leset

Leset is a term used in linguistics and computational linguistics to denote a curated subset of a language's lexicon selected for a particular analytical purpose or domain. The concept is not uniformly defined and appears most often as a descriptive label in niche scholarly discussions rather than as a formal theory. The etymology is uncertain; the word is sometimes described as a blend of lexicon and set, and in some texts it appears alongside variants such as lexset or semantic set.

A leset is typically defined by explicit selection criteria, such as semantic field (for example emotion terms), part of speech, frequency thresholds, or task-specific constraints. It serves to focus analysis on a manageable portion of the lexicon and to study properties such as lexical similarity, polysemy, or acquisition patterns within a controlled subset. Lesets are used in corpus linguistics, lexical semantics, and natural language processing to build or test resources like glossaries, semantic nets, or domain-specific lexica.

Examples include a leset of color terms (red, blue, green, yellow) or a leset of emotion terms (happy, sad, angry, fearful). Researchers also compare lesets across languages to explore cross-linguistic variation in domain coverage. The lack of standard criteria for constructing lesets means findings can be sensitive to how the set is formed, limiting cross-study comparability.

See also: Lexical field, lexical set, semantic field, lexical semantics, corpus linguistics, natural language processing.