | New terms flagged |
| 10 | providence (journey begins) | Reviewed — reuses ch. 9 terms |
| 11 | complain_grumble, craving, the_spirit_ot_empowering | New terms flagged |
| 12 | prophet (Moses vs. Miriam), holy/ceremonial purity (leprosy) | New term flagged (prophet); ceremonial_purity reused |
| 13 | Spies sent into Canaan | Reviewed — no new gap; sets up ch. 14 |
| 14 | unbelief, intercession, steadfast_love_chesed, test_god, despise_spurn, gods_patience | Core doctrinal chapter — five new/elevated terms flagged |
| 15 | deliberate_defiant_sin, unintentional_sin | New terms flagged — legal hinge for ch. 21’s severity |
| 16 | rebellion, glory, wrath_of_god, atonement, intercession (Aaron’s incense) | New/elevated terms flagged |
| 17 | Aaron’s staff confirms priesthood | Reviewed — no new gap; reinforces covenant/priesthood vocabulary already flagged |
| 18 | covenant (Levitical), inheritance-adjacent priestly portion | Reviewed — reuses covenant, inheritance terms |
| 19 | ceremonial_purity, water_of_purification (red heifer) | New terms flagged |
| 20 | unbelief, holy, test_god (Meribah); Edom’s refusal | Reviewed — reuses ch. 14/20 terms; Meribah is a second unbelief data point requiring the same pastoral-care note |
| 21 | bronze_serpent, fiery_serpents, pole_standard, look_see, live_survive (vv. 4-9, core passage); providence (Sihon/Og victories); devote_to_destruction_cherem (vv. 1-3) | Core passage chapter — highest-density new-term chapter in the book |
| 22 | divination, prophecy_oracle (Balaam introduced) | New terms flagged |
| 23 | bless_curse, prophecy_oracle (Balaam’s oracles) | Reuses ch. 22 terms; doctrinal payload |
| 24 | prophecy_oracle, Messianic-adjacent oracle language (a star out of Jacob) | Reviewed — reuses ch. 22-23 terms; flag for cross-reference to Messianic Promise doctrine in the Romans baseline |
| 25 | zeal_for_god, covenant (Phinehas), atonement, idol (Baal-Peor) | New terms flagged |
| 26 | census_numbering (second census) | Reviewed — reuses ch. 1 terms |
| 27 | inheritance, commissioning_lay_hands (Zelophehad’s daughters; Joshua) | New terms flagged |
| 28 | burnt_offering, grain_offering (offering calendar) | Reviewed — reuses ch. 7 terms |
| 29 | feast_of_tabernacles, day_of_atonement | New terms flagged |
| 30 | vow_general (women’s vows) | Reviewed — reuses ch. 6 terms |
| 31 | devote_to_destruction_cherem (Midianite war) | Reviewed — reuses ch. 21 term |
| 32 | Reuben/Gad request land east of Jordan | Reviewed — no new gap; inheritance vocabulary reused |
| 33 | idol (destruction of idols in the land); journey summary | Reviewed — reuses ch. 25/33 idol term |
| 34 | Land boundaries defined | Reviewed — no new gap; inheritance vocabulary reused |
| 35 | cities_of_refuge, avenger_of_blood | New terms flagged |
| 36 | Zelophehad’s daughters’ marriage law; inheritance | Reviewed — reuses ch. 27 terms |
Coverage confirmation: All 36 chapters of Numbers have been reviewed. 22 chapters contribute new vocabulary or elevated-risk terms; 14 chapters are confirmed as reusing previously flagged vocabulary with no new gap, and are recorded here explicitly rather than silently omitted, per the full-book coverage mandate.
6. Ranked List of Highest-Risk Ambiguities
Ranked by combined severity of (a) doctrinal centrality to this curriculum’s seven named doctrines, (b) grounded cultural-collision strength in Bengal specifically, and (c) risk that a single mistranslation could produce a syncretistic misreading rather than mere awkwardness.
-
পিতলের সাপ (bronze serpent) vs. Manasha (মনসা) snake-goddess healing folk practice. Highest rank because the collision is not merely lexical but situational: the exact real-world scenario (venomous snakebite + a serpent image + healing) is the object of active devotional practice in rural Bengal today. A Bengali reader encountering Numbers 21:8-9 without the mandatory translator note could plausibly read the bronze serpent as a Manasha-type healing totem rather than a sign pointing away from itself to সদাপ্রভু’s own provision — directly undermining the Bronze Serpent as a Type of Christ doctrine and its fulfillment in John 3:14-15.
-
দেখা (“look”) vs. বিশ্বাস করা (“believe”) in Numbers 21:8-9. Second-ranked because the failure mode here is theological over-correction by well-meaning translators: substituting “believe” for “look” to make the John 3 typology explicit would collapse the OT type into its NT fulfillment and misrepresent what the Numbers text itself actually asks of the dying Israelite (a bare physical act of obedient looking, not a developed faith-response). This is the single most likely self-inflicted translation error in the core passage.
