Children use Others’ Emotional Expressions to Predict Their Behaviors in Classic False Belief Tasks

• In Event: 1-022 - Affective Cognition: Bidirectional Influences of Emotion and Social Cognition

Integrative Statement

Researchers have proposed that children construct an intuitive theory of others’ mental states in which representations of beliefs, desires, and emotions are causally linked (Harris, 2008; Wellman, 2014). Considerable research