Looked But Failed to See

Runway Collisions, Invisible Gorillas, and the Illusion of a Complete Visual Field

At NASA Ames Research Center in California in 1980, aerospace psychologists Richard Haines and Edith Fischer placed active commercial airline captains inside a full-motion Boeing 727 flight simulator to evaluate a breakthrough in cockpit instrumentation. Engineers had developed an early windshield Head-Up Display that projected glowing green symbols for airspeed, altitude, and glide-slope alignment directly onto the forward glass. Because the symbols were collimated at optical infinity, pilots could fly precision instrument approaches through thick clouds without ever looking down at their instrument panel. Every aviation authority expected the glass projection to eliminate runway accidents by keeping the pilot's eyes aimed straight down the flight path during the critical final seconds before touchdown.

After several routine landings through simulated fog where the runway appeared clear at two hundred feet above the ground, Haines and Fischer introduced an unannounced emergency on a test approach. As the simulated Boeing 727 broke out of the cloud ceiling two hundred feet above the threshold, a massive wide-body jetliner sat parked squarely across the middle of the active runway waiting for takeoff. Two of the veteran airline captains steered their simulated airliners straight through the center of the parked jet without touching the go-around throttles. When Haines halted the simulator and showed the captains video replays of the wide-body fuselage they had stared directly through for several seconds, the pilots were stunned. Their eyes had been wide open and aimed dead ahead through the windshield. Locking their cognitive focus onto the glowing green numbers in the near focal plane had erased the physical airplane behind them.

Throughout the 1990s, psychologists Arien Mack at the New School for Social Research and Irvin Rock at the University of California, Berkeley, brought this cockpit failure into the psychophysics laboratory across thousands of controlled trials. They instructed volunteers to fixate on the center of a small cross flashed briefly on a screen and report whether its horizontal or vertical arm was longer. On the fourth trial, while the observer stared straight at the intersection, an unexpected word or geometric shape appeared directly at the central fixation point. Nearly seventy-five percent of observers reported seeing nothing other than the cross. Calling the phenomenon inattentional blindness, Mack and Rock established that retinal stimulation and healthy visual pathways cannot produce conscious sight unless executive attention is actively allocated to the object.

In 1999, Daniel Simons and Christopher Chabris at Harvard University tested whether inattentional blindness could erase a live, moving figure across an extended real-world scene rather than a millisecond screen flash. They filmed two teams of three undergraduate students—one dressed in white shirts, the other in black shirts—weaving in a circle outside an elevator lobby while tossing orange basketballs to one another. Volunteers watching the sixty-second film were given a demanding selective-attention task: silently count the exact number of bounce passes and aerial passes completed by the white-shirted team while ignoring the black-shirted players.

Halfway through the clip, a student wearing a full-body dark brown gorilla suit walked casually into the middle of the passing circle, turned to face the camera, thumped her chest with both fists, and walked off the opposite side of the frame after nine full seconds on screen. When Simons and Chabris polled their observers afterward, nearly half of all neurotypical participants reported an accurate pass count while insisting with complete confidence that no gorilla had ever entered the hallway. By contrast, when observers were asked to count passes made by the black-shirted team instead, detection of the dark-furred gorilla jumped past eighty percent. Their top-down attentional template was now tuned to admit dark moving shapes while suppressing white ones. When sports scientist Daniel Memmert later repeated the basketball-pass study using infrared corneal eye trackers, the observers who swore no gorilla had entered the room had fixated their foveas directly on the gorilla's chest for an average of one full second. That gaze duration matched the exact fixation count of those who spotted her immediately.