Lexiconeither

Lexiconeither is a theoretical framework and data model for annotating lexical items whose meanings depend on context and may realize as one of two primary senses. The name combines lexicon with either, signaling a built-in duality within a single entry. It is used in computational linguistics and lexicography to capture how context selects a sense and, in some cases, how two senses relate within discourse.

Structure and annotation: Each entry records the lemma, part of speech, two primary senses (Sense A and Sense B), and contextual cues that favor one sense. A translation layer provides language-specific equivalents for each sense, and a probability or preference score expresses the expected sense given a context. Entries can cover gradient or adjacent senses, not only strict opposites.

Example: For the English word bank, Sense A denotes a financial institution (cue: banking transactions; translation: banque), and Sense B denotes the land beside a river (cue: by the river; translation: berge). A sentence such as The bank approved the loan versus The fisherman sat by the bank would trigger different sense annotations and translations.

Applications and status: Lexiconeither supports word sense disambiguation, machine translation, cross-lingual information retrieval, and lexicography by providing aligned sense pairs and context-sensitive signals. It remains an active area of research with ongoing discussions about standardization and annotation schemes.

See also: Lexicon, Word sense disambiguation, Semantic annotation, Multilingual NLP.