Linguistic Gap Analysis
Linguistic Gap Analysis — Titus (English → Hindi)
Scope and Method
This analysis covers the full book of Titus (chapters 1–3), with Titus 2:11–3:8 as the theological anchor passage. It builds directly on 07_semantic_analysis.md and 08_core_glossary.md and must not contradict any rendering already fixed in the baseline translation_memory.json or in the Titus-specific additions recorded in assets/bible_term_registry.json. Where this document identifies a gap or a fencing requirement, the recommended strategy is stated as a requirement for Phase 2 segment translation, not as a new term decision — new Critical/High term decisions remain the property of the glossary and registry files.
Three structural facts drive this analysis and distinguish Titus from the Romans/Galatians baseline:
- Titus is short, dense, and repetition-driven. Six terms recur as fixed refrains (good works ×6, sound doctrine ×5, Savior title ×6, submission-family ×3+, appearing ×3, purity-family ×3 near-synonyms). Any inconsistency in a refrain term is far more visible to a Hindi reader than a single-occurrence inconsistency would be in a longer letter.
- Titus concentrates its highest-risk material in a single fourteen-verse core passage (2:11–3:8). Nearly every Critical-risk term in the book’s glossary occurs there. This creates unusual density of collision risk in one continuous unit of text, rather than the more distributed risk profile of Romans.
- Titus is a household/civic-order letter, unlike the argumentative theology of Galatians. This foregrounds a category of risk barely present in the earlier baseline runs: social-hierarchy vocabulary (submission, master/slave, elder-as-office vs. elder-as-age) intersecting live caste and gender hierarchies in the Indian context.
1. Doctrine Vocabulary Matrix
| Doctrine | Available Hindi Term(s) | Weaknesses of Each | Recommended Strategy |
|---|---|---|---|
| Grace That Trains for Godly Living | अनुग्रह (grace, TM); प्रशिक्षित करना (train); भक्ति (godliness); तपस्या (ascetic discipline, available but forbidden) | अनुग्रह alone does not signal agency — Hindi readers default to reading a nearby verb’s subject as the human actor unless grammar forces otherwise; तपस्या is the natural indigenous word for “spiritual training/discipline” but implies self-effort earning spiritual merit, the exact inversion of Titus 2:12. भक्ति is doctrinally serviceable but is simultaneously the primary term for Hindu devotional religion. | Keep अनुग्रह as the unambiguous grammatical subject of प्रशिक्षित करना in 2:12 (verified at the sentence level, not just the word level, during Phase 2 QA). Never introduce तपस्या or its cognates (तप, साधना) anywhere in this doctrine’s vocabulary. Attach a standing translator note to भक्ति on first occurrence per curriculum document distinguishing Spirit-produced Christlike character from bhakti-mārga devotion. |
| Qualifications for Elders | प्राचीन (elder-office); अध्यक्ष (overseer); बुज़ुर्ग (generic elderly); गुरु (available, forbidden) | बुज़ुर्ग describes age only, no office; village-panchayat “elder” carries unquestionable customary authority by virtue of age/lineage, unlike Paul’s checkable moral/doctrinal criteria; गुरु is the dominant Indian model of a revered, quasi-permanent spiritual authority-figure whose status rarely depends on itemized, revocable qualifications. | Reserve प्राचीन/अध्यक्ष exclusively for the church office defined by 1:6–9’s criteria; never let बुज़ुर्ग or गुरु leak into this doctrine’s vocabulary even in informal glosses. Introductory material accompanying 1:5–9 should state explicitly that the office is qualification-based and revocable, not lineage- or age-conferred. |
| Sound Doctrine and Good Works | खरी शिक्षा / स्वस्थ शिक्षा (sound doctrine); भले काम (good works, TM-extended); भले कर्म (forbidden variant) | The single available “works” pair — काम (neutral action) vs. कर्म (action understood within a merit-ripening cosmology) — is a crowded neighborhood: common spoken Hindi frequently uses कर्म for ordinary “good deeds” without any technical karma intent, so a translator drafting naturally will be pulled toward कर्म by default register, not by theological error. | Absolute fencing: काम only, for all six occurrences of the “good works” refrain, with no exceptions for register or euphony. Apply a single fixed rendering (खरी शिक्षा, not alternating with स्वस्थ शिक्षा) across all five “sound doctrine” occurrences to preserve the refrain’s audibility. |
