Contemporary predictive-processing accounts characterize perception as “controlled hallucination,” yet they leave unspecified the mechanistic transformations through which intrinsically generative, high-entropy cortical activity becomes stable, structured, and cognitively usable. This article argues that this gap is structural rather than peripheral, and that perceptual inference presupposes a distinct mechanistic domestication layer operating prior to interpretation and decision-making. The Hallucinatory Domestication Theory (HDT) advances a unifying framework in which spontaneous cortical activity is progressively constrained through three neurophysiological operations: inhibitory stabilization, oscillatory organization, and attractor formation. Together, these operations reduce entropy, impose temporal structure, and stabilize recurrent dynamics into discrete perceptual constructs that are neither raw sensory inputs nor symbolic representations, but intermediate, meaning-ready units. Integrating hierarchical neural coding with a revised functional role for the orbitofrontal cortex (OFC), HDT further proposes that OFC serves as a semantic grounding hub, binding stabilized perceptual constructs with contextual, motivational, and counterfactual significance. Developmental trajectories, electrophysiological signatures, and clinical phenomena are interpreted as reflecting distinct stages and failure modes of the domestication process. The framework yields explicit, falsifiable predictions across electrophysiology, development, psychopathology, and computational modeling. By articulating a mechanistic domestication layer that precedes inference, HDT reframes perception as the progressive stabilization and structuring of intrinsic cortical dynamics into coherent, meaningful states. This article is offered as a target article to invite critical evaluation and interdisciplinary discussion across neuroscience, cognitive science, and theories of meaning.