The idea that 'reality is a controlled hallucination' has been recently popularised by figures such as neuroscientist Anil Seth. But this claim, which purports to be hard, down-to-earth science, is, in fact, bad philosophy. Philosopher and author of The Blind Spot: Why Science Cannot Ignore Human Experience, Evan Thompson, here argues the 'controlled hallucination' hypothesis brings nothing new to the table regarding the problem of consciousness. And there is a circularity problem: if the theory claims that reality is a hallucination, then that theory itself is part of the hallucination.

“Reality is a controlled hallucination” may look like cutting-edge science but actually is debatable philosophy from the nineteenth century. Popularized recently by neuroscientist Anil Seth, it traces back to Hermann von Helmholtz (1821-1894), a brilliant German scientist and philosopher who made fundamental contributions to physiology, physics, mathematics, epistemology, and the philosophy of science. In the twentieth century, AI scientist Max Clowes, psychologist Chris Frith, and neurophysiologists Humberto Maturana and Rodolfo Llínas put forward versions of the idea. Unfortunately, it’s replete with problems. A better viewpoint is that perception discloses and gives us access to how things are in the environment tailored to our bodily agency.

Today’s version of the reality-as-controlled-hallucination idea comes from the “predictive processing” theory of perception. Suppose you wake up in the night and it’s pitch black in your room. (This example comes from Alva Noë.) You can’t see anything and there’s a power outage. As you’re moving around, looking for a flashlight, you stub your toe on a piece of furniture. Your mental map of where things are isn’t quite right, so you update it and try to carry on without bumping into things again. You predicted the layout to be a certain way but were forced to revise your map, and now you try to minimize the likelihood of making another mistake. That cycle of prediction, updating, and acting, while trying to minimize prediction error, is how perception in general works, according to the predictive processing theory.

Of course, we hardly ever experience perception as being like this. But your brain is always going through that kind of cycle, according to the theory. Your brain actively meets incoming sensory signals on its own terms based on its having a pre-existing model of the outside world. It constantly makes predictions about the sensory signals it will receive, and works to minimize the difference between predicted and actual sensory information. To minimize prediction error your brain revises its internal model and generates motor actions that change sensory stimuli to bring them closer to what it predicts. Sensory signals function as error signals in relation to your brain’s own self-generated predictions.

___

Thinking that perception happens inside the brain is like thinking that flying happens inside the bird’s wings or that dancing happens inside the dancer’s nervous and musculoskeletal systems. This is a category error.

___

If we stopped here, we’d have a general framework in which to develop testable dynamical-statistical models of the brain’s role in perception. The problems arrive when our perceptual experience or consciousness is said to be nothing other than the content of our brain’s predictive model. This is to identify perceptual content, which resides at the level of the whole animal or person standing in relation to the world, with neurophysiological transactions, which take place inside the brain.