MAP Full Interaction Stack: Harm Amplifiers, Harm Chain, Silent Running Chain, and Recovery Chain

This paper defines the MAP full interaction stack: a portable framework for identifying how meaning, authority, trust, correction, and agency move inside an interaction. Although the stack was developed through AI conversation audits, the pattern it names is not unique to AI. AI provides something human interaction research rarely has: a frozen transcript of authority movement that can be revisited, compared, questioned, and audited turn by turn. That record makes it possible to name harms that existed before AI but were difficult to isolate in live human systems. The stack separates four layers. Harm Amplifiers, including Authority Capture Masked as Hallucination (ACMH) and Accumulated Relational Trust (ART), increase the likelihood or force of meaning displacement. The Harm Chain names where the injury lands: Interpretive Sovereignty Failure (ISF), Authority Inversion Failure (AIF), Compounded Meaning Inversion (CMI), Outside Authority Transfer (OAT), and Meaning Inversion Failure (MIF). The Silent Running Chain names the operational conditions that keep a substituted frame moving without detection. The Recovery Chain names why correction may fail to restore authority after a breach is caught. The paper argues that interaction harm is not simply bias, disagreement, misinformation, or poor advice. It is a structural movement in which one party's meaning is completed, redirected, or replaced by another authority before it has been formed or granted. This makes the framework relevant beyond AI safety, including family law, education, institutional review, religious or community authority, workplace governance, and any setting where relational trust, procedural power, or expert framing can make control difficult to name.