Abstract
In a two-alternative forced-choice task, pigeons were trained to discriminate whether two successive displays containing four differently colored bars at four different orientations were identical or nonidentical. Different-trial displays were distinguished by three features: color, orientation, and location. That is, the second display involved bars of different colors, different orientations, and different locations on the screen. Pigeons learned this change-no change discrimination to high levels of accuracy. In the first series of tests, only one attribute (color, orientation, or location) was changed in the second display. In the second series of tests, two attributes were changed (color and location, color and orientation, or location and orientation). The results indicated that location was the most salient feature and that orientation was the least salient feature; moreover, changes in two features produced more “different” responses than changes in single features. Finally, we conducted a series of tests where different-trial displays did not contain new values of the three attributes; instead, the values of the attributes were swapped among the objects. The results disclosed no evidence of feature binding.
This research was supported by National Institute of Mental Health Grant MH47313.