Visual discrimination ofabstract mirror - reflected three - dimensional objects at many orientations LAWRENCE

in detail (e.g., one's own body; Parsons, 1987b), the rate of imagined reorientation for different planes or axes can vary by a factor of at least 3. In addition, the rate can vary by a factor of at least 2.5 for different directions within a plane of orientation difference (i.e., for clockwise and counterclockwise rotations about an axis; Parsons, 1987b). For other stimuli (e.g., those used by Shepard & Metzler, 1971), the rate may not vary with the plane or direction of reorientation. Finally, the rate of imagined reorientation is apparently influenced by properties of the imaginedobject. Rates can vary by more than an order of magnitudedepending on the object's complexity or familiarity (cf. results in Cooper, 1975; Kaushall & Parsons, 1981; Parsons, 1987b; Shepard & Hurwitz, 1984; Shepard & Metzler, 1971). These and related results (e.g, Cooper, 1976; Pinker, 1980) are taken to imply that the imagined spatial transformations produce an approximately continuous series of intermediate internal representationsof a shape that correspond to its intermediate physical orientations. In addition, these results are thought to imply that objects are internally represented in three dimensions, rather than in two dimensions of projected three-dimensional information (as in a literally "pictorial" representation). This study examined people's discrimination of identical and mirror-reflected pairs of abstract three-dimensional objects like those used by Shepard and Metzler (1971).In contrast to previous research, the axis of the orientation difference (OD)between a pair of objects was quite unconstrained by experimental design. Each stimulus pair differed in orientation by an angle of rotation up to 180 0 about one of13 disparate axes. The axis ofOD was randomly ordered across trials. Subjects reported attempting to imagine one object at the orientation of the other and then comparing their shapes. Reaction time (RT) depended on the extent of OD, confirming earlier results. However, RT also depended on the axis or plane of OD between a stimulus pair. Slopes for RTOD functions for trials with identical objects varied by a factor of 3 for different axes. The rank order of functions from steepest to slightest slope was: (1) axes not in any principal plane of the observer's visual reference frame, (2) frontal diagonal axes, (3) horizontal diagonal axes, (4) the line-of-sight axis, (5) midsagittal diagonal axes, (6) the vertical axis, and (7) the horizontal axis perpendicular to the line ofsight. There was no difference in performance between subjects achieving high scores on pretests of spatial ability and those achieving moderate scores. The variation in slopes is likely due to variation in rate, path length, and/or initiation time of imagined reorientations. Various geometrical procedures could underlie the imagined spatial transformations usedin the task. RT patterns are much more consistent with the class ofprocedures producing paths of relatively efficient length than the class of inefficient procedures, such as that using rotations through Euler angles.