Abstract: Introduction Children younger than 4 years old often have difficulties in understanding others’ mental states (e.g., beliefs). One explanation for these difficulties is what is referred to as “the curse of knowledge” or a tendency to be biased by one’s own knowledge when appreciating an ignorant perspective. It is different from egocentric bias in that the curse of knowledge is asymmetric. The present study aimed to investigate the characteristics of the curse of knowledge in different tasks (knowledge-attribution and desire-attribution) and, particularly, to identify whether there is also an asymmetric bias in children’s desire for understanding.
Method
Two experiments were conducted based on 108 participating children ages 3 to 5 years old. In Experiment 1, a knowledge-attribution task was administered to these children. The children were asked to judge whether a puppet would know what was in a container (a box or an envelope), with which the children were either familiar or unfamiliar. Two conditions for children’s knowledge were designed. One was that children were knowledgeable of what was inside the container. The other condition was that children were unknowledgeable. In Experiment 2, a desire-attribution task was administered. The children were told of presents a puppet liked or disliked and were asked to choose a present for the puppet according to its likes and dislikes.
Results
The results showed that children demonstrated an egocentric bias in the knowledge-attribution task. When the children themselves were knowledgeable of the contents inside the container, they tended to overestimate the puppet’s knowledge of it. However, when the children were ignorant of the contents, they tended to underestimate the puppet’s knowledge. In contrast, children demonstrated an asymmetric bias in the desire-attribution task. Younger children at the age of 3 tended to assume their own favorite as that of the puppet’s even though the reality was otherwise. However, all participating children made much more successful inferences when their own preferences did not interfere.
Discussion and conclusion
The inconsistent findings in the two tasks might result from the differences of the task requirements and the contradiction between children and the character in the tasks. In the knowledge-attribution task, the conflict between children and the character was the only knowledge state about the reality. However, in the desire-attribution task, the conflict involved the desire. Because knowledge state was a kind of neutral state, whereas desire involved emotion, the degree of conflict differed between the two tasks. Furthermore, desire - attribution might require much more cognitive control to inhibit one’s own preference, which might be related to the asymmetric bias. Thus, in contexts requiring less cognitive control, children tended to be able to take others’ perspective much more easily than in a more conflicting context. In conclusion, the results suggest that perspective bias in younger children’ s mental state reasoning is context specific

Key words: mental state, theory of mind, perspective bias, children