During this talk I will present some ongoing and unfinished
research. I will briefly investigate the notion of semantic autonomy,
i.e. lack of centralized semantics, but presence of a type of semantic
coherence (formalised as a type of consistency). The analysis will
focus on rather inexpressive ontologies (similar to taxonomies) and
investigate the case where they and the mappings between them are
The motivating practical example that could benefit from this theoretical framework is Distributed Knowledge Management (DKM). I will briefly mention the organisational benefits of a DKM system that has semantic autonomy. Selected related research will be mentioned, and some of it originates from the field of the Semantic Web.
A layered framework will be presented, that includes epistemological assumptions and a logical formalization and a system that integrates very simple reasoning with interaction. Finally, the question will be asked if a system built using this framework scales, and I will touch the issue of preliminary experimental results.