Abstract: Perceptual global precedence referred to a priority in perceptual processing for the global features of a form or object. This phenomenon was firstly reported by Navon (1977) with compound stimuli (larger letters were constructed from smaller letters, e.g., a large H made up of small Ss). Navon required observers to identify either the large (i.e., global) or the small (i.e., local) letters and found that global letters were identified more rapidly than local letters and also that irrelevant global letters were more difficult to ignore than irrelevant local letters. Perceptual global precedence had been widely replicated within certain boundary conditions (Dalrymple et al., 2007; Roalf et al, 2006; Schatz & Erlandson, 2003; Han Shihui & Chen Lin, 1996; Kimchi, 1992). However, it was still unclear that whether global precedence could exist in other higher cognitive processes, such as mental rotation.
Mental rotation was one typical transformation of mental images, which was reported to be functional equivalence with visual perception. Whether there was an effect of global precedence on mental rotation was explored in the present research.
Compound stimuli combined with rotation task were used in two experiments. Participants were required to judge whether the large letter or small letter of the compound stimulus was in its regular format or left-right mirror. In Experiment 1, 2 (task: identification of normal image /mirror-reversed image for large or small letters )×2 (stimuli : congruence or incongruence of large-small letters) was designed to eliminate perceptual global precedence through prolonging the duration of target stimuli. Experiment 2 investigated the global precedence on mental rotation with a 4 (rotation degree: 0°, 60°, 120°, 180°) ×2 (rotation pattern: rotation of large letters or small letters) ×2 (stimuli: congruence or incongruence of rotation angles of large-small letters) within-subject factorial design.
Repeated measures analyses of variance were conducted separately for the two experiments. In experiment 1, perceptual global precedence could be eliminated by prolonging targets’ duration. There was no significant difference between the RTs of identification of normal image for large and small letters, and the congruity of large-small letters had no effect on identification of normal image for large and small letters. However, in experiment 2, when mental rotation was added to the task, large letters were responded more quickly than small letters, though the congruity of rotation angles of large-small letters didn’t influence the rotation of large and small letters; in addition, the difference between the rotation speeds of the large letters and small letters reached a marginal significant level.
Based on these results, the conclusion was that global precedence did exist in mental rotation even when perceptual global precedence was excluded in mental rotation task. Time required in global rotation was much shorter than that in local rotation. In addition, neither large letters rotation nor small letters rotation was influenced by the congruity of rotation angles of large-small letters. That was to say, the global precedence in mental rotation was of a little difference from the perceptual global precedence. The present findings not only extended the field of mental image, but also shed light on the research of global precedence in higher cognitive processes, and the method of eliminating perceptual global precedence was also of great significance for future research.

Key words: mental rotation, global precedence, compound stimuli