ABSTRACT
Attention is related to the way the cognitive system processes information (→INFORMATION). A two-stage processing view, in which an initial parallelprocessing stage is followed by a second, sequential stage was first proposed by Donald Broadbent in 1958. Later, in the 1970s, under the impetus of researchers such as Michael Posner, Richard Shiffrin, and Walter Schneider, this view gave way to a new one based on the distinction between two types of processes: automatic processes, which are rapid parallel processes that do not require attention (→AUTOMATISM), and controlled processes, which are slow, serial, and strategically determined (→CONTROL). Anne Treisman’s model of feature integration exemplifies this approach. Target detection is said to be automatic when it is defined by a single elementary property, on a single dimension such as color, brightness, size, or shape that can be analyzed by prewired, specialized detectors functioning in parallel (the target “pops out” no matter how many distractors are simultaneously present). In contrast, when the target is defined by a combination of two elementary properties belonging to two different dimensions and for which there is no dedicated analyzer, detection is controlled. Acting as a sort of “glue,” attention temporarily holds the combination of expected properties together in a coherent representation of the object.