| Salvation by Grace not Works | उद्धार (salvation, TM); धार्मिकता (righteousness, TM); धर्म (forbidden); मुक्ति/मोक्ष (forbidden) | Titus 3:5 stacks three of the curriculum’s most collision-prone words in one clause (“not because of works done by us in righteousness, but according to his own mercy, he saved us”): काम, धार्मिकता, and उद्धार all sit within one sentence, each individually fenced but jointly forming the single highest-density karma-collision sentence in the letter. | Treat Titus 3:5 as a unit requiring combined sentence-level (not just term-level) theologian review — verify the full clause reads as an antithesis to any merit-fruit framework, not merely that each word individually avoids its forbidden counterpart. |
| Regeneration by the Holy Spirit | नया जन्म (new birth, TM-extended per baseline new_birth); पुनर्जन्म (forbidden); पवित्र आत्मा (TM); स्नान (forbidden for baptism-adjacent washing) | παλιγγενεσία is etymologically “again” + “birth” — the single Greek word most likely to seduce a translator into the literally correct-looking but doctrinally forbidden पुनर्जन्म by direct morpheme-matching rather than by any conceptual error. This is a uniquely acute version of the gap the baseline already flagged once for “new birth” in Romans/Phase 2 v1. | Never construct “washing of regeneration” compositionally from Greek roots. Use the fixed phrase नये जन्म का धोना exactly as fixed in the registry; forbid any translator-generated alternative phrasing for this term, including at draft stage. Mandatory theologian sign-off before any Phase 2 segment containing this term is marked complete. |
| Submission to Authority | अधीन होना (submission, TM-extended); आज्ञा मानना (obey); दास/दासत्व (slavery, TM); दासता के अधीन जन्म से (forbidden implication) | Hindi has no single unmarked word for “voluntary submission within a God-ordered structure” that is not also readable as permanent, birth-conferred subordination — precisely the semantic field occupied by caste (jāti/varṇa) hierarchy and by patriarchal household custom. There is no existing Hindi Christian idiom that cleanly separates “submit because God has ordered it, for this season/relationship” from “submit because you are inherently lesser.” | Every occurrence (2:5 wives, 2:9 slaves, 3:1 civil rulers) requires a co-located translator note (not merely a glossary entry) clarifying that the submission is relational/positional, voluntary, and reciprocal within the passage’s own household-code logic (cf. 2:9–10’s “well-pleasing,” not coerced), not an assertion of the submitting party’s lesser being. Route every occurrence to combined theologian + native speaker review per the registry. |
| Avoiding Divisive Controversies | विभाजन करनेवाला (divisive person); कल्पित कहानियाँ (myths); मूर्खतापूर्ण विवाद (foolish controversies); पाखण्डी/विधर्मी (heretic-family, forbidden) | Available Hindi vocabulary for “false teacher” or “divisive person” is dominated by words built on धर्म (विधर्मी = “against dharma/religion,” पाखण्डी = “hypocrite/heterodox”) which imports a right-religion-vs-wrong-religion framing entirely foreign to Paul’s narrower pastoral concern: an individual disrupting one specific congregation’s relational unity. | Use only विभाजन करनेवाला, built on विभाजन (“division/split”) rather than any धर्म-root word. Reject विधर्मी/पाखण्डी/नास्तिक categorically for this doctrine across all documents in the curriculum. |
| Deity of Christ (cross-cutting via Savior-title sharing) | उद्धारकर्ता (Savior, TM-extended); महान परमेश्वर (great God); त्रिएकता (Trinity, TM, background) | Hindi syntax permits (and casual translation habit favors) inserting a comma or “और भी” (“and also”) between paired divine titles, which would read the single Sharp’s-rule referent of 2:13 as two separate beings — a misreading that the Hindi religious environment, accustomed to distinguishing a supreme deity from a lesser divine emissary/avatar, would find entirely natural rather than obviously wrong. | Fix the single-referent Hindi syntax of 2:13 and 1:4/3:6 in one register-checked template (हमारे महान परमेश्वर और उद्धारकर्ता यीशु मसीह) applied verbatim at every occurrence; forbid any inserted conjunction or punctuation that would sever the single-referent reading. Verify with back-translation specifically probing whether a naive Hindi reader would count one person or two. |
2. Missing Vocabulary vs. Crowded Semantic Neighborhoods
Titus’s linguistic risk profile divides cleanly into two distinct failure modes, and Phase 2 processes must treat them differently.
