GMEOW Language Core Module

• IRI: https://blackcatinformatics.ca/gmeow/slices/language
• Tier: core
• Group: core

What This Slice Covers

This slice owns 13 terms and contributes 17 mapping or projection rows. Use it when its terms match the native fact you want to preserve; use the linkage tables to see how those facts leave GMEOW for consumer vocabularies.

Dependencies

• gmeow:slices/kernel
• gmeow:slices/names

Consumers

• Every language-tagged literal (private-use-language-tag); names' romanizations; software's languages (P16 slim core).

Local Map

Examples

Multilingual Document

• Source: slices/core/language/examples/multilingual-document.ttl
• GMEOW terms: gmeow:InformationObject, gmeow:Language, gmeow:bcp47Tag, gmeow:langEnglish, gmeow:languageTag, gmeow:title, gmeow:writtenInLanguage

# SPDX-FileCopyrightText: 2026 Blackcat Informatics® Inc. <paudley@blackcatinformatics.ca>
# SPDX-License-Identifier: CC-BY-4.0
#
# Worked example: language as a first-class entity. A gmeow:Language is a
# reified individual carrying its gmeow:bcp47Tag — not a bare string column — so a
# document can be gmeow:writtenInLanguage SEVERAL co-equal languages at once
# (a bilingual manual), with no primary language privileged. The Language objects
# are minted (there is no closed enum of languages); the document is an
# InformationObject, the domain of gmeow:writtenInLanguage.
@prefix gmeow: <https://blackcatinformatics.ca/gmeow/> .
@prefix ex:    <https://blackcatinformatics.ca/gmeow/examples/language/> .
@prefix rdfs:  <http://www.w3.org/2000/01/rdf-schema#> .

# --- Languages: first-class, BCP-47-tagged. English already exists as the core
#     seed gmeow:langEnglish — reused, not re-minted. Japanese has no seed, so it
#     is minted, declaring its internal private-use tag for @lang annotations.
ex:japanese a gmeow:Language ;
    rdfs:label        "Japanese"@en ;
    gmeow:bcp47Tag    "ja" ;
    gmeow:languageTag "ja" .

# --- A bilingual document: two co-equal languages, neither primary. The title is
#     carried as separate co-equal language-tagged literals, one per language.
ex:manual a gmeow:InformationObject ;
    gmeow:title             "Installation Manual"@en , "インストールマニュアル"@ja ;
    gmeow:writtenInLanguage gmeow:langEnglish , ex:japanese .

Terms

Classes

Properties

Individuals

Linkages

• Rows: 17
• Projection profiles: ontolex, schema-org
• External vocabularies: lime, ontolex, rdf, schema, wd

Guide

Language — first-class languages, registry-independent by design

Slice: https://blackcatinformatics.ca/gmeow/slices/language · tier: core The slim language foundation every slice depends on: Language, WritingSystem, the private-use-language-tag tag machinery, and the seed languages the framework itself speaks.

The standard pattern — inLanguage "ja" — makes a registry code the identity of a language. GMEOW inverts this: a language is a first-class information object with a self-minted IRI, and every registry (BCP-47, ISO 639, Glottolog, Wikidata) is an optional alignment, never identity (Principle 5: bridge by reference). The payoff is anti-colonial in both directions (Principle 9): a language is not less real for lacking a committee-issued code — a conlang, an Indigenous language absent from a Western registry, or an AI-generated language is exactly as first-class as English — and a richly-coded language is not flattened to whichever single tag a schema happened to standardize. Language sits in core (Principle 16) because every GMEOW literal depends on it: this is the slim half of the slice-dependency doctrine dependency split, with the rich sociolinguistic machinery (proficiency, varieties, diachronic states, conlang lineage) in the languages extension.

Internally, GMEOW's own literals carry private-use tags (@en, never @en) so that the canonical graph never pretends a registry tag is ground truth; standard BCP-47 is reconstructed on projection (Principle 4: one canonical source, lossy projections generated outward).

