Abstract
According to recent developments in predictive processing and Bayesian cognition, our thinking is fantastic: we grasp the world through predictive, probabilistic models that we compare against the feedback from the actual environment, which really only plays a significant role if it proves our predictions to be wrong and forces us to revise the probabilistic model. Rather than our eyes and brains registering every detail of the environment and configuring it into the larger whole of the percept, perception works the other way around. We already know what we are likely to perceive on the basis of our predictive probabilistic models (or ‘fantasies’) and correct these only if discrepancies with the environment create prediction errors. The predictive processing model of cognition, with its feedback loops of virtual models and prediction errors, has in recent years been extended into many other areas of the cognitive sciences, from perception to motor control, from emotions to our sense of self (see Clark and Hohwy for comprehensive overviews). Not only perception but cognition more generally might turn out to be nothing short of fantastic. The predictive, probabilistic models of perception and cognition which Frith and Friston gloss as “fantasies” are, as we shall see, usually not noticed in everyday life. Tzvetan Todorov’s definition of the literary genre of “the fantastic”, as a genre that makes us hesitate between “truth” and “illusion”, however, suggests that literary texts might often test boundaries and highlight the workings of such cognitive models. As I will go on to show, the fantastic in literature throws “fantastic cognition” into relief.
This is an accepted version of a chapter published in Cognitive Literary Science: Dialogues between Literature and Cognition, eds. Michael Burke and Emily Troscianko. © 2017 Oxford University Press