A growing number of neuroscientists claim that “reality is a controlled hallucination,” a projection of our brains. But in a recent article for the IAI, the philosopher Evan Thompson argued that this is nonsense: if brains hallucinate reality, then the brain itself must be part of the hallucination, so the theory falls apart. Now, Oxford philosopher Jan Westerhoff, author of Reality: A Very Short Introduction, replies to Thompson, arguing that his objection misses the radical core of the neuroscientists’ claim. The idea is coherent, he says, but only if we embrace its implication that our theories aren’t descriptions of the world, but rather provide models for intervening in it.

Some contemporary neuroscientists describe the brain as a “fantastic organ,” able to create the fantasy—or hallucination—of the world around us, a world that appears to impress itself on our senses, with photons bouncing off objects and hitting our retinas, sound-waves from the radio entering our ears, and olfactory molecules from food being registered by our noses, but which really is only something the brain has constructed. This brain-based construction, the story goes, is not entirely unmoored from extra-cranial reality: there are the sensory data against which the brain’s fantasies are tested, thereby constraining the hallucination within the boundaries of the real. The most successful contemporary incarnation of this approach is predictive processing theory: the “hallucinations” are the brain’s best guesses for explaining the perceptual data, which are tested and adapted as more data come in.

Critics were quick to point out a curious circularity that seems to follow from this form of neuroconstructivism: if the world is the model produced by the brain, and the brain is part of the world, then the brain is a part of the model that generates itself. This appears to lead to an absurd situation, “like a reflection generating its object or a film producing the projector.” Less metaphorically, this picture of the relation between the brain and the world seems to undermine itself both ontologically and epistemologically, leading to problems with respect to existence and with respect to knowledge.

SUGGESTED READING Reality is not a controlled hallucination By Evan Thompson

If the brain produces the world-model, the brain, as a cause, seems to have to be there first, before the effect, the world-model, is produced. But the brain is part of the world-model, so how can it exist before having produced itself? Moreover, we study the brain and develop a specific account of how it creates the fantasy of the world around us. But once we have accepted this, all our knowledge of the brain also turns out to be part of this fantasy, so how can we trust that our initial claims about the brain and its model-creating capacity actually constitute knowledge?

How might one respond to this criticism? An immediate possibility is to give up brain-based constructivism, arguing that because it lands us in these inextricable puzzles about reality and knowledge, we should let go of the idea that the world we live in is just a construction of our brains. Reality is not a controlled hallucination. Computational models of cognition like the ones offered by predictive processing theory are idealized models, but cannot be literally true as theories of the mind.

___

The solution is to let go of the realist assumptions. We know the tables and chairs in-the-model, but not the tables and chairs themselves, and we also do not know the brain itself, but merely its representation.

___