taivutuslähde

Taivutuslähde is a term used in linguistics and language technology to refer to a data source that provides the inflectional paradigm for a lexeme or a class of words. It contains information about how a word changes form to express grammatical categories such as number, case, tense, mood or voice, and may also indicate related stem changes or irregular forms. In practice a taivutuslähde serves as the basis for generating or recognizing all inflected forms of a word.

A taivutuslähde can be part of a dictionary, a morphological analyzer, a spell checker or a machine translation system. It typically includes the lemma (the base form), part of speech, and a set of inflected forms or the rules needed to derive them. In Finnish, where inflection is rich and features like vowel harmony and stem mutation (gradation) play a major role, a taivutuslähde may encode multiple inflection classes and the corresponding derivation rules.

There are manual and automatic approaches to building taivutuslähteet. Manual creation ensures coverage of irregularities and exceptions, while automatic methods can scale to large lexicons but may require validation. Taivutuslähde is closely related to concepts such as taivutusparadigma and morphological databases; together they enable robust word form generation, parsing, and linguistic analysis.

Limitations include the need for regular updates to cover new words, loanwords, and evolving usage. As a specialized resource, a taivutuslähde reflects the inflectional patterns used in a language at a given time and may vary between linguistic resources and software systems.