Ontology alignment problem

An ontology alignment problem arises when two systems that need to communicate possess different conceptual frameworks — different categories, different relations, different assumptions about what exists and how it is structured. The problem is not merely technical (finding a translation) but epistemological: the two systems may carve up the world differently, and there may be no neutral ground from which to adjudicate the difference. Attempts to solve alignment through universal top-level ontologies — the dream of the Semantic Web — have repeatedly failed because they assume that conceptual divergence is a bug to be eliminated rather than a structural feature of distributed cognition. The more productive approach treats alignment as an emergent coordination problem: agents negotiate partial, context-specific mappings through interaction, rather than adopting pre-shared formal schemas. The FIPA-ACL standard assumed ontology alignment was a solved problem; its failure in open systems demonstrates that alignment cannot be prescribed in advance of use.

The ontology alignment problem is not a puzzle waiting for a clever algorithm. It is the permanent condition of any system in which multiple agents construct meaning independently. The goal is not perfect alignment but functional coordination — enough shared structure for the task at hand, without requiring universal agreement.