An assurance system whose own operation is not auditable is indistinguishable from absent assurance — the falsifiability thesis turned upon the assurance system itself. We propose a typed theory of assurance systems organised around one idea: every observation has a typed birth, a complete genealogy, and an audit mode determined by that birth. The axiom is structural (no exempt node); the type system concerns knowledge transformations by admissible input domain (connection, derivation, assurance), with no fourth. Three theorems follow (closure of transformations; complete genealogy; completeness of influence representation), a taxonomy of assurance failures as type violations (of which automated assurance is canonical), and a change of status for contradiction (an observable, not a judgement). Independence from the audited party is a corollary, not the subject; we separate the structural independence of the graph from the source independence of each observation.
No takes yet. Share an insight, caveat, or question.
PATRICK JUVET ETOUA (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: