collokasjoner

Collokasjoner, or collocations, are word combinations that occur together more often than chance would predict. They are a central concept in corpus linguistics, lexicography, and language teaching because they influence how natural a text sounds to native speakers. While idioms are often fixed and semantically non-literal, collocations are usually compositional, even though their habitual co-occurrence makes some pairings feel almost inseparable.

Common types include lexical collocations, such as adjective + noun (strong tea), noun + noun (traffic jam), and verb + noun (make a decision). Grammatical collocations involve predictable combinations with function words, such as depend on, rely on, or take a chance on. The term collokasjoner is used in some Scandinavian languages to refer to the same idea.

Measuring collocations relies on corpus data. Researchers analyze large text corpora to identify word pairs that co-occur more frequently than expected. Statistical measures such as mutual information, likelihood ratio, and t-scores are used to rank candidate collocations, often with thresholds that depend on corpus size and research goals. Collocation strength can vary by genre, register, and language, so findings are context-dependent.

Applications include language education, where teaching common collocations improves fluency and naturalness; lexicography, where collocations guide usage notes and examples; and natural language processing, where collocation information enhances language models, search, and machine translation. Understanding collokasjoner helps explain why certain word choices sound right in everyday language and why others feel awkward.