Source-audited platform manualበምንጭ የተመረመረ የplatform መመሪያ

Texts, Sources, Methodology & Coverageጽሑፎች፣ ምንጮች፣ ዘዴ እና የመረጃ ሽፋን

A platform-wide guide to textual authority, corpus status, pronunciation/audio policy, coverage boundaries and research-claim limits.

የጽሑፍ ስልጣን፣ የcorpus ሁኔታ፣ የአነባበብ/ድምፅ ፖሊሲ፣ የመረጃ ሽፋን ወሰኖች እና የምርምር የይገባኛል ገደቦችን የሚያብራራ የplatform አቀፍ መመሪያ።

PT1904primary teaching textዋና የማስተማሪያ ጽሑፍ

83Grammar chaptersየGrammar ምዕራፎች

5,460vocab lemmasየVocabulary lemmas

2Interlinear authority statesየInterlinear authority states

The platform is layered: teaching text, reference corpora and research evidence are not the same thingየማስተማሪያ ጽሑፍ፣ የማጣቀሻ corpus እና የምርምር ማስረጃ አንድ ነገር አይደሉም።

GTM Biblical Greek deliberately distinguishes a primary teaching text, secondary textual comparators, lexical/reference datasets, integrated token corpora, source-supplied grammatical analysis and exegetical research layers. The manual must preserve these boundaries because a result is only as strong as the layer that produced it.

GTM Biblical Greek የተለያዩ የጽሑፍና የምርምር ምንጮችን በተለያዩ ዓላማዎች ይጠቀማል። lexical/reference datasets፣ token corpora፣ source-supplied grammar እና exegetical research layers እያንዳንዳቸው የተለየ የምርምር ሚና አላቸው። ምንጩንና source/layer ሁኔታውን ሳይለዩ ውጤት መተርጎም የለበትም።

Greek text policyየግሪክ ጽሑፍ ፖሊሲ

Layer	Audited policy
Primary GTM teaching text	1904 Greek Orthodox Patriarchal / Antoniades tradition (PT1904) is identified by the platform as the primary teaching text.
Secondary critical comparators	NA28 / UBS5 are presented as secondary comparison witnesses in selected Grammar/Interlinear contexts, not as the primary running GTM teaching text.
Exegetical scriptural witnesses	The Exegetical Concordance can expose Greek OT/Swete-oriented witness material and other source evidence. Such comparison is evidence for research; it does not by itself prove an exact Vorlage.

Future-version policy: if GTM later adds NA29/UBS6 comparators, the manual should name them only after the live application and source policy are actually updated.

ዋናው የማስተማሪያ ጽሑፍ PT1904/Antoniades ነው። NA28/UBS5 በተመረጡ የጽሑፍ ንጽጽሮች ላይ እንደ secondary comparators ይጠቀማሉ። Exegetical Concordance የGreek OT/Swete witness evidenceን እና ሌሎች የምንጭ ማስረጃዎችን ሊያሳይ ይችላል። እነዚህ ምንጮች እንደ አንድ የተቀላቀለ ጽሑፍ መወሰድ የለባቸውም።

Pronunciation and audio policyየአነባበብ / የድምፅ ፖሊሲ

Modern Greek pronunciation is normative for GTM teaching. Erasmian-style pronunciation belongs only in historical/comparative discussion.

Application	Audio behavior
NT Greek 310	Approved fixed recordings only when an approved recording exists. The learner UI does not substitute unapproved browser/device speech for missing fixed audio.
Vocabulary administration	Device speech may be auditioned as a non-authoritative preview; fixed candidates pass explicit listening/technical QA before approval.
Interlinear	Hear Greek uses browser/device Greek speech synthesis (el-GR) as a reading aid. This is not the same audio authority as NT Greek 310.

Modern Greek pronunciation የGTM መደበኛ የአነባበብ ፖሊሲ ነው። NT Greek 310 የተፈቀዱ ቋሚ ቀረጻዎችን ይጠቀማል፤ Interlinear ደግሞ Hear Greek በbrowser/device የel-GR voice ሊጠቀም ይችላል። ስለዚህ ድምፁ ከdevice ወደ device ትንሽ ሊለያይ ይችላል።

GOLD Grammar corpus identity and scholarly boundaryየGOLD Grammar corpus ማንነት

The audited Grammar repository verifies a read-only scholarly corpus with fixed expected counts:

Chapters	83	Sections	1,264
Paradigms	120	Quiz records	823
Terms	765	Chapter-term links	851

The application displays a SHA-256 corpus identity and reads the Grammar SQLite database in query-only/read-only mode. Student reading progress is stored separately in the learning platform; it does not modify the scholarly Grammar corpus.

