LMF

LMF, short for Lexical Markup Framework, is an ISO standard designed to model lexical resources such as dictionaries and lexicons in a shared, machine-readable form. Its aim is to enable interoperability and reuse of lexical data across languages, tools, and applications in linguistics, lexicography, and natural language processing.

The framework provides a modular meta-model with a Core package and optional extensions for morphology, syntax, and semantics. At the top is a LexicalResource, which contains multiple LexicalEntry elements representing individual lemmas. Each LexicalEntry typically comprises a Lemma (the canonical form) and one or more Form objects describing inflected or related forms, including written representation and sometimes pronunciation. Senses describe the meanings associated with an entry, and may include definitions, usage examples, and semantic relations to other entries. Additional packages can capture subcategorization frames, word-form derivations, morphological features, and syntactic behavior.

LMF emphasizes standardized data categories to support consistent annotation of features such as part of speech, gender, number, tense, and mood, as well as relationships like synonyms or hypernyms. The standard is designed to be extensible, allowing language-specific extensions and mappings to other representations, and is typically serialized in XML using the LMF schemas, with various implementations also experimenting with RDF or JSON-based formats.

Adoption and use: LMF has been adopted in academic and industrial settings to model and exchange lexical resources, providing a common backbone for dictionaries, lexical databases, and NLP pipelines. It is published by ISO as Lexical Markup Framework (ISO 24613).