lext

Lext is an open, interoperable data format and schema designed for lexical data used in dictionaries, language resources, and natural language processing. It provides a consistent representation of lemmas, senses, and lexical relations across languages, with an emphasis on both human readability and machine parsability. Lext aims to complement existing standards by offering a compact, extensible model that can be adopted by publishers, researchers, and software projects.

Its core data model defines entries with fields such as id, language, lemma, orthography, phonetic transcription, part of speech, and inflectional paradigms. Each entry contains one or more senses, each sense carrying a definition, example usages, sense-level notes, etymology, and semantic relations (synonyms, antonyms, hypernyms). Optional fields support frequency data, domain labels, regional variants, and translations to other languages. The model supports multilingual data by linking cross-language senses and senses across languages.

Lext was proposed by the Lext Consortium in the early 2020s as a lightweight alternative to heavier lexical standards. A reference implementation and validator were released, with ongoing contributions from academic and industry partners. The format is documented with versioning to manage compatibility and extension.

Lext relates to established standards such as the Lexical Markup Framework and TEI-based lexical resources, providing a simpler, more pragmatic option for small to medium-sized lexicons and NLP pipelines. Tooling includes validators, converter utilities, and language-specific bindings. The format is designed to be extension-friendly and openly licensed, encouraging community contributions and cross-project interoperability.

See also Lexical Markup Framework, TEI Lexical, JSON-LD.