Six studies exploring children’s awareness of how learning occurs found that children’s understanding of learning developed during the preschool and early elementary school years and was correlated with their emerging ToM ability. Referring to the
research questions proposed in Chapter Three, the main findings of the current study can be summarized as the following:
1. Children’s understanding of learning as knowledge change developed between 4 and 6 years of age, during preschool and early elementary years.
Between 4 and 6 years, children came to understand that learning is the process of representational knowledge change. Young children started with a behavioral
understanding of learning. They understood that moving from not knowing to knowing is learning. However, they tended to over-attribute learning to failed attempts, and predict everyone needed to learn no matter whether the person knew the knowledge of interest or not. They seemed to follow a behavioral hint when making predictions that those who had learned before would do it again in the future. Younger children also found it difficult to differentiate learning that happened recently from learning that happened a while ago, similar to Taylor et al.’s (1994) findings. With age, children gradually realized that learning is a mental representational change; and one does not need to learn what one already knows. Their learning judgments and predictions improved with age.
2. Children found it extremely difficult to differentiate learning from mere actions that resemble learning but without the mental representational change; even
6-year-olds could not tell the difference between genuine learning and behavioral change.
Children’s difficulty with failed learning was magnified in Study 2 and Study 3 when a conflict was presented between the behavioral outcome and the mental
representation. They were drawn to the resemblance between a behavioral change and learning, and tended to over-attribute learning to events without a genuine knowledge change. Children tended to conclude that a person who could draw a circle that perfectly resembles a letter O, yet without the mental representation of the letter knew how to write a letter O and learned the knowledge in the action. Even though 6-year-olds finally responded that the person in the Coincidence tasks did not know at the end of the stories, they still thought the person learned, which revealed an incoherent, behavioral concept of learning independent of knowledge change.
3. Familiarity with the learning contents affected children’s learning judgments; the effect varied for children of different ages.
There was a significant improvement in children’s learning judgments in the familiar learning contexts from 4 to 6 years of age in Study 3. Children’s learning judgments in novel learning contexts, however, remained unchanged during the same period. Younger children might assume everybody else knew exactly what they knew due to epistemic egocentrism and therefore concluded both persons with or without genuine knowledge change learned. They picked between the two randomly. Older children, on the other hand, might be able to simulate their own experience of learning in their
resort to neither epistemic egocentrism nor simulation to help with the learning judgments.
4. Children began to appreciate that people’s learning intention was based on their beliefs about knowledge states; yet even 6-year-olds still had trouble
understanding learning intentions based on false beliefs.
Study 4 and Study 5 found it was easier for children to understand the implications of true belief in learning than false belief. In Study 4, even though the beliefs about the knowledge states were explicitly stated in the stories, children tended to reprocess the belief information in case of false belief. Learning necessity was easier than learning intention when controlling for yes bias and belief status, since the former is based on knowledge state and the latter is based on belief. Study 5 found that children up to 6 years of age failed to recognize the subjectivity of false belief by treating false belief the same as ignorance in learning intention prediction. The idiosyncrasy of the false belief might have made the contrast between true knowledge and ridiculous false belief more salient for children, which probably explains why it was easier for children to judge who would need to learn between the knowledgeable person and the person with false belief than between the knowledgeable person and the ignorant person.
5. Children appreciated learning intention as a causal force of the learning outcome. However, even 6-year-olds still found the conflict between learning intention and learning outcome hard to comprehend.
Study 6 found that children did better with tasks where the intention and the learning outcome were in consistent with each other than with tasks where the two were in conflict with each other. They were more likely to reprocess one aspect, usually the
learning intention instead of the learning outcome, to make it in line with the other in conflict tasks. Children did recognize that intention affects learning outcome; they understood that positive intention is usually associated with positive outcome, and vice versa. The fact that children could make perfect learning judgments based on physical cues independent of intention suggested that they understood to some extent that intention is neither a sufficient nor a necessary condition for learning. However, even 6- year-olds found the conflict between learning intention and learning outcome hard to comprehend.
6. Children’s learning understanding was correlated with their emerging ToM ability, particularly false belief understanding represented by the sum of
Content False Belief and Knowledge Access.
Across the six studies, various indicators of ToM were significantly correlated with children’s learning understanding when holding the effect of age constant. However, the most consistent correlations were between learning understanding and the sum of
Content False Belief and Knowledge Access, rather than the overall ToM score. The two tasks share a similar format and similar false belief construct. It seems false belief is the main player in mindful learning development, which is cross validated in Study 5 where common sense knowledge, most evident self knowledge such as one’s own sex and name, as well as factual knowledge in classic false belief stories all revealed the same outcome patterns in the learning prediction questions. The agreement suggested children’s
difficulty in learning tasks was inherently the same with that in ToM development as represented by false belief.
7. ToM measure with both mainland Chinese children and Hong Kong children generated reasonably scalable progressions, which largely replicated the sequence from other samples. Nevertheless, the differences between the current data and those from the literature did suggest culture specific characteristics of ToM development.
Adopting the well-documented ToM scale enables comparison between current data and that in the literature. The current samples were one year older than comparable samples in the literature; yet they shared similar passing rate on ToM items, indicating Chinese children develop ToM understanding later than their Western counterparts. HK children scored lower than CQ children on the scale, even though the difference was not statistically significant. Knowledge Access was easier than Diverse Belief for both CQ and HK children, similar to what was documented in the literature with Chinese children. Younger children in both mainland and Hong Kong samples demonstrated more
advanced emotion understanding compared to false belief understanding. HK children even demonstrated a reversed sequence between Hidden Emotion and False Belief, similar to Peterson et al.’s (2005) autistic children’s pattern. The evidence suggested that children in different cultures might develop ToM understanding via different routes.