Active Ignorance: Self-Evidencing, Confirmation Bias, and Model Performativity in Predictive Processing

Event DateEvent Date

Event LocationLocation

Off Campus

Part of the Center for Discursive Inquiry's 'At the Conjuncture: Art and the Imagination'

Active Ignorance: Self-Evidencing, Confirmation Bias, and Model Performativity in Predictive Processing, as part of the strand At the Conjuncture: Art and the Imagination.
The structure of our sessions this semester will be focussed on dialogues and conversations where our guests will present for a few minutes each so as to encourage conversation amongst all attendees.
 
All are welcome!
 
Saturday, Oct. 12, 2024
11:30 am–2 pm PST


Predictive processing (PP) and active inference constitute a novel embodied computational theory of mind that unites perception, cognition, and action under a single explanation: the reduction of uncertainty through acting on the world and constantly updating a generative model of the causal etiology of sensory change. However, most accounts construe such uncertainty reducing activity to naturally tend toward increasing knowledge, or as optimizing the fit between the generative model and the world. This presentation introduces the concept of active ignorance to elucidate how unwarranted epistemic entitlements are pervasive in thought and follow directly from the premises of PP. Drawing on agnotology, we argue that the tacit Enlightenment conception of mind as progressively banishing ignorance is inadequate. Active ignorance is characterized as an ‘agno-inferential’ strategy whereby epistemic communities ignore, deny, or alter evidence that conflicts with their preferences, potentially reinforcing maladaptive systemic affordances which sustain social cognition as a whole. In the midst of the massive misinformation and the rise of fascism we consider the concept of active ignorance to be an important intervention into the nascent image of thought proposed by PP.