In this view, if you built a robot whose brain worked like the human brain, and then asked computer scientists to describe its workings, they’d say that its brain consists of lots of partly overlapping modules, and modules within modules, and the robot’s circumstances determine which modules are, for the moment, running the show.