2.1 Missing Vocabulary (true gaps — no adequate existing Hindi word)
These are concepts for which Hindi (including established Hindi Christian usage) simply has no single existing word that carries the required sense; a compound, gloss, or newly-fixed phrase is unavoidable.
| Concept | Why no single word exists | Chosen strategy |
|---|---|---|
| Washing of regeneration (λουτρὸν παλιγγενεσίας) | No Hindi word combines “washing/cleansing” with “once-for-all new-birth” without either defaulting to a ritual-bathing word (स्नान) or a reincarnation word (पुनर्जन्म). | Fixed compound नये जन्म का धोना (registry-fixed); treat as an irreducible two-word technical phrase, not a single lexical item. |
| Savior applied identically to Father and Son | Hindi has titles for a supreme deity (परमेश्वर, ईश्वर) and titles for a saving intermediary figure, but ordinary religious Hindi usage does not have a habit of applying one shared title to both without distinction — Hindu theology typically differentiates a supreme unmanifest deity from a saving avatar/incarnation by title as well as by nature. | Fixed identical rendering उद्धारकर्ता for both referents in all six occurrences (registry-fixed); the sameness of the word across referents is itself the doctrinal payload and must be preserved as a formal pattern, not just a lexical choice. |
| ”Great God and Savior Jesus Christ” as one grammatical referent | Hindi has no native syntactic convention paralleling the Greek single-article-governing-two-nouns construction (Sharp’s rule); Hindi coordination of two honorific titles for a person defaults toward ambiguity between one-referent and two-referent readings. | Fixed template phrase (registry-fixed) applied without variation; explicit translator note required. |
| The custodial-but-terminated role of the law (relevant background from Galatians, echoed in Titus’s “quarrels about the law,” 3:9) | Not itself present in Titus as a term, but the letter assumes readers understand law-quarrels as a closed controversy Titus should decline to re-enter — Hindi has ample vocabulary for ongoing legal/religious disputation (विवाद, झगड़ा) but no ready single word for “a controversy one should refuse to re-litigate because it has already been doctrinally settled.” | Handled by context/note rather than a lexical fix; flagged here so Phase 2 drafters do not attempt to invent a single-word solution where none is needed — a descriptive clause suffices. |
2.2 Crowded Semantic Neighborhoods (too many concepts competing for too few safe words — require fencing, not invention)
These are the inverse problem: Hindi has words available, but the same word (or a small cluster of near-synonyms) is being asked to carry multiple distinct theological concepts that Titus itself keeps structurally separate. Fencing means assigning each concept its own word and prohibiting drift between them.
