Language / Terminology
Describe a governed linguistic reference package connecting language and script profiles, glossary editions, terminology usage, translation assessments and locale selections.
Bundle → Layer → Finding → Questions Filled
6 bundles · 12 layers · 12 findings · 48 questions
Linguistic reference Typed language and script evidence without invented authority.
Package identity and remit
Root identity separates the governed package from every registry, glossary edition and linguistic subject.
Package identity and remit
Root identity separates the governed package from every registry, glossary edition and linguistic subject.
- Which master identifier, namespace and revision identify this reference package? identity
- Which package components and domains are included, optional or expressly excluded? classification
- Which steward controls this package and which authorities control its external records? ownership
- What distinguishes a working package from a released, superseded or retired edition? lifecycle
Language and script references
Record submitted tags and typed scheme references; separate syntax, registry validity and content-language evidence.
Language and script references
Record submitted tags and typed scheme references; separate syntax, registry validity and content-language evidence.
- Which original tag or typed language and script codes were supplied and from which registry snapshot? identity
- Does the selected profile distinguish well-formed, registered, deprecated and private-use values? validation
- What supports the asserted language or script of this content and what remains unknown? evidence
- How are mixed, undetermined, absent and unregistered variety descriptions represented without guessed tags? exception
Expression and usage Keep community usage evidence separate from globally stable concept identity.
Variety and writing context
Treat variety classification, orthography and writing context as sourced assertions within a named remit.
Variety and writing context
Treat variety classification, orthography and writing context as sourced assertions within a named remit.
- Which variety, register or orthography is described and who recognizes that description? classification
- Which languages, scripts and regions are linked by evidence rather than assumed one-to-one? relationship
- During which period and in which community context does the usage assertion apply? temporal
- How are contested names, multiple scripts and absent registry codes retained for review? exception
Concept references and designations
A designation belongs to a language and usage context; equal text cannot merge concept identities.
Concept references and designations
A designation belongs to a language and usage context; equal text cannot merge concept identities.
- Which externally mastered concept and term versions does each package designation reference? relationship
- Which definition or scope-note version explains the intended sense for this package? definition
- Which designation is preferred, admitted, hidden or deprecated under the declared usage profile? state
- How are homonyms, synonyms and competing preferred labels kept distinct in each export projection? constraint
Glossary stewardship Versioned membership and scoped editorial authority.
Glossary editions and membership
Release an explicit membership snapshot rather than allowing live upstream edits to silently rewrite an edition.
Glossary editions and membership
Release an explicit membership snapshot rather than allowing live upstream edits to silently rewrite an edition.
- Which concept and term references belong to this glossary edition and which are deliberately absent? composition
- Which release date, validity interval and upstream versions describe the edition? temporal
- How are coverage counts, missing languages and unresolved definitions reported for this exact snapshot? quality
- Which scheme links and hierarchy projections are exported without treating them as formal domain axioms? interoperability
Editorial authority and revision
A local edit proposal cannot acquire authority merely because an agent or importer supplied it.
Editorial authority and revision
A local edit proposal cannot acquire authority merely because an agent or importer supplied it.
- Who may approve a designation change or glossary release within this package? authority
- Which proposal, review and decision records justify a change from the prior edition? process
- Which contributors or consumers may read restricted notes, pending terms or licensed extracts? access
- What retires an edition and governs retention or removal of private contribution evidence? retention
Translation assessment Contextual cross-language assessments with separate work and workflow masters.
Contextual equivalence
Store the purpose and evidence of a proposed cross-language relation; matching text or tags supplies no equivalence proof.
Contextual equivalence
Store the purpose and evidence of a proposed cross-language relation; matching text or tags supplies no equivalence proof.
- Which source and target concept or term versions are compared in the stated direction? relationship
- Which profile defines exact, approximate, partial, absent or disputed equivalence for this use? classification
- Which examples and counterexamples support the equivalence assessment? evidence
- What prevents the assessment from becoming universal identity or approval outside its scope? constraint
Translation review and change impact
Keep imported workflow state, local approval and continued applicability separate.