The underlying corpus contains paradigm and quiz records, but the audited chapter UI currently shows counts rather than rendering those records as full interactive content. The manual must describe the UI, not merely the database.

GOLD Grammar 83 ምዕራፎች፣ 1,264 sections፣ 120 paradigms፣ 823 quiz records እና 765 የሰዋሰው ቃላትን ይዟል። ይህ corpus በSHA-256 የተረጋገጠ ሲሆን databaseው በquery-only/read-only ሁኔታ ይጠቀማል። የተማሪ እድገት ግን ከGrammar corpus ተለይቶ በplatform ላይ ይቀመጣል።

Vocabulary corpus and editorial-status transparencyየVocabulary corpus እና የአርትዖት ሁኔታ

The Complete Vocabulary Reference exposes a 5,460-lemma NT Greek reference corpus with Greek, English, Amharic, Strong’s, frequency and location information where supplied. It is distinct from the 310-word learning corpus.

Amharic editorial status is displayed as carried by the source. Labels such as review-pending or machine-assisted-draft must not be presented as final human review.

Complete Vocabulary Reference 5,460 lemmas ያሉት ሙሉ የNT Greek ቃላት ማጣቀሻ corpus ነው። NT Greek 310 ግን በጥናት ዓላማ የተመረጡ 310 ቃላት ያሉት የመማሪያ corpus ነው። የeditorial status ምልክቶች እንደ `review-pending` ወይም `machine-assisted-draft` ያሉ የይዘት ሁኔታዎችን ይገልጻሉ፤ እንደ final human-reviewed status መቆጠር የለባቸውም።

Interlinear authority: Production FINAL vs Research FixtureየInterlinear ምንጭ ሁኔታ፦ Production FINAL እና Research Fixture ተመሳሳይ የauthority ደረጃ አይወክሉም። Production FINAL የተጠናቀቀውን የማስተማሪያ ጽሑፍ ይወክላል፤ Research Fixture ግን ለalignment/research አወቃቀር የተዘጋጀ መረጃ ነው።

Authority	Meaning
Production FINAL	Finished teaching corpus profile used for learner-facing Greek/Amharic/English teaching layers.
Research Fixture	Retained to exercise alignment/source-order architecture or research evidence. It is explicitly not completed linguistic review.

Production FINAL በሕዝብ ፊት የሚቀርብና የተረጋገጠ የGTM corpus ወይም የapplication ሁኔታ ነው። Research Fixture ግን ለalignment/research architecture የሚያገለግል የምርምር መዋቅር ነው። Research Fixture እንደ ሙሉ linguistic review የተጠናቀቀ የመጨረሻ ውጤት መቆጠር የለበትም።

Lexical Concordance: two modes, two coverage claimsLexical Concordance፦ ሁለት ሁነታዎች፣ ሁለት የሽፋን ይገባኛሎች

Mode	Coverage
Lemma Concordance	Full-NT lemma reference layer with verse-location data.
Forms & Morphology	Only currently integrated token occurrences; can filter by authority such as production-final and research data.

Token counts and verse counts are not interchangeable: one lemma/form can occur multiple times in one verse. The lemma’s total NT occurrence count can therefore exceed the number of distinct verse locations.

Lemma mode ዋናውን የlemma ማጣቀሻ ያሳያል። Forms & Morphology ደግሞ ከInterlinear የሚመጡ የtoken corpus ቅርጾችን እና morphologyን ይመረምራል። የNT morphology coverage በተለየ የcoverage ሁኔታ ይመዘናል። token count እና verse count የተለያዩ መለኪያዎች ናቸው፤ እርስ በርሳቸው መቀየር የለባቸውም።

Grammatical Concordance evidence layers የGrammatical Concordance የማስረጃ ንብርብሮች

The Grammatical Concordance combines morphology, source-supplied syntax, lemma/form/occurrence comparison, and advanced research/QA layers. These evidence layers must not be treated as interchangeable.

Lemma layer: the ordinary Lemma view uses the server-precompiled PT1904 lemma index. The verified index contains 5,383 lemmas derived from the full 140,488-token PT1904 corpus covering 27/27 New Testament books and 260/260 canonical chapters.
Parsed Forms: use when the research question concerns inflected forms and morphology rather than only lemma distribution.
Occurrences: use when the question requires occurrence-level evidence, reference context, or comparison of individual token instances.
Source-supplied syntax: structured syntactic data supplied by an external/source layer. It is valuable for finding and comparing patterns, but it is not automatically an independently human-reviewed GTM judgment.
Comparison / QA layers: advanced research-maintenance tools for agreement, ambiguity, alignment, provenance, syntax comparison and related auditing tasks. They are not the normal starting point for ordinary grammatical study.

Precompiled production architecture:Precompiled production architecture፦ ordinary Lemma browsing does not require the browser to reconstruct the full PT1904 corpus at startup. The lemma index is precompiled on the GTM server and served as ready-to-use data. Full occurrence-level corpus processing is invoked only when a mode or research task requires it.መደበኛ lemma browsing ሲደረግ browserው በstartup ጊዜ ሙሉ PT1904 corpusን እንደገና መገንባት አያስፈልገውም። Lemma indexው በGTM server ላይ ቀድሞ የተዘጋጀ ሲሆን ready-to-use data እንደሆነ ይቀርባል። ሙሉ occurrence-level corpus processing የሚጀምረው mode ወይም research task ሲፈልገው ብቻ ነው።

Source-supplied syntax is evidence, not an automatic GTM judgment.Source-supplied syntax ማስረጃ ነው፤ በራሱ የGTM ውሳኔ አይደለም። Use it to locate and compare patterns; important grammatical and exegetical claims still require contextual evaluation.Patternsን ለማግኘትና ለማነጻጸር ይጠቀሙበት፤ ጠቃሚ grammatical እና exegetical claims ግን አሁንም contextual evaluation ይፈልጋሉ።

Grammatical Concordance የmorphology፣ syntax እና lemma/form/occurrence መረጃን ያጣምራል። የQA ምልክቶች የcoverage ወይም የsource ሁኔታን ለመግለጽ ይጠቅማሉ። Syntactic label ከsource-supplied syntax የመጣ ከሆነ በGTM የሰው ግምገማ የተረጋገጠ ነው ብሎ መገመት የለበትም። ሁልጊዜ የsource እና context ሁኔታውን ይመልከቱ።

Exegetical Concordance: evidence claims and limitsExegetical Concordance፦ የማስረጃ ይገባኛሎች እና ወሰኖች

Curated means editorial/source checking within the GTM research workflow; it does not mean independently peer-reviewed scholarship.
Machine discovery provides candidate-ranking assistance. Similarity scores are not literary-dependence probabilities.
Allusion/echo relationships are graded research judgments and should not be conflated with explicit quotations.
Research-RC / release-ready language describes an internal research release state, not external peer-reviewed finality.
Greek OT/Swete comparison evidence does not by itself assert that Swete is the exact source text behind the NT author.

Reader View should remain the default scholarly-use surface. R21–R33 Research Audit layers document methodology, claim boundaries, witness conditions, bibliography, validation, freeze and external-review preparation.

`Curated` የሚለው በGTM የተመረጠና የተደራጀ ምርምር መረጃን ያመለክታል፤ independently peer-reviewed ማለት አይደለም። `Machine discovery` በማሽን የተገኙ እጩ ግንኙነቶችን ያሳያል፣ እነዚህም probability ወይም ሌላ የእጩ ሁኔታ ሊኖራቸው ይችላል። Allusion/echo ከquotation የተለየ የማስረጃ ደረጃ አለው። `Research-RC ready` የምርምር መልቀቂያ ሁኔታን ያመለክታል፤ independent peer review እንደተጠናቀቀ አያሳይም።

Methodological reproducibility: server, browser and snapshotsእንደገና ማረጋገጥ፦ server፣ browser እና snapshots

Account-backed learner records are server-persistent. Some static research workspaces use browser localStorage/session state. Grammatical workspace snapshot/export mechanisms are therefore important when research state must be portable or reviewable outside the current browser.

የተማሪ መዝገቦች server-backed ናቸው። አንዳንድ static research workspaces ግን stateን በbrowser localStorage/session ውስጥ ሊያስቀምጡ ይችላሉ። ምርምር ሁኔታን ከአንድ device/browser ውጭ ለመውሰድ ካስፈለገ workspace snapshot/export መጠቀም ይገባል።

What must be verified before final publicationከመጨረሻ ህትመት በፊት ምን መረጋገጥ አለበት?

□ Verify all live corpus labels against current production data.
□ Update any comparator policy if NA29/UBS6 are actually introduced.
□ Correct hard-coded or stale authority wording before screenshot annotation.
□ Finalize legal/privacy policy text.
□ Freeze a manual version against a known application release so screenshots and controls cannot silently drift.

□ Live corpus ከproduction data ጋር እንደሚጣጣም ያረጋግጡ። □ Comparator መረጃ ከlive application ጋር እንደሚስማማ ይፈትሹ። □ የመረጃ ብዛቶች hard-coded ሳይሆኑ ከትክክለኛው source እንደሚመጡ ያረጋግጡ። □ የምንጭ እና የcoverage መግለጫዎችን ያረጋግጡ። □ Manual ከapplication release ጋር መስማማቱን ይፈትሹ።