Tyyppisana

Tyyppisana is a term used in Finnish linguistics to denote the underlying lexical item or canonical form of a word, from which inflected forms are derived. It is closely related to, and often used as, the lemma or perusmuoto. In corpus linguistics, the tyyppisana is the unit represented by a distinct lexical item, in contrast to a token, which is any occurrence of a word form in running text.

For example, the Finnish noun talo has the tyyppisana talo, while its inflected forms talon, taloa, taloja, taloissa and others are tokens derived from that same tyyppisana. A verb such as kirjoittaa has the tyyppisana kirjoittaa, and its inflected forms like kirjoitan, kirjoitat, kirjoittaa, kirjoitamme, etc., are tokens formed from that base form.

Determining the tyyppisana typically requires lemmatization or stemming, depending on the application. In dictionaries, the tyyppisana corresponds to the lemma listed as the headword. In natural language processing, identifying the tyyppisana supports efficient search, language modeling, and frequency analysis, especially when computing type-token statistics.

The concept helps distinguish between vocabulary diversity (types) and the number of word occurrences (tokens). The term is widely used in Finnish linguistic literature and Finnish-language NLP resources, and it aligns with the broader idea of a lemma or canonical form used in many languages.