attention (Treisman, 1986, 1988; Treisman & Gelade, 1980). To demonstrate that
attention is necessary to feature integration, researchers often divert or over-
load their subjects’ attention. Under such circumstances, errors in feature com-
binations may occur, known asillusory conjunctions.
Researchers have produced illusory conjunctions by briefly flashing (for
less than one-fifth of a second) three colored letters with digits on both
sides of them.5 XOT 7
The subjects’ task is to report the digits first and then to report all of the
color-letter combinations. On a third of the trials, subjects report seeing
the wrong color-letter combination. For example, they report a yellow X
instead of a blue X or a yellow O. They rarely make the mistake of reporting
any colors or letters that were not present in the display, such as a red X
or a blue Z.
Subjects were also likely to report that they saw a dollar sign ($) in the
briefly flashed display containing S’s and line segments shown in figure
7.15. The same effect was obtained even when the display contained S’s
and triangles. This result demonstrates that the subjects did not combine
the lines of the triangles right away; the lines were floating unattached at
some stage of perceptual processing, and one of the lines could be bor-
rowed by the visual system to form the vertical bar in the dollar sign
(Treisman & Gelade, 1980).Figure 7.14
Search for the conjunction of two colors. (Yellow appears as light gray, red as medium gray, and
blue as dark gray.)
(A) Find the yellow-and-blue item.
(B) Find the yellow house with blue windows.
(A) Search is very inefficient when the conjunction is between the colors of two parts of a target. (B)
However, search is much easier when the conjunction is between the color of the whole item and
the color of one its parts.
156 Philip G. Zimbardo and Richard J. Gerrig