Segmenting action for computers and humans: possible links to intentional understanding

The ability to identify discrete actions within a continuous stream of motion interests researchers across a range of fields, including cognitive science, engineering and robotics. The present chapter reports on a study investigating adults´ segmentation of human action for a computer and a human. Consistent with observers using different levels of analysis -bottom-up versus top-down-for the two audiences, participants inserted more segments for the computer than for the human.