CONCEPT
Direct Perception
Gibson's thesis that perception is detection of information already structured in the environment, not construction from impoverished sensory data — a claim that replaces the entire inferential tradition running from Helmholtz to contemporary cognitive science.
The standard account of perception, dominant in Western thought since Helmholtz's 1860s formalization, treated the retinal image as ambiguous and insufficient, requiring the brain to supplement it through unconscious inference, stored knowledge, and computational reconstruction. Gibson rejected every element of this picture. The retinal image is not the stimulus for vision — the ambient optic array is, and the ambient optic array is not impoverished but extraordinarily rich with texture gradients, occlusion patterns, and optic flow that specify the environment with mathematical precision. The organism's task is not to construct a meaningful world from fragmentary data but to detect information that is already present, structured, and available. The implications for understanding the natural language interface are severe: for the first time in computing history, the obstruction between perceiver and problem has been removed, restoring something like direct perception — but with consequences the framework cannot fully settle.
In The You On AI Field Guide
Gibson's rejection of inferential perception