Predictive Processing: an introduction to the unifying proposal of human cognition

Abstract

This article aims to provide a critical, comprehensive, and previously unprecedented, Portuguese presentation of Predictive Processing (PP) – a theoretical framework for comprehending cognition that proposes an inversion of our standard understanding of action, perception, sensation, and their relation. Here, our primary objective is to introduce the main concepts and ideas behind PP, treating it as a moderately embodied model of cognition and analysing its credentials as a unifying theoretical proposal. In order to do so, we will start from a historical contextualization of some currents of thought that might have fostered its initial development, then we will offer a non-mathematical description of the Free Energy Principle, which underlies and substantiates the activity of its specificities, and finally clarify the role that, according to PP, Bayesian inference, prediction error minimization and the so-called Active Inference have in the homeostatic maintenance of our predictive brains and bodies. As a conclusion, we will provide a synthesis of some of the consequences PP could bring to our current understanding of the human brain and behaviour, alleging that, although its description of cognition as a single and continuous predictive process has the potential to eventually unify different explanatory paradigms and levels of analysis, for now, perhaps it is better to think of it in more modest terms, as a tool or heuristic to help us rethink many of those topics that are central to the scientific and philosophical study of the mind.