Papers

This paper proposes absorption as a predictive constraint that modulates when experience functions as evidence capable of updating higher-level predictive models. Absorption is defined as a configuration of attention in which competing predictions are attenuated, interpretive alternatives are temporarily suppressed, and learning signals are integrated with greater relative weight. Framed as a constraint rather than a distinct state or novel computational mechanism, absorption alters the inferential context in which existing predictive processes operate. We argue that this extension accounts for variability in learning and change outcomes by specifying a gating condition for durable predictive updating.

This paper introduces the Unified Absorption Model (UAM), a mechanistic framework proposing that absorbed experience functions as a central enabling condition under which predictive systems update. Within this model, absorption is defined as a configuration of attention in which competing predictions are attenuated, learning signals are amplified, and identity-relevant models become temporarily more plastic. UAM distinguishes absorbability as a property influencing the rate of system updating rather than the depth or ultimate potential of change. The model offers an account of why moderate absorption repeated over time can produce durable change, why insight and effort alone often fail to update entrenched patterns, and why intensity without absorption yields inconsistent outcomes.