Basisform

Basisform is a term used in linguistics and computational linguistics to denote the base or dictionary form of a word, commonly referred to as the lemma or citation form. The basis form represents the canonical form from which inflected or derived variants are generated, and it is typically the form stored in lexical resources and dictionaries.

In natural language processing, basisforms are used to normalize tokens for tasks such as lemmatization, morphological analysis, and part-of-speech tagging. By mapping each surface form to its basis form, systems can consolidate related word forms and improve search, indexing, and linguistic analysis. The concept is especially important for languages with rich morphology, where a single lemma may have many inflected forms.

Terminology and usage vary by language and field. In English, the preferred term is usually lemma or dictionary form; basisform is more common in German-language linguistic literature and some cross-linguistic works. The basis form is not always unique across languages or POS categories, and different lemmatization schemes may choose different canonical forms for multi-form words.

Relation to related concepts: basisform is distinct from stemming, which often outputs an incomplete or non-dictionary form. It is also related to canonical forms and root forms, though definitions may differ by linguistic framework. Examples include mapping went and going to the basis form go, or cameras to camera.

See also: lemma, dictionary form, canonical form, lemmatization, morphology.