CaseNom

CaseNom is a framework for labeling nominal case information in linguistic annotation and natural language processing. It provides a compact, language-agnostic tag set designed to indicate the syntactic role of nouns across languages, facilitating consistent annotation and data exchange among corpora and tools. The core set includes NOM for nominative, ACC for accusative, GEN for genitive, DAT for dative, LOC for locative, INS for instrumental, and VOC for vocative, with an UNKNOWN option for unclear cases. The tag set is intended to be extensible to accommodate language-specific phenomena and cliticized or multi-part case systems.

Implementation and usage: In annotation pipelines, CaseNom tags are attached to tokens alongside lemmas, e.g., dog-NOM or ball-ACC. The framework supports both manual annotation and automated tagging, with mapping rules that align language-specific case markings to the CaseNom core. It is designed to interpolate with existing annotation schemes and morphological analyzers.

Origins and scope: CaseNom was proposed as a standardization concept in cross-language grammar and corpus work to improve interoperability among datasets and NLP models dealing with case information. It is not tied to a single language family and aims to adapt to analytic and inflectional languages alike.

Applications and limitations: It serves corpus linguistics, syntactic parsing, morphological analysis, lexicon development, and machine translation. However, CaseNom cannot fully capture all language-specific case mappings and relies on supplementary annotations for certain constructions.

Example: English: the dog-NOM chased the cat-ACC. German: Der Hund-NOM jagt die Katze-ACC. Turkish: Kedi-NOM köpeği-ACC gördü. See also: grammatical case, nominative case, annotated corpora, multilingual NLP.