| Crowded Neighborhood | Concepts Competing | Fencing Rule |
|---|---|---|
| प्रकट / प्रकाशन family (“appear/manifest/reveal”) | (a) ἐπιφάνεια — grace’s/Christ’s/kindness’s threefold structural “appearing” (2:11,13; 3:4); (b) φανερόω — God’s word “manifested” through preaching (1:3); (c) ἀποκάλυψις-family “revelation” (background from Galatians baseline, not in Titus itself but liable to bleed in via cross-reference). | प्रकट होना reserved for the ἐπιφάνεια inclusio only; प्रकट किया used for φανερόω (1:3) as a distinct, non-inclusio event; प्रकाशन reserved for the Galatians-baseline revelation entry and not introduced into Titus vocabulary at all unless a direct cross-reference requires it. |
| पवित्र / शुद्ध / सतीत्व purity-family | Three distinct Greek purity roots converge on adjacent Hindi vocabulary: ἅγιος (holy, TM पवित्र), καθαρός (clean/pure, TM शुद्ध, and Titus 1:15’s paradox “to the pure all things are pure”), ἁγνός (chaste, specifically marital-sexual purity, 2:5). | Each Greek root keeps its own dedicated Hindi word with zero substitution: ἅγιος→पवित्र, καθαρός→शुद्ध (never used for “holy”), ἁγνός→सतीत्व (never collapsed into either of the other two). Titus 1:15 additionally requires a translator note since it is a proverb subverting ritual-purity thinking, not affirming it — a live risk given that Hindi readers carry a strong prior śuddha/aśuddha (ritual purity/impurity) framework that could cause the verse to be read as endorsing exactly the legalism it dismantles. |
| व्यवस्था-उल्लंघन vs. पाप vs. अधर्म (“lawlessness/sin” family) | ἀνομία (2:14, opposition to God’s revealed order) sits adjacent to the baseline’s already-fenced पाप (general sin) and to the already-forbidden अधर्म (rejected for “sin” because it imports dharma-cosmology). Titus adds a third near-neighbor without adding a third safe word automatically. | व्यवस्था-उल्लंघन (built compositionally on the already-safe व्यवस्था) is reserved specifically for ἀνομία; पाप remains reserved for general ἁμαρτία contexts if any occur; अधर्म remains categorically forbidden across both terms, extending the baseline prohibition rather than creating a new exception. |
| काम vs. कर्म (“works/deeds” family) | This is the single most crowded neighborhood in the entire curriculum. Ordinary spoken/written Hindi treats कर्म and काम as near-interchangeable synonyms for “deed, action, work” in non-technical registers (e.g., “अच्छा कर्म करो” is completely natural, unmarked Hindi for “do a good deed”). Titus’s six-fold “good works” refrain sits exactly inside this overlap zone, where the theologically wrong word is also the stylistically more natural one. | Absolute, register-overriding fencing: काम is non-negotiable for every ἔργα occurrence in Titus, even where कर्म would read more idiomatically to a native speaker. This must be stated as an explicit override of normal stylistic preference in the Phase 2 system prompt, not left to translator discretion, because natural register is precisely the vector by which drift would occur. |
| गुरु / प्राचीन / बुज़ुर्ग / अध्यक्ष (“teacher/elder/authority” family) | Four distinct authority-models compete: the church office of 1:5–9 (qualification-based, plural, local), generic age-based respect (बुज़ुर्ग), and the pervasive Indian guru-model (personal charismatic/ascetic-attainment authority, often singular and quasi-permanent). Titus’s age-category “older men” (πρεσβύτης, 2:2) additionally shares a root with the office-word πρεσβύτερος, risking accidental conflation within Hindi itself. | गुरु is excluded from this doctrine’s vocabulary entirely. प्राचीन/अध्यक्ष reserved strictly for the qualified office (1:5–9); बूढ़े पुरुष/बूढ़ी स्त्रियाँ reserved strictly for the age-categories of 2:2–3, with an explicit glossary note (already recorded in 08_core_glossary.md) flagging the root-sharing risk for QA reviewers. |
| स्वामी (“master/lord” family) | The baseline explicitly forbids स्वामी for κύριος/God specifically because “Swami” is a common honorific for Hindu ascetic teachers/renunciants. Titus reintroduces a legitimate human referent (δεσπότης, 2:9, a slave’s earthly master) that needs a word — and स्वामी is the natural one — creating tension between the baseline’s category-level prohibition and this passage’s need. | स्वामी is permitted only for the human household-master referent of 2:9, never for God/Christ, and every occurrence must sit inside a sentence whose grammatical subject/object structure makes the human referent unmistakable without relying on the reader’s prior theological knowledge — do not rely on capitalization or context alone, since Devanagari carries no case distinction equivalent to English capitalization of “Lord” vs. “lord.” |
| भक्ति (“godliness/devotion” family) | The single Hindi word carries the entire weight of Titus’s core ethical vision (εὐσέβεια, 1:1; 2:12) while simultaneously being the primary vernacular and theological term for Hindu bhakti-mārga devotional religion — the crowding here is not with another biblical term but with the destination culture’s own dominant religious vocabulary. | No substitute word is proposed (भक्ति remains the established BSI rendering and no adequate alternative exists), but every occurrence carries a mandatory standing translator note distinguishing Spirit-produced, grace-trained Christlike character from generic devotional bhakti toward any chosen personal deity (iṣṭa-devatā). This is a fencing-by-annotation strategy rather than a fencing-by-substitution strategy, since substitution is not available. |
3. Transliteration vs. Paraphrase Decisions
Titus offers few new transliteration candidates (most proper names and established Christian technical terms are already fixed in the baseline), but the following decisions are recorded for completeness and consistency with the baseline’s transliteration standard.