Translation review and change impact
Keep imported workflow state, local approval and continued applicability separate.
- What translation workflow state was imported and what local review decision actually exists? state
- Which source and target content versions, transformation method and reviewers produced this assessment? provenance
- Which domain qualification and acceptance criteria are required before a translation can be used? requirement
- Which upstream change suspends the assessment and triggers another review? event
Text and locale profiles Versioned processing policies preserve originals and distinguish inference from observation.
Lexical preservation and normalization
Canonical or compatibility transformations are declared projections, never silent semantic deduplication.
Lexical preservation and normalization
Canonical or compatibility transformations are declared projections, never silent semantic deduplication.
- Which original lexical form and encoding were retained before normalization? provenance
- Which normalization form and Unicode version produced the searchable projection? process
- Which collision and round-trip checks expose distinctions lost by the chosen normalization? validation
- Which separate rendering and confusable-text checks are required for this use? security
Locale selection and fallback
Distinguish language matching, locale-data inheritance and formatting configuration.
Locale selection and fallback
Distinguish language matching, locale-data inheritance and formatting configuration.
- Which language priority list and matching algorithm select a supported language? interoperability
- What explicit default or refusal occurs when no supported language matches? decision
- Which locale-data version and explicit calendar, numbering or formatting overrides are applied? composition
- Which requested, inferred and selected values remain visible to prevent locale guesses becoming user facts? quality
Exchange and assurance Controlled projections and inspectable evidence lineage.
Interchange contract and loss
A named format alone is insufficient: each export has a profile, mapping and permitted scope.
Interchange contract and loss
A named format alone is insufficient: each export has a profile, mapping and permitted scope.
- Which TBX dialect, SKOS profile or XLIFF version is targeted and which components are mapped? interoperability
- Which identities, statuses, notes or language distinctions cannot survive this projection? constraint
- Which schema and semantic fixtures were run on the exact export bytes? validation
- Which license and recipient policy permit each included term, definition and evidence extract? access
Provenance and review assurance
Track sources, revisions and responsible roles without equating attribution with correctness.
Provenance and review assurance
Track sources, revisions and responsible roles without equating attribution with correctness.
- Which evidence entities and activities generated each admitted package assertion? provenance
- Which observation, effective and revision times are known and which remain uncertain? temporal
- Which conflicts, source checks and independent reviews remain open for this edition? quality
- How are sensitive contributor details minimized while preserving permissible audit evidence? privacy
Classifiers Filled
- Family
- World Models
- Category
- Information and virtual systems
- Entry kind
- aggregate
- Navigation path
- NAV.INF.KNW.LNG
- Domain
- INF.KNW.LNG
- Industry
- Cross-industry
- Tags
- languageterminologyinf.knw.lng
- Also called
- N9
What it is Filled
One independently versioned reference package is the aggregate root. It groups typed local profile, membership, assessment and edition records while referencing external language registries and separately mastered concepts and terms. A package need not contain every optional component. It is neither the language itself nor a universal termbase or translation engine.
In scope
- Package identity, language and script references, varieties and uncertain language observations
- Glossary membership and edition policy with references to concept and term masters
- Context-specific usage and translation assessments with review evidence
- Locale selection, normalization policy and interchange projections with explicit loss reports
Out of scope
- Authoritative global language registration and political recognition of varieties
- Concept and term master identity or general ontology axioms, delegated to WM-KNW-006 and WM-KNW-001
- Creative works, full translation memories, unrestricted corpora, translation service execution and general rights or territory masters
- Operational translation approval, certified linguistic correctness and automatic inference of a person's identity from language
Why it exists Filled
Describe a governed linguistic reference package connecting language and script profiles, glossary editions, terminology usage, translation assessments and locale selections.
