Does the Brain Really 'Decide'? A Neuroscientist Says No

For decades, cognitive neuroscience has treated decision-making as a distinct step between sensing the world and acting on it, with information flowing from perception to thought to action, each stage tied to its own brain process. Tom James, a professor in Indiana University's Department of Psychological and Brain Sciences, argues in a new study published in the Journal of Cognitive Neuroscience that this "sandwich model" does not match what is known about the brain: researchers have identified neural systems for sensing and for movement, but no comparable brain mechanism that serves as a dedicated decision-making stage.

James proposes instead that purposeful-looking behaviour emerges from continuous interaction among sensory, sensorimotor and motor processes, a concept he calls "action selection" rather than decision-making. He points to a simple robot built only from sensory, motor and sensorimotor components: even with no capacity to decide or strategize, the robot displays "wall-following" behaviour that looks intentional and goal-directed. "The robot does not have decisions built into it," James said. "It just senses its environment and moves around accordingly... it looks like the robot is making decisions. And yet, it is not."

James draws on the "physicalist" philosophy associated with Daniel Dennett, under which only physical processes can cause physical events; because a decision is a nonphysical, abstract concept, he argues it cannot itself directly cause an action, much as an object's centre of mass is a mathematical concept, not something that can be moved independently of the object itself.

James also argues that proposing a central brain "controller" that monitors and directs behaviour risks the philosophical problem Dennett called the "Cartesian Theatre": the controller would itself need an inner observer to explain its own workings, and that observer would need another, in an infinite regress.