Spontaneous Perspective Taking in Humans?

Abstract

A number of social cognition studies posit that humans spontaneously compute the viewpoint of other individuals. This is based on experiments showing that responses are shorter when a human agent, located in a visual display, can see the stimuli relevant to the observer's task. Similarly, responses are slower when the agent cannot see the task-relevant stimuli. We tested the spontaneous perspective taking theory by incorporating it within two classic visual cognition paradigms (i.e., the flanker effect and the Simon effect), as well as reassessing its role in the gaze cueing effect. Results showed that these phenomena (e.g., the Simon effect) are not modulated according to whether a gazing agent can see the critical stimuli or not. We also examined the claim that previous results attributed to spontaneous perspective taking are due to the gazing agent's ability to shift attention laterally. Results found no evidence of this. Overall, these data challenge both the spontaneous perspective taking theory, as well as the attentional shift hypothesis.

Keywords: perspective taking; social attention; theory of mind; vision.

Figure 1

A stimulus used in Experiment 1. Note that in order to have parity with our other experiments, the barriers were also present, but always allowed the avatar to see the lateral wall.

Figure 2

Mean RT (reaction time) and error rates from Experiment 1 together with standard errors. Consistent/Inconsistent refers to the location of the flanker with respect to where the avatar was looking. Congruent/Incongruent refers to the identity of the flanker with respect to the target letter, i.e., the manipulation that generates the classic flanker effect.

Figure 3

The stimuli employed in Experiment 2. In the left panel, the target is on the left side for both the participant and the avatar. By contrast, in the right panel the target is on the left side for the participant but on the right side for the avatar. Note that we again placed barriers either side of the avatar but the avatar could always see the targets.

Figure 4

Mean RT and error rates from Experiment 2 together with standard errors. Compatible/Incompatible refers to the manipulation that generates the classic Simon effect.

Figure 5

Example of a stimulus employed in Experiment 3. The figure shows a seeing condition valid trial.

Figure 6

Mean RT and error rates in Experiment 3. Standard error bars are also shown.

Figure 7

Stimuli used in Experiment 4. The example shows a seeing condition, valid trial.

Figure 8

Mean RT and error rates from Experiment 4 together with standard errors.

Figure 9

Mean RT and error rates from Experiment 5 together with standard errors.

Figure 10

Examples of the stimuli used in Experiment 5. In the left panel the cue is the (‘valid’) avatar facing the target. In the right panel, the cue is an object onset (i.e., the grey line) that appears adjacent to the target.