The Answerability Protocol: A Minimal Standard for Synthetic Coherence

This paper proposes the Answerability Protocol: a minimal structural standard for testing whether a claim, output, institution, or intelligence-like system remains answerable to reality in the age of synthetic coherence. The paper argues that the central verification problem of the AI era is not merely the circulation of falsehoods, but the cheap production of coherent form. Large language models can generate fluent answers without stable grounding; institutions can produce polished reports, dashboards, compliance artifacts, and values language while exporting cost; public narratives can feel complete while avoiding the conditions that would correct them. The result is a general crisis of signal: coherence, polish, and internal consistency no longer reliably indicate intelligence, legitimacy, or truth. The Answerability Protocol shifts verification from surface adequacy to contact. It asks whether a system can name its claim, ground, falsifier, cost horizon, revision trigger, binding mechanism, and external witness or trace. The paper integrates three layers of the Structural Intelligence corpus: the AI layer, where fluency can exceed grounding; the institutional layer, where standardization can substitute representation for contact; and the protocol layer, where answerability becomes a portable test across AI systems, institutions, policy, and high-stakes reasoning. The core claim is simple: in an age of synthetic coherence, intelligence is not the ability to appear right, but the capacity to be corrected.