-
প্রায়শ্চিত্ত (atonement) vs. Hindu karmic self-penance. Ranked third because this term touches the same theological territory as Grace and Justification, already Critical in the Romans baseline, and appears repeatedly across Intercession doctrine (Aaron, Phinehas) and the Day of Atonement — a high-frequency term whose collision, if unmanaged, would reframe God-provided atonement as a human merit-accumulation project running through the entire book.
-
অবিচল প্রেম (chesed) risk of collapsing into দয়া or অনুগ্রহ. Ranked fourth because Moses’ entire intercessory argument in 14:18 depends on chesed’s specific covenant-relational logic; if translators default to a neighboring word for convenience, the Faithfulness of God’s Promises despite Human Failure doctrine loses its textual anchor at its single clearest proof-text.
-
গণনা (census) vs. its own astrological/fate-calculation homonym. Ranked fifth because the term is unavoidable (it is the book’s fixed Bengali title) yet directly borders live folk-astrological practice; the risk compounds at ch. 22-24 where Balaam’s own false গণনা/divination sits in the same narrative as God’s true census-numbering, requiring readers to hold both senses apart deliberately rather than by accident.
-
আশীর্বাদ (blessing) vs. Hindu-devotional বর (“boon”) logic in the Balaam narrative. Ranked sixth because Balak’s entire scheme in ch. 22-24 is a transactional attempt to purchase a curse/blessing — precisely the boon-for-payment logic the text itself mocks — and Bengali আশীর্বাদ risks importing that very logic back in if not fenced as sovereign, non-negotiable divine pronouncement.
-
নৈবেদ্য (grain offering) and হোমবলি (burnt offering) vs. Hindu puja food-offering and Vedic fire-sacrifice. Ranked seventh — high-frequency (chs. 7, 15, 28-29) and structurally close to living Bengali Hindu ritual practice, but lower doctrinal centrality than items 1-6 since these are background cultic-law terms rather than terms carrying one of the seven named curriculum doctrines directly.
-
ঈর্ষা (envy) vs. উদ্যম (zeal) at Phinehas’s decisive act (25:11). Ranked eighth because a single wrong word choice would morally invert the entire Intercession doctrine’s third embodiment, turning a commended act of covenant-zeal into a condemned act of sinful jealousy — a severe error, but confined to a narrower textual footprint than items above.
-
শুচি/অশুচি (ceremonial purity) vs. পবিত্র (holiness) at Numbers 19. Ranked ninth: real risk of Hindu-background readers collapsing biblical ritual purity into biblical moral holiness (or vice versa), but the baseline’s own precedent (rejecting শুচি for holy, reserving it correctly here) already substantially manages this risk; residual risk is in maintaining the distinction consistently, not in an absence of guidance.
-
প্রতিজ্ঞা (vow) vs. ব্রত/মানত (Hindu merit-vow / folk shrine-bargain vow). Ranked tenth: real collision (chs. 6, 30) but lower doctrinal centrality among the seven named curriculum doctrines; primarily a background-law term rather than a load-bearing doctrinal term for this specific curriculum.
7. Recommendations Carried Forward to Phase 1 Step 9 (Core Glossary) and Phase 2
- Every term ranked 1-6 above requires mandatory human theologian review at every occurrence, not sample-based review, consistent with the doctrine_risk_registry.json “Critical” routing standard.
- The bronze-serpent translator note (rank #1) and the look/believe discipline (rank #2) should be built into the Phase 2 AI system prompt as hard-coded validation checks, analogous to the Romans package’s forbidden-substitution list, given how easily both could fail silently. 2a. Recommend adding to the Phase 2 validation checklist: “Verify দেখা is not rendered as বিশ্বাস করা in Numbers 21:8-9” and “Verify পরিত্রাণ/অনুগ্রহ do not appear in the base narrative translation of Numbers 21:4-9, only in the commentary layer.”
- The চেসেদ (chesed) compound অবিচল প্রেম should be added to translation memory as a new Critical-adjacent High-risk entry before Phase 2 begins, per the recommendation already logged in
08_core_glossary.md. - The গণul-Balaam divination root-sharing (গণনা/ভাগ্যগণনা) should be exploited pedagogically in teaching materials for ch. 22-24, not merely flagged defensively — it offers a natural teaching hook contrasting true and false “counting/reckoning” that reinforces the Balaam doctrine’s core claim (God’s sovereign blessing cannot be manipulated by divination).
- All findings in this document are incorporated into
analysis/08_core_glossary.mdandassets/bible_term_registry.json; no term decision recorded here may be altered downstream without a corresponding update to this file’s Section 6 ranking rationale.
This document must be loaded alongside 08_core_glossary.md before Phase 2 segment translation of Numbers. It extends but never contradicts the Romans-baseline translation_memory.json and bible_term_registry.json.