nonlemma

Nonlemma is a linguistic term used to describe any surface word form that is not the lemma of its lexeme. The lemma, or dictionary form, is the canonical form under which the word's related inflected or derived forms are grouped. Nonlemmatized forms include inflections produced by grammar (for example, run, runs, ran, running) and irregular or derived variants that do not constitute separate lemma entries in all dictionaries (goose vs geese; child vs children). In many languages with rich morphology, a single lemma can yield dozens or hundreds of nonlemma forms.

In lexicography and natural language processing, distinguishing lemmas from nonlemmas supports tasks such as search, indexing, and part-of-speech tagging. Lemmatization or morphological analysis attempts to map nonlemmas to their lemmas, yielding a normalized representation. This helps reduce data sparsity and connect inflected forms to a common lemma across documents or corpora.

The term nonlemma is not uniformly standardized; some researchers prefer “inflected form,” “surface form,” or simply “non-canonical form.” Nevertheless, nonlemma serves as a concise label for references to non-headword forms in discussions of morphology, lexicon design, and linguistic annotation.