Un-debunking ordinary objects with the help of predictive processing

Abstract

Debunking arguments aim to undermine common sense beliefs by showing that they are not explanatorily or causally linked to the entities they are purportedly about. Rarely are facts about the etiology of common sense beliefs invoked for the opposite aim, that is, to support the reality of entities that furnish our manifest image of the world. Here I undertake this sort of un-debunking project. My focus is on the metaphysics of ordinary physical objects. I use the view of perception as approximate Bayesian inference to show how representations of ordinary objects can be extracted from sensory input in a rational and truth-tracking manner. Drawing an analogy between perception construed as Bayesian hypothesis testing and scientific inquiry, I sketch out how some of the intuitions that traditionally inspired arguments for scientific realism also find application with regards to proverbial tables and chairs.