specifitet

Specifitet is a theoretical quality used in linguistics, information theory, and cognitive science to describe the degree to which a description or dataset uniquely identifies a referent within a given context. In practice, higher specifitet means the statement reduces ambiguity more effectively.

Specifitet is conceptually related to specificity but focuses on the identification power of a description rather than its content breadth. The term is a coinage of recent scholarship and is used mainly in discussions of disambiguation, precision in data annotation, and query formulation.

Measurement: One formalization defines specifitet as the log-ratio of possible referents before and after applying the description: S = log2(N_before / N_after). When normalized to a 0–1 scale, 0 indicates no reduction in ambiguity and 1 indicates a single unambiguous referent in context. In practice, N_before depends on domain complexity and context constraints; N_after is the subset consistent with the description.

Applications: Specifitet is relevant to natural language processing, knowledge graphs, search systems, cataloging, and user interfaces, in any setting where precise disambiguation matters.

Limitations: The concept depends on context and background knowledge, is not yet standardized, and can be subjective. Its measurement requires explicit definitions of the contextual space and the referents considered, which may vary across domains. See also specificity and disambiguation.