“Sensing” the future: From Kant to the brain as a predictor

The simple act of perceiving, planning an action, predicting others’ intentions, and imagining the events of a day in the future are cognitive processes with something in common, at least according to the Active Inference Hypothesis: They are all based on a model constructed on the basis of our inner structures as well as our past experiences, themselves formed on the basis of our perceptions of internal and external stimuli. As a conceptual tool, that hypothesis can be traced back to Kant’s metaphysical assumption about the way in which we form our perceptions from sensory data and Helmholtz’s view of perception as a process of deducing “causes”—the objects in the outside world—from the “effects” that they have on our sensory organs. In the first part of the paper, we trace a historical epistemological genealogy of this powerful tool up to its appropriation and application to goal-oriented actions by the soviet “physiological cyberneticist” Nikolay Bernstein. In the second part, we show how this tool is employed in contemporary neuroscience. These ties between philosophical roots, theoretical and computational models, and neuroscientific experiments underline the importance of philosophical tools in contemporary neuroscience.