Distinguishing features Derived, awaiting review
- Unlike WM-KNW-006 Concept / Term: Registry parent_ids makes WM-KNW-002 a candidate organizational parent. Keep semantic concept and term master identities external; this aggregate owns package membership, usage assertions and edition-specific projections, not a second concept master.
- Unlike WM-KNW-001 Model / Ontology: Vocabulary arrangements and external mappings are references, not imported formal domain axioms or evidence that the described world has a given structure.
- Unlike WM-XCT-011 Identifier Scheme: Language and script code schemes remain external masters. Store scheme, version and typed code references; display names cannot become identifiers.
- Unlike WM-MED-001 Creative Work: Translation assessment links source and target record versions; the creative work and its renditions remain externally mastered.
- Unlike WM-KNW-003 Intellectual Property Right and territory reference schemes: Store permission and region references only. A locale region does not establish nationality, residence, territorial sovereignty or reuse rights.
Note: Derived from boundary notes against neighbouring models.
What robots and AI may and may not do Derived, awaiting review
Must not
- Create a proposal when authority, source evidence or language is unknown; never manufacture registration or approval
- Deny access unless the package policy grants a recipient and purpose; release only the permitted projection.
May
- Check supplied tag references: Proposed, unimplemented local operation. Evaluate syntax and registry-reference status without identifying content by guess.
- Propose glossary membership: Proposed, unimplemented local operation. Stage a concept/term reference and scoped usage assertion for review.
- Record translation assessment: Proposed, unimplemented local operation. Record direction, intended use, counterevidence and review state against pinned versions.
- Preview language and locale selection: Proposed, unimplemented local operation. Produce a local selection trace using an explicit algorithm and data version.
- Preview lexical projection: Proposed, unimplemented local operation. Derive a reversible reference to original text and report projection loss.
- Prepare an authorized edition projection: Proposed, unimplemented local operation. Stage a local export with pinned membership, recipient scope and a loss report.
Note: Derived from functions, policies, CRUD and access rules; prohibitions were not authored for agents as such.
Moral aspects Missing, in the backlog
Not described yet. This gap is in the card backlog.
Owners Filled
Steward
Name a neutral knowledge steward and accountable glossary editor by role; source publishers are not package owners
Roles
- Knowledge steward
- Approve package remit, dependencies and access/retention policies
- Terminology editor
- Propose glossary membership and usage changes with evidence
- Language reviewer
- Assess language and translation evidence within declared competence and scope
- Release custodian
- Check edition pins, permissions, mappings and hold visibility before authorized release
- Evidence auditor
- Review lineage, source checks, conflicts and access logs without self-certifying independent review
Links to other meta-models Filled
references
- WM-KNW-006 - Concept and term masters; required by the terminology component, optional for a language-only package. Registry parent hint is not automatic inheritance or shared ownership.
- WM-XCT-011 - Identifier scheme authority for language/script codes; adopter must pin a compatible binding before operational use.
- WM-KNW-001 - Optional ontology or modeling context; vocabulary relations do not automatically assert formal axioms.
- WM-MED-001 - Optional external source/translated work identity; no work or rendition ownership imported.
- WM-KNW-003 - Optional external rights evidence; a reference alone never grants terminology reuse.
aligned
- https://www.w3.org/TR/skos-reference/ - Candidate scheme, label and mapping projection, with explicit integrity and loss checks.
- https://www.tbxinfo.net/tbx-about/ - Termbase exchange lead only; select and verify an exact dialect and normative specification.
- https://docs.oasis-open.org/xliff/xliff-core/v2.1/os/xliff-core-v2.1-os.html - Translation workflow references; imported state remains separate from package approval.
neighbor
- WM-KNW-006 Concept / Term - Registry parent_ids makes WM-KNW-002 a candidate organizational parent. Keep semantic concept and term master identities external; this aggregate owns package membership, usage assertions and edition-specific projections, not a second concept master.
