[If] we can experimentally isolate a process and show that it can still work well, then it cannot require complex interaction with other parts of vision and can therefore be understood relatively well on its own.

Vision, 1982