Lexemens

Lexemens are minimal units of lexical meaning proposed as building blocks in a semantic theory. A lexemen captures a core sense or semantic feature that can combine with other lexemens to form lexical items, phrases, or expressions. Unlike traditional lexemes, which are tied to a word form in a language, lexemens are abstract, language-neutral primitives realized in each language through its phonology and morphology.

Origin and naming: The term blends lexicon and sememe and is intended as a universal primitive for meaning, separating stable semantic content from surface form.

Structure and properties: Lexemens carry a semantic type (such as motion, agentive role, or perceptual quality) and may encode argument structure. They combine according to simple rules, and a single word can realize multiple lexemens across senses.

Applications: In computational linguistics, lexemens underlie multilingual lexica and semantic parsers, enabling cross-lingual mapping of words to shared semantic primitives and aiding sense induction and translation.

Examples: In a hypothetical lexicon, L_motion encodes motion events (go, run) and L_agency encodes agentive roles. A verb may realize several lexemens depending on context.

Limitations: The abstraction is difficult to validate empirically, and matching real words to a fixed set of lexemens can be data-intensive and prone to overfitting.

See also: Lexeme, Sememe, Semantic frame, Word sense disambiguation, Lexical semantics.