The Predictive Processing premise is false: The brain does not predict, it compares.

Predictive Processing (PP) holds that the brain continuously predicts future sensory states to organize perception. This paper argues that premise is false due to logical and physical impossibilities: predicting a system that includes oneself creates a self-referential paradox, and both quantum uncertainty and classical chaos prevent accurate prediction at any timescale. The paper distinguishes prediction as a high-level cognitive function from its proposed role as a low-level operating mechanism, arguing PP mistakenly universalizes it. An alternative is proposed: a comparison mechanism that continuously compares current sensory input with the immediately prior state stored in short-term memory, generating a result that drives the stream of consciousness and triggers exogenous attention when change exceeds an adaptive threshold.

Study at a glance

Abstract

AbstractPredictive Processing (PP) is a dominant framework in contemporary cognitive science for explaining perception and attention. This paper argues that a foundational premise of PP — that the brain continuously predicts future states to organize perception — is false, due to both logical and physical impossibilities.Logically, accurately predicting the future state of a system that includes oneself leads to an inescapable self-referential paradox: the act of prediction itself changes the system being predicted. Physically, both quantum uncertainty and classical chaos (the butterfly effect) impose fundamental limits on any claim of accurate prediction, regardless of the timescale involved.Furthermore, this paper distinguishes between prediction as a high-level cognitive function (e.g., planning, anticipation, reasoning) and prediction as a proposed low-level operating mechanism. While the former exists and is useful, PP erroneously elevates it to a universal, moment-to-moment principle — an error this paper corrects.Based on this critique, this paper proposes an alternative low-level operational framework: a comparison mechanism. This mechanism does not rely on any prediction of future states. Instead, it is based on the real-time comparison between past and present states. Specifically, the brain continuously compares the current sensory input with the immediately prior sensory state stored in short-term memory, generating a comparison result (e.g., same, different, or a specific magnitude of change). This comparison result simultaneously serves two functions: (1) it drives the stream of consciousness from one moment to the next; and (2) when the comparison result exceeds an adaptive threshold, it triggers exogenous attention.The comparison mechanism is superior to Predictive Processing in logical coherence, physical feasibility, and computational efficiency. This paper argues that this mechanism should replace prediction-based frameworks as the foundational operational principle for explaining the dynamics of consciousness and exogenous attention.Keywords: Predictive Processing; Comparison mechanism; Stream of consciousness; Exogenous attention; Self-reference; Chaos theory