Spontaneous Structural Alignment Contributes to Children’s Learning of Theory of Mind

• In Event: Analogy and Cognitive Development

Abstract

Theory of mind (ToM) refers to the ability to reason about the mental states of others and oneself, including desires, beliefs, emotions, intentions, and knowledge. One critical insight that is learned over development is the distinction between true-beliefs and false-beliefs. Children under 4 routinely fail false-belief tasks, as though they implicitly assume that beliefs match the state of the world. The stages of ToM understanding are well-documented (Wellman & Liu, 2004), but there is no consensus on what learning processes drive this development. Here we ask whether analogical processes contribute to children’s ToM learning.

Children (and adults) can gain insight by carrying out explicit analogical mappings, noting the object correspondences and the structural alignment (Gentner, 2010). Here, we take a different tack: we ask whether children can gain insight into ToM from invisible analogies—structural alignments that occur spontaneously in their experience. Here, we used a sequence designed to promote spontaneous alignment—the repetition-break sequence (RBS) (Loewenstein & Heath, 2009). In this sequence, two highly alignable events are followed by another alignable event that differs in some key way. The idea is that if the first two events are sufficiently similar, they will spontaneously align, fostering schema abstraction. When the final event is aligned with this schema, the difference between them will pop out as an alignable difference. This pattern is often used in children’s stories (e.g., The Three Little Pigs).

To test whether the RBS could lead to insight into ToM, we used a Pretest-Training–Posttest design. The Pretest and Posttest were three standard false-belief tasks using different specific content from each other and from the training materials. In the Training task, 4-year-old children were shown a sequence of short vignettes in which a character was shown to have a belief about the contents of a box, and then the actual contents were revealed (Figure 1). The first two vignettes depict a true belief; the third depicts a false belief. If children spontaneously align these vignettes, structural alignment between the first two situations will highlight their common true-belief structure; and alignment with the third (false-belief) event will render the critical difference between true and false beliefs salient. To test for the critical role of alignment, half the children received high-surface-similarity vignettes, which are especially likely to be spontaneously aligned (HA); the other half received low-surface-similarity vignettes (LA). Children were asked questions about individual vignettes but were never asked to compare the vignettes.

The results showed that, as predicted, (1) children gained in false-belief understanding from pretest to posttest; (2) the gain was greater for children who received high-similarity (easily aligned) vignettes than for those who received low-similarity vignettes—consistent with the claim that structural-alignment processes are driving this learning. In a further experiment, children failed to learn from the same vignettes in reversed order.

These results show that analogical comparison processes operating over social experiences can contribute to ToM development. We suggest that cognitive development is driven partly by invisible analogies—structural alignments that occur spontaneously in children’s experience.

Authors

• Dedre Gentner, Northwestern University
  Presenting Author

• Christian Hoyos, Exponent
  Non-Presenting Author

• William Horton, Northwestern University
  Non-Presenting Author