kaassõnad

Kaassõnad is an Estonian linguistic term used to describe words that frequently occur together with another word in natural language data. The concept captures fixed expressions, habitual collocations, and other pairings or short phrases that shape how language is used. Kaassõnad are not standalone meanings; their significance emerges from their relationship to the word they accompany and can vary across genres, registers, and contexts.

In practice, kaassõnad are identified through corpus analysis. Researchers examine large text collections to find words that co-occur with a target word more often than would be expected by chance. They typically employ statistical measures such as mutual information, t-score, or log-likelihood, often using a defined window of surrounding text. The resulting lists help build dictionaries of collocations and phraseology, and they support language learning resources and various natural language processing tasks.

Kaassõnad relate closely to concepts like collocations and phraseology. They highlight syntactic patterns (for example, common verb–noun or adjective–noun pairings) and semantic associations that recur in language use. While some kaassõnad form highly stable expressions, others are more flexible and context-dependent, reflecting both linguistic structure and speaker behavior.

History and usage of the concept have grown with corpus linguistics and computational methods. Kaassõnad are used in Estonian language research, in compiling dictionaries, in language education, and in computational applications such as language modeling, machine translation, and information retrieval.

See also: collocation, phraseology, corpus linguistics.