A situation in bibliographic or administrative data where two or more distinct persons or entities share identical or indistinguishable name strings, producing ambiguity, conflation, or misattribution in records and discovery systems.

Principle

A name string alone is insufficient to disambiguate agents; reliable attribution requires supplemental disambiguating attributes (e.g., birth/death dates, affiliation, persistent identifiers) or authority control to map identical names to distinct identities.

Demonstration

Illustrative scenario — Situation: Two different authors publish under “A. Garcia.” Recognition: citation counts and search results mix works by both individuals. Action: apply authority control by adding ORCID/ISNI to each author record, separate bibliographic entries, and update discovery facets. Consequence: corrected attributions, improved discoverability, and accurate metrics for each author.

Misapplication

Assuming a middle initial or affiliation at a single point in time suffices for permanent disambiguation. That error ignores name changes, shared initials, and evolving affiliations that can cause future conflation.

Consequence

Name collisions cause misattribution of works, incorrect citation metrics, patron confusion, legal or reputational errors, and inefficient provenance tracking; these consequences arise because systems resolve identity on inadequate identifiers (name strings) rather than persistent identity metadata.

Reversal

In contexts where the shared name intentionally denotes a single legal or corporate entity (e.g., corporate authorship) or where a curatorial policy groups works under a collective name, identical name strings are a correct aggregation rather than a collision; the qualification depends on the intended entity of attribution.

Boundary

Clearly within: personal name fields in author/patron records that lack persistent identifiers or distinguishing dates. Boundary case: pseudonyms, transliterations, or culture-specific naming conventions that require specialist authority control. Clearly outside: identical titles or subject headings that are not agent names.

Semantic Tension

Disambiguation ↔ Privacy — collecting richer identifiers (dates, affiliations, identifiers) improves disambiguation but may impinge on privacy or data‑protection constraints; systems must balance identification needs against legal and ethical limits.

Synthesis

Name collision shows that strings are not identities: robust disambiguation requires persistent, managed identity metadata and governance (authority files, identifiers, provenance) rather than relying on surface name forms.