| Term | Decision | Rationale |
|---|---|---|
| Trustworthy saying formula (πιστὸς ὁ λόγος, 3:8) | Paraphrase: यह वचन विश्वासयोग्य है | No transliteration candidate exists; this is a discourse-marker formula, not a proper noun or unique theological technical term. Fixed phrasing required for cross-document consistency if other Pastoral Epistles are added later. |
| Titus (proper name) | Transliterate: तीतुस | Standard BSI proper-name convention, parallel to Paulus (पौलुस), Timothy, etc. Not previously listed in the Romans/Galatians baseline since Titus does not appear there; establishing here as the letter’s addressee. |
| Crete / Cretans (1:5,10,12) | Transliterate: क्रेते / क्रेतीय | Geographic proper names follow standard BSI transliteration; the Cretan-liar proverb (1:12) is a quoted Greek cultural saying and should be rendered as reported speech, not adapted into an equivalent Hindi proverb, to preserve its status as Paul’s citation of a pagan source rather than his own teaching. |
| Zenas, Apollos, Artemas, Nicopolis (3:12-13) | Transliterate per BSI proper-name conventions | Standard practice; no theological content at risk. |
| παιδαγωγός-family vocabulary (law_as_guardian, carried over conceptually from Galatians but not textually present in Titus) | Not applicable — no occurrence in Titus | Noted only to confirm this Galatians-specific transliteration/paraphrase decision does not need re-litigation for this curriculum. |
| ἐπιφάνεια (appearing) | Paraphrase, not transliteration: प्रकट होना | A transliterated “एपीफेनिया” was considered and rejected: transliteration would strand the term as an opaque technical import with no inherent meaning to a first-time Hindi reader, whereas the doctrinal risk here is precisely about implied meaning (avatar-collision) — the safer path is a plain-meaning paraphrase carrying a translator note, not an opaque loanword that would require the note to do all the work with no semantic anchor at all. |
| παλιγγενεσία (regeneration) | Paraphrase (fixed compound), not transliteration | Same logic as above, with added urgency: a transliterated “पालिन्गेनेसिया” would avoid the पुनर्जन्म collision entirely but would communicate nothing to the target reader and would likely be silently “fixed” by a well-meaning downstream editor back toward a natural-sounding but forbidden calque. The paraphrase नये जन्म का धोना is judged safer specifically because it pre-empts that failure mode by giving editors a natural-sounding, already-correct alternative to gravitate toward. |
| δεσπότης / master (human) | Paraphrase using existing word: स्वामी (contextually restricted) | Not a transliteration case; recorded here to confirm no invented neologism is preferred over the existing (context-restricted) word. |
General transliteration principle for Titus: transliterate only proper names and already-established liturgical formulas (आमीन); paraphrase or fix a descriptive compound for every theological technical term, because Titus’s short letter format gives readers very little surrounding context to infer the meaning of an unfamiliar loanword, unlike a longer epistle where repeated exposure across many verses could gradually anchor an opaque transliteration.
4. Ranked List of Highest-Risk Ambiguities
Ranked by combined severity (doctrinal centrality × probability of natural drift toward a forbidden reading × frequency/structural weight in the letter).
- Titus 3:5’s stacked clause (“not because of works done by us in righteousness, but according to his own mercy”) — three fenced terms (काम/धार्मिकता/उद्धार) co-occurring in one sentence that is itself the letter’s flattest karma-collision point; any single lapse collapses the entire grace/works antithesis into “duty comes through duty performed in duty.” Sentence-level, not term-level, theologian review mandatory.
- “Washing of regeneration” (3:5) — the unique risk of arriving at the forbidden पुनर्जन्म by direct, individually-defensible morpheme translation (πάλιν + γένεσις) rather than by any obviously wrong word choice; the error is invisible at the single-word level and only detectable by someone who already knows the target risk.
- Titus 2:13’s single-referent “great God and Savior” — Hindi coordination habits and the destination culture’s default supreme-deity/avatar distinction make the two-referent misreading the unmarked, natural reading rather than an aberration; a translator would need active resistance to produce the correct one-referent reading.
- The six-fold “good works” refrain vs. कर्म’s superior naturalness — the theologically correct word (काम) is less idiomatic than the theologically wrong word (कर्म) in ordinary registers, meaning ordinary stylistic instinct actively works against doctrinal accuracy across all six occurrences.
- Submission vocabulary across 2:5 (wives), 2:9 (slaves), 3:1 (civil rulers) — no existing Hindi word cleanly separates God-ordered relational submission from birth-conferred caste/gender subordination; the ambiguity is structural to the language, not a translator’s error, and recurs in three distinct high-sensitivity social contexts within one short letter.
- भक्ति for godliness — the identical word simultaneously names the target Christian virtue and the dominant competing religious system’s core devotional category; there is no substitute word, only annotation, making this a permanently unresolved (mitigated, not eliminated) risk for the life of the curriculum.
- ἐπιφάνεια’s threefold “appearing” inclusio (2:11,13;3:4) — avatar-collision risk compounded by the fact that the term structures the entire core passage; an error here does not corrupt one verse but the passage’s organizing skeleton.
- Titus 1:15’s purity paradox (“to the pure all things are pure… to the defiled, nothing is pure”) — uniquely inverted risk: this is the one place in the letter where the danger is not drift toward a Hindu-adjacent reading but a reader’s correct prior familiarity with śuddha/aśuddha categories causing them to hear the verse as affirmation rather than subversion of ritual purity thinking.
- व्यवस्था-उल्लंघन vs. अधर्म for “lawlessness” (2:14) — lower frequency (single occurrence) but high stakes, since अधर्म is the more natural, more frequently encountered Hindi word in general religious discourse, creating the same “natural-but-wrong” drift pattern seen in the काम/कर्म case at smaller scale.
- “Divisive person” vs. the धर्म-rooted heretic-vocabulary family (3:10) — lower frequency but categorically important, since drift here would retroactively reframe the entire letter’s Crete-controversy material (chapter 1) as a right-religion/wrong-religion dispute rather than Paul’s narrower concern with congregational factiousness and false teaching.
- στέκω/πρεσβύτης vs. πρεσβύτερος root-sharing (“older men” vs. “elder office”) — an internal-Hindi-only risk (both English source terms are visually similar in Hindi: बूढ़े पुरुष / प्राचीन are dissimilar enough, but drafting shortcuts could conflate the underlying qualification logic of 1:6-9 with the age-virtue-list of 2:2 if reviewers are inattentive).
- स्वामी for the human master of 2:9 against the baseline’s God-referent prohibition — lower probability of doctrinal catastrophe (context almost always disambiguates a slave/master household scene) but included because it is the one place in Titus where a category-level baseline rule (never स्वामी for the divine) must be locally suspended for a human referent, creating a rule-exception that downstream automated validation scripts must be specifically configured to permit rather than flag as an error.
5. Summary Recommendations for Phase 2
- Load this document alongside
08_core_glossary.md,assets/translation_memory.json, andassets/bible_term_registry.jsonbefore any Titus segment translation begins. - Elevate Titus 3:5 and Titus 2:13 to sentence-level (not merely term-level) mandatory theologian review, given the stacked-risk findings in Section 4, items 1 and 3.
- Configure automated validation to flag कर्म and अधर्म as forbidden-on-sight anywhere in Titus output, in addition to the baseline’s existing forbidden-term list, given the “natural-but-wrong” drift pattern documented for both.
- Attach standing (not one-time) translator notes to भक्ति, अधीन होना/submission-family occurrences, and प्रकट होना/appearing-family occurrences — these are not one-off footnote items but recurring annotation requirements across the whole letter.
- Treat the नये जन्म का धोना phrase as a locked string with no permitted paraphrase variation, given its unique morpheme-level collision risk with पुनर्जन्म documented in item 2 of Section 4.
This document extends but never contradicts the baseline Romans/Galatians Language Package. Where this analysis and the baseline appear to diverge, the baseline’s established rendering governs, and this document’s role is limited to identifying the additional Titus-specific risk context around that established rendering.