AbstractTrauma exerts persistent regulatory effects through distortion of precision-weightedhierarchical inference. Within the predictive processing framework, psychological injury isunderstood not primarily as a memory disorder or emotional wound in any colloquial sense, butas a systematic miscalibration of precision — the confidence weighting assigned to predictionerror signals at multiple levels of a generative model. This paper formalizes trauma as ahigh-level prior violation that corrupts authority-installed semantic tokens, producinggain-amplified trigger cascades, recursive rumination loops, and characteristic disruptions tosleep architecture and interoceptive coherence. Semantic tokens — words, names, relationalcategories — are modeled as high-level latent variables whose precision weights are installedduring developmentally sensitive periods through authority relationships. Violation of thoseauthority-granted expectations generates catastrophic prediction error that propagatesdownward through the generative hierarchy, locking affected priors into high-precision,low-flexibility attractor states. Subsequent exposure to contextually similar stimuli producesrapid attractor capture, experienced phenomenologically as involuntary reactivity. Ruminationis formalized as a failed error-resolution policy: a recursive loop in which narrativereconstruction, counterfactual simulation, and meta-monitoring generate sustained internalprediction error without convergent resolution, maintaining high sympathetic arousal andsuppressing the precision recalibration that occurs during consolidated sleep. Sleep ispositioned as a core regulatory mechanism — a nightly downscaling and reweighting cyclewhose disruption sustains and amplifies the pathological attractor landscape. Dissociation ismodeled as emergency hierarchical decoupling: a temporary reduction in interoceptiveprecision weighting that preserves local functional coherence at the cost of global integration.Depression is characterised by rigidly pessimistic high-level priors that constrain policyexploration and reinforce rumination dominance. Psychosis risk is carefully framed asprecision instability under sustained allostatic load — a regime transition in which salienceoverflow and boundary dissolution represent quantitative perturbations of the metastableconscious regime rather than categorical breaks. Boundary repair and therapeutic change areformalised as precision recalibration and source reattribution: the correction of misassignedauthority precision weights, the restoration of accurate self-other boundary modelling, and the gradual semantic reweighting of high-gain tokens. The paper argues that mechanistic,non-metaphysical accounts of these dynamics are sufficient and preferable to archetypal ormoral framings, which risk reifying maladaptive priors. Implications for measurement,early-warning detection, and non-prescriptive therapeutic design principles are outlined. Keywords: predictive processing; precision weighting; trauma; attractor dynamics;rumination; dissociation; regulatory coherence; source monitoring; sleep reconsolidation;metastable consciousness; active inference; boundary repair
Smith et al. (Wed,) studied this question.