- WM-KNW-001 Model / Ontology - Vocabulary arrangements and external mappings are references, not imported formal domain axioms or evidence that the described world has a given structure.
- WM-XCT-011 Identifier Scheme - Language and script code schemes remain external masters. Store scheme, version and typed code references; display names cannot become identifiers.
- WM-MED-001 Creative Work - Translation assessment links source and target record versions; the creative work and its renditions remain externally mastered.
- WM-KNW-003 Intellectual Property Right and territory reference schemes - Store permission and region references only. A locale region does not establish nationality, residence, territorial sovereignty or reuse rights.
What else AI and robots need to interact with it Incomplete
Identity and identifiers required Filled
- Authoritative master-system identifier and revision
- Scheme-qualified external artifact identifier with issuer context
- Package-scoped stable local identifier when no authoritative identifier exists
Direct properties not applicable Not applicable
Not applicable
Institutional or informational subject: no invented physical properties.
Recognition optional Missing, in the backlog
Not described yet. This gap is in the card backlog.
Capabilities and actions required Filled
- Check supplied tag references: Proposed, unimplemented local operation. Evaluate syntax and registry-reference status without identifying content by guess.
- Propose glossary membership: Proposed, unimplemented local operation. Stage a concept/term reference and scoped usage assertion for review.
- Record translation assessment: Proposed, unimplemented local operation. Record direction, intended use, counterevidence and review state against pinned versions.
- Preview language and locale selection: Proposed, unimplemented local operation. Produce a local selection trace using an explicit algorithm and data version.
- Preview lexical projection: Proposed, unimplemented local operation. Derive a reversible reference to original text and report projection loss.
- Prepare an authorized edition projection: Proposed, unimplemented local operation. Stage a local export with pinned membership, recipient scope and a loss report.
Hazards and failure modes optional Missing, in the backlog
Not described yet. This gap is in the card backlog.
Standards and interfaces required Derived, awaiting review
- PROV-O: The PROV Ontology
Context of use required Filled
- No national language policy, dominant language or universal glossary authority is assumed
- BCP 47 region and locale inference are processing metadata, not statements about citizenship, residence or territorial legitimacy
- The package may use any language and script supported by its reviewed profile; no ASCII restriction is imposed on actual instance data
Sources Filled
- Tags for Identifying Languages - Internet Engineering Task Force
- Language Subtag Registry - Internet Assigned Numbers Authority
- Matching of Language Tags - Internet Engineering Task Force
- SKOS Simple Knowledge Organization System Reference - World Wide Web Consortium
- Unicode Normalization Forms - Unicode Consortium
- Unicode Locale Data Markup Language - Unicode Consortium
- About TBX - TBX Council
- PROV-O: The PROV Ontology - World Wide Web Consortium
- XLIFF Version 2.1 - OASIS
Open questions
- Complete independent external review before any canonical or publishable-draft promotion.
- Verify source responses, pinned payloads, revisions, licenses and precise normative adoption profiles, including TBX dialects and language/script registry coverage.
- Implement nested instance schemas and execute mapping, normalization, matching, staleness, access and loss fixtures with exact dependencies.
- Review community and domain-specific terminology coverage, contested varieties and qualified translation acceptance without assigning universal linguistic authority.
- No exhaustive language/script inventory, dialect recognition adjudication, corpus linguistics or endangered-language coverage
- No executable instance schemas, parsers, Unicode rendering/security implementation, matching engine or certified translation workflow
- Detailed ISO 639, ISO 15924 and current ISO 30042 normative texts and TBX dialect schemas were not reviewed; their conformance is not claimed
- Exact neighbor version bindings, rights clearance and qualified domain acceptance remain adopter tasks
- Independent external review and direct HTTP source snapshot checks remain open
Machine files
Provenance
world-models research · reviewable-draft
Built from: models/wm-knw-002-language-terminology/spec.yaml