Hierarchical Bits — an investigation: from a representation paradigm to a property (FCIR), and its provisional conclusion

Hierarchical Bits (BH) began as a strong hypothesis — that it introduced a new, general paradigm of representation — and is reported here at the honest size the evidence supports. The investigation tested the hypothesis across nine measured angles and a 20-domain survey. Result: the universal-paradigm claim was NOT confirmed — a shared immutable substrate plus selective reading is already mature state of the art across most domains (DICOM, COG/STAC, lakeFS, CRAM/tabix, S-LoRA, MAM). What survived is a narrower, sharper property: keeping multiple — possibly contradictory — interpretations as first-class, co-registered, persistent entities over one substrate, with adjudication deferred and optional. We give it a working name, the First-Class Interpretation Representation (FCIR), formalized as the decoupling of coexistence from adjudication. Honestly: FCIR is not a new mechanism — RDF named graphs (with provenance) and standoff annotation already implement it within their domains; the contribution is a cross-domain name, a falsifiable test, and an algebra (a specification, not a verified theory). This version contains the measured study, the principle, the formal algebra, and a provisional conclusion with its limitations stated. The work stands as an investigation and as method as much as any single finding.