ABSTRACT
Since the time of Gestalt psychology, it has been known that stimulus structure plays an important role in visual object recognition (e.g., Gottschaldt, 1926). Whereas the Gestaltists were largely concerned with the description of phenomena, more recent approaches have endeavored to discover the underlying information processes. For doing so, reaction time (RT) measures have proved to be a useful instrument (Lachman, Lachman, & Butterfield, 1979). It soon became clear that the influence of visual stimulus structure depends on the task to be performed (Ashby & Maddox, 1990; Pomerantz & Pristach, 1989). For instance, in a partwhole detection task, components of a complex figure were revealed to be part of its internal representation only if these components were relevant to the task (Stins & van Leeuwen, 1993). However, task-irrelevant aspects of visual structure are not always ignored, as evidenced in demonstrations of interference by irrelevant stimulus dimensions in Stroop and Garner tasks (Lockhead, 1972; Pomerantz & Pristach, 1989; van Leeuwen & Bakker, 1995). The influence of task-irrelevant aspects of structure is particularly striking with complex visual stimuli, which, by definition, are characterized by multiple stimulus dimensions.