The classes

gmeow:Language

A system of signs and rules for communication or computation, under gmeow:InformationObject. Identity is the IRI; codes are data (languageCode), registry coreference is skos:exactMatch + gmeow:authorityLink. Its names are co-equal gmeow:Appellations — endonym and exonym, never preferred-vs-alternate (the names doctrine); the scripts it is written in bind co-equally via the extension's WritingSystemUsage.

gmeow:WritingSystem

A script as a first-class object — Latin, Han, Hiragana, Arabic, Braille, or a bespoke conlang/AI script — with scriptCode (ISO 15924 when one exists), a type, and a text direction. Declared here (one-defining-slice rule) so names and documents can reference scripts; bound to languages co-equally and simultaneously through the extension — a language with three concurrent scripts (Japanese) is the normal case, not an edge case.

The tag machinery

gmeow:languageTag

The internal private-use BCP-47 tag ("en") that anchors @lang annotations on the framework's literals to a first-class Language individual. Functional: one internal tag per language — this is the join key between literal-space and IRI-space.

gmeow:bcp47Tag

The optional external tag ("en", xsd:language) used only when projecting into vocabularies that demand standard tags. Non-functional: one language legitimately exposes several BCP-47 tags across script/region/variant contexts — another reason a single flat tag was never going to be identity.

gmeow:languageCode

Any registry code — BCP-47 "ja", ISO 639-3 "jpn", a Glottocode — as a see-also alignment value, never identity. Non-functional: a language carries codes in several registries at once, and a code-less language carries none and loses nothing.

The seed languages

gmeow:langEnglish · gmeow:langMandarin · gmeow:langFrench

The languages the framework itself speaks — the individuals anchoring en, zh, and fr on GMEOW's own labels and definitions. They are ordinary data, not schema: any slice or dataset mints further languages the same way (the exhaustive reference catalog is the design).

ex:langKlingon a gmeow:Language ;            # code-less, fully first-class
    rdfs:label "tlhIngan Hol"@und ;
    gmeow:languageTag "und" ;
    gmeow:languageCode "tlh" .               # alignment, not identity

Formal languages and transliteration

gmeow:FormalLanguage · gmeow:ProgrammingLanguage

Grammar-defined languages with machine modality and no native speakers — programming, markup, query, logic, schema languages. A thin structural split (the sociolinguistic facets simply don't apply), with ProgrammingLanguage as the first-class target of writtenInLanguage for the software slice's source trees.

gmeow:transliterationScheme · gmeow:TransliterationScheme

How a romanization was derived — Hepburn vs Kunrei, Pinyin vs Wade-Giles — attached to a gmeow:Appellation beside its gmeow:romanization. This is the bridge into the names slice's doctrine that a romanization relates a name only to a transliteration of itself; recording the scheme makes the derivation reproducible. The major schemes are catalogued as FnO functions in the projection layer (projections/transforms.fno.ttl).

gmeow:writtenInLanguage

The generic written-in relation (FRBR's language-of-expression, domain-wide): any InformationObject — a document, an expression, a source tree, an inscription reading — is written in a first-class Language IRI, never a code literal. Non-functional: content mixes languages, and a codebase uses several. The registry-independent replacement for the inLanguage "ja" anti-pattern GMEOW's own doctrine names.

Solver and projection notes

Tag work is computation, not assertion (Principle 12): reconstructing a standard BCP-47 tag from bcp47Tag (+ script/region context), executing a transliteration scheme, and matching a name's language to an audience locale all happen in the solver/projection layer — the canonical graph stores the first-class objects and their alignments, nothing derived. Projections to schema.org (schema:inLanguage), Dublin Core (dcterms:language), and Wikidata are downcasts; the private-use-tag discipline is canonical and internal.

Dependencies

Depends on kernel and names (Appellation, romanization). Depended on by effectively everything: every private-use-language-tag literal in every slice resolves here, names' romanizations cite its schemes, and software's writtenInLanguage targets its classes (P16 slim core).