Every vocabulary contains decisions that can appear neutral because they are expressed through technical forms. Someone chooses the preferred name. Someone defines its scope. Someone decides whether two concepts are equivalent, which version is current, which relation is legitimate and when an older formulation should be retired. These operations shape what can be searched, compared, counted and reused. Semantic organization therefore always distributes authority. A centralized ontology can make a domain highly coherent. Shared definitions and controlled relations are indispensable where precise coordination matters. But one semantic centre can also convert a situated interpretation into an apparently universal order. This is especially problematic in interdisciplinary fields, historical vocabularies and community knowledge, where several legitimate definitions may coexist. Complete decentralization creates the opposite difficulty: every actor can define independently, but the resulting fragmentation makes shared use increasingly difficult. Between these extremes lies a polycentric possibility. Different institutions, communities or authors can maintain distinct but connected conceptual records while sharing minimum rules for identity, provenance, versioning and mapping. Coherence then emerges through inspectable relations rather than compulsory sameness. Two definitions can coexist without pretending to be identical. A concept can branch into another line of development while preserving its genealogy. A translation can expose its losses instead of hiding them behind an equivalence sign. This approach does not remove authority. It makes authority attributable and revisable. We can see who maintains a definition, within which domain, under what mandate and with what history of decisions. Ostrom’s work on polycentric governance is useful here because it shows that coordination does not require a single sovereign centre. Bowker and Star similarly demonstrate that classifications are infrastructures with consequences, not neutral descriptions of a pre-given world. Semantic governance should therefore resemble a negotiated system of coordinates more than one final map. Shared standards remain necessary, but they can operate as minimum protocols rather than total ontologies. The challenge is to create enough common structure for exchange while preserving the differences through which concepts remain historically and culturally meaningful. A well-governed semantic system does not eliminate disagreement; it makes the location, authorship and consequences of disagreement visible.
Bowker, G. C. and Star, S. L. (1999) Sorting Things Out. MIT Press.
Foucault, M. (1969) L’archéologie du savoir.
Gruber, T. R. (1993) ‘A Translation Approach to Portable Ontology Specifications’, Knowledge Acquisition, 5(2), 199–220.
Ostrom, E. (1990) Governing the Commons.
Ostrom, E. (2010) ‘Beyond Markets and States’, American Economic Review, 100(3), 641–672.