Catalogers and metadata professionals are frequently labeled traditionalists—or occasional "curmudgeons"—for insisting on controlled vocabulary and precise terminology. Yet without intentional lexical discipline, large-scale discovery systems inevitably succumb to noise and entropy.
The Tension Between Precision and Natural Language
As social tagging and broad keyword indexing gained popularity throughout the late 2000s, proponents argued that raw user queries would obsolete structured thesauri. However, natural language search introduces fundamental failure modes:
- Synonymy: Searching for "automobiles" misses articles indexed solely under "cars", "vehicles", or "motor transports".
- Polysemy: Homographs dilute results with irrelevant hits across completely unrelated subject areas.
- Temporal Drift: Technical terms and cultural descriptors evolve over decades, leaving historical documents undiscoverable under modern query syntax.
Structured Vocabularies in the Era of Linked Open Data
The solution was never to abandon controlled vocabularies, but rather to decouple them from rigid legacy catalog cards. By assigning stable dereferenceable URIs to concepts (utilizing SKOS and OWL ontologies), we allow natural language interfaces to map directly onto unambiguous conceptual nodes.