fraseleksikografi

Fraseleksikografi, or phrase-lexicography, is a branch of lexicography that studies and compiles dictionaries and databases of multiword expressions (MWEs) such as idioms, collocations, phrasal verbs, and fixed or semi-fixed sequences. It seeks to describe how these units function as entries in a language's lexicon and to provide resources for users, teachers, translators, and NLP systems.

The field covers identification, documentation, description, and organization of MWEs, including their form, meaning, morphosyntax, usage, and variation. It distinguishes idiomatic expressions from compositional phrases and addresses cross-linguistic equivalences, sense distinctions, and frequency information. It also concerns standardization and representation of MWEs in lexicographic data models.

Data sources include large corpora, learner corpora, and web data, while methods combine corpus-based statistics (frequentist measures and collocation statistics such as mutual information and t-score), pattern-based extraction, and semi-automatic annotation. Fraseleksikografi produces resources such as phraseological dictionaries, collocation dictionaries, idiom dictionaries, and phraseme inventories, usually linked to lemmas, with notes on usage, grammatical variants, examples, and, where relevant, translation equivalents.

Applications span language education, lexicography, dictionary publishing, natural language processing, machine translation, and information retrieval. Challenges include detecting irregular or opaque idioms, modeling varying forms and inflections of MWEs, cross-linguistic alignment, and ensuring data quality and interoperability.