Abstract
The work of Alfred Yarbus (1967) is often cited in discussions of ‘bottom-up’ vs. ‘top-down’ oculomotor control strategies. Many authors use Yarbus' experiment in which subjects viewed Repin's painting of an “unexpected visitor” returning home after a long absence to demonstrate the task-dependence of eye movement patterns. Yarbus reported that subjects' eye movement records varied dramatically with instruction, depending, for example, on whether subjects were simply examining the painting or were preparing to answer a question such as “surmise what the family had been doing before the arrival of the ‘unexpected visitor.’” Yarbus eyetracking apparatus was primitive by today's standards, but more important were the conditions under which the experiments were performed. Extraordinarily long (e.g., three minute) viewing times, severely restricted movements, and awkward (and likely painful) optical stalks attached to the sclera via suction surely affected the oft-reported results. While there is no doubt that eye movement patterns are influenced by high-level tasks, it is not clear to what degree Yarbus' results reflect the peculiar circumstances of his study. Because of the role of Yarbus' work in the literature, it is important to determine the extent to which the results reflect oculomotor behavior under more natural conditions. Twenty subjects' gaze patterns were monitored with a head-free eyetracker as they viewed Repin's painting (and other images) on a large-field display. As with Yarbus' experiment, subjects either free-viewed the images or were asked one of a number of questions about each image. The presentations were self-paced; viewing times were typically an order of magnitude shorter than the times Yarbus imposed. While subjects' eye-movement patterns were clearly task dependent, the patterns are much less dramatic than those shown in Yarbus' now classic illustrations.