Abstract
The visual world extends around us, but due to acuity limitations we can only process to a high degree the objects or regions that we are currently fixating. Accordingly, if an object of interest falls outside our fovea, we must move our eyes over to it for further processing. According to the saccade target object theory (Currie et al, 2000; McConkie & Currie, 1996), prior to moving our eyes to a target object, we select critical locating features of the target that will be used at the start of the subsequent fixation to verify that the eyes have landed correctly on the intended target. However, what constitutes the set of critical locating features is undetermined. Previous work has demonstrated the retention of various features (e.g., orientation, Henderson & Siefert, 1999; Pollatsek et al., 1984; color, identity and location, Henderson et al, 1989; Irwin & Andrews, 1996). The implicit assumption on the basis of this evidence is that these features are therefore always retained. However, given that attention precedes the eyes to the saccade target (e.g., Deubel & Schneider, 1996; Hoffman & Subramaniam, 1995), and given that attentional allocation is flexible (Hawkins, Hillyard, Luck, Mouloua, Downing, & Woodward, 1990; Posner, 1980; Wolfe, 1994), the purpose of the present experiments is to demonstrate that the encoding of one feature of the saccade target (color) is flexible. In Experiments 1 and 2, we demonstrate that subjects do not retain the color of the saccade target when they can select the target on the basis of location. In contrast, in Experiment 3, when color is used to define the location of the saccade target, subjects retain color information, even after it is no longer relevant for the task. Thus, features of the saccade target (e.g., location and color) are differentially salient, consistent with a hierarchical encoding (Jiang, Olson, & Chun, in press), with salience in this case determined by the goals of the task.