Questions: Mental Rotation and Spatial Transformation

Question 1

If mental rotation were a purely symbolic process — like looking up whether two patterns are identical in a mental database — what would the Shepard-Metzler reaction time data look like?

• A. Reaction time would increase linearly with rotation angle, since larger angles require longer lookup times
• B. Reaction time would be roughly constant regardless of rotation angle, since a symbolic lookup doesn't depend on angular distance
• C. Reaction time would decrease with rotation angle, because larger differences are easier to detect
• D. Reaction time would vary randomly with no systematic relationship to angle

Answer: (B)

The whole force of the Shepard-Metzler finding is that a symbolic lookup system has no reason to care about angular disparity — if you're just checking whether two descriptions match, it should take the same amount of time regardless of how the objects are oriented. The fact that RT increases linearly with rotation angle implies the mind is actually stepping through intermediate orientations at a constant rate — an analog process that simulates the rotation rather than computing it symbolically.

Question 2

Two subjects each judge whether pairs of 3D block figures are identical or mirror images. Subject A sees pairs oriented 160° apart; Subject B sees pairs oriented 40° apart. Based on the analog model of mental rotation, what does the RT data predict?

• A. Both subjects should respond in roughly the same time, since the judgment is binary (same or different)
• B. Subject B should be faster because small angular differences are harder to discriminate precisely
• C. Subject A should take about four times as long as Subject B, reflecting the proportional rotation rate
• D. Response time depends on whether the figures are same or mirror-image, not on the rotation angle

Answer: (C)

The analog model predicts a linear RT-angle relationship. If rotating 40° takes time T, then rotating 160° (four times the angle) should take approximately 4T. This linear proportionality is exactly what Shepard and Metzler found — rotation rates are roughly constant, so larger angles require stepping through more intermediate states, taking proportionally longer. Option A is the symbolic prediction; option D is false because both same-pair and different-pair responses show the linear RT-angle function.

Question 3

The finding that reaction time in mental rotation tasks increases linearly with the angular difference between two figures supports an analog rather than a propositional model of mental representation.

• T. True
• F. False

Answer: True

This is precisely the evidential logic of the mental rotation paradigm. Propositional representations are abstract symbol structures with no intrinsic spatial properties; rotation angle should be irrelevant to a symbolic lookup. The clean linear RT-angle function implies the representation preserves spatial geometry — that the 'rotation' is a real traversal of intermediate states, not a computation. This is the defining feature of an analog representation.

Question 4

Individual differences in mental rotation ability reflect innate, fixed biological capacities that remain stable across the lifespan regardless of experience.

• T. True
• F. False

Answer: False

Mental rotation ability is trainable with practice. While robust individual differences exist — including some of the largest cognitive sex differences consistently found in psychology — these are not fixed biological limits. Practice systematically improves rotation speed and accuracy. The causes of baseline individual differences are actively debated (biological, experiential, and motivational factors have all been implicated), but the malleability of the ability is well established.

Question 5

Why does the linear relationship between reaction time and rotation angle in mental rotation experiments challenge propositional theories of mental representation?

Model answer: Propositional theories hold that mental representations are abstract symbol structures — like sentences in a mental language — with no intrinsic spatial properties. If mental rotation were just a symbolic lookup, angular disparity would be irrelevant to response time: matching two descriptions doesn't take longer just because one object is more rotated. The linear RT-angle function shows that spatial geometry is directly reflected in the time course of cognition — the mind appears to step through intermediate orientations at a constant rate. This 'passing through' intermediate states is only possible if the representation preserves metric spatial structure, which is the hallmark of an analog (not propositional) representation.

The key move is understanding what propositional theories predict (angle-independent RT) versus what was found (linear RT increase with angle). The linear relationship is direct evidence that the representational format encodes spatial properties, not just abstract relational descriptions. This is why the Shepard-Metzler experiments were so influential in the imagery debate: they gave behavioral evidence for a genuinely spatial representational medium.