Ordhyppighed

Ordhyppighed, in linguistics, refers to how often words occur within a text or corpus. It can be expressed as absolute frequency (the count) or relative frequency (the proportion or per-million words). Frequency can be measured for tokens (each actual word form) or types (distinct word forms). A frequency list orders words from most to least frequent and is a common resource in lexicography, language teaching, and natural language processing.

Frequency data come from corpora—text collections intended to represent a language or domain. Results depend on corpus size, genre, time period, and language. To compare texts, counts are often normalized by the number of tokens and converted to relative or per-million frequencies. Steps such as lowercasing, lemmatization, or stemming are common; in Danish, compounding creates many forms, which affects whether related forms are counted separately or as lemmas.

Word-frequency distributions typically follow Zipf's law: a small set of function words accounts for a large share of occurrences, while many words are rare. This pattern informs language modeling and algorithm design.

Applications include lexicography (prioritizing entries), language teaching (focusing on high-frequency vocabulary), and natural language processing tasks such as search, information retrieval, text classification, and readability assessment.

Limitations include sampling bias, domain differences, and diachronic change. Frequency lists describe a corpus, not a fixed language, and must be interpreted with metadata about genre, time period, and source.