Symmetry Lasts Longer Than Random, but Only for Brief Presentations

Abstract

Previous research has shown that explicit emotional content or physical image properties (e.g., luminance, size, and numerosity) alter subjective duration. Palumbo recently demonstrated that the presence or absence of abstract reflectional symmetry also influenced subjective duration. Here, we explored this phenomenon further by varying the type of symmetry (reflection or rotation) and the objective duration of stimulus presentation (less or more than 1 second). Experiment 1 used a verbal estimation task in which participants estimated the presentation duration of reflection, rotation symmetry, or random square-field patterns. Longer estimates were given for reflectional symmetry images than rotation or random, but only when the image was presented for less than 1 second. There was no difference between rotation and random. These findings were confirmed by a second experiment using a paired-comparison task. This temporal distortion could be because reflection has positive valence or because it is processed efficiently be the visual system. The mechanism remains to be determined. We are relatively sure, however, that reflectional patterns can increase subjective duration in the absence of explicit semantic content, and in the absence of changes in the size, luminance, or numerosity in the images.

Keywords: arousal; internal clock; subjective duration; symmetry; time.

Figure 1.

Schematic diagram of scalar expectancy theory (SET: Gibbon, Church, & Meck, 1984).

Figure 2.

Illustration of the image types used in the experiments. Left panel: reflectional symmetry. Middle panel: random. Right panel: rotational symmetry. Every trial used a different example without any repetition.

Figure 3.

Mean verbal estimates (ms) plotted against the standard duration (ms). Error bars show standard error.

Figure 4.

Mean proportion of long responses for reflection and random in the short (400–600 ms) and long (1400–1600 ms) conditions. Error bars show standard error. *p < .05.

Figure 5.

Mean proportion of “comparison long” responses plotted against the comparison or standard. Upper panel shows data from the short duration range and lower panel shows data from the long duration range.