Safety–critical autonomous systems require causal reasoning mechanisms that remain au-ditable, deterministic, and traceable under degraded sensing, adversarial disturbances, andregime transitions. Conventional causal frameworks are typically implemented using proba-bilistic graphical models whose validity depends on distributional assumptions and stochasticindependence.This paper introduces a deterministic causal architecture in which causal structure emergesfrom trust–controlled structural dynamics rather than probabilistic dependency. The architec-ture integrates deterministic structural growth, trust–gated interaction rules, and causal graphconstruction to produce a locally finite directed acyclic graph with explicit causal provenance.Within this framework, causal connectivity becomes a function of trust dynamics, induc-ing a trust-controlled causal topology (TCCT) over the event history of the system. We showthat deterministic trust thresholds induce sharp structural regime dynamics (SRD) — abrupttransitions in causal connectivity — and illustrate that these transitions serve as deterministicindicators of system-level behavioral changes.This work focuses on the architectural and structural foundations of deterministic causaltopology, providing theoretical analysis and computational illustrations rather than evaluationon a specific operational dataset.The resulting architecture provides explicit causal traceability, deterministic replayability,and bounded structural behavior, properties that align naturally with certification-oriented ob-jectives in safety–critical autonomy systems. Deterministic trust dynamics produce measurablestructural regime transitions in causal graph connectivity. Reproducibility: All empiricalsimulations are reproducible using the dsfb-srd reference implementation available on GitHub.
Riaan De Beer (Thu,) studied this question.