The Reality Totem: Human Epistemic Anchoring and Hybrid Vigilance in Human-AI Systems

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This conceptual preprint proposes the Reality Totem as a human epistemic anchor formed through repeated, corrected exposure to an observable domain. It distinguishes the human anchor from its possible instrumentation and from hybrid human-AI vigilance. The framework emphasizes representational discordance, symmetric doubt, relative independence of the monitoring channel, and audit-trigger status rather than truth-verdict status. It includes two paired thought experiments and a falsifiable C0-C4 validation program. This version corresponds scientifically to the manuscript submitted to Minds and Machines on 20 September 2026. The preprint has not been peer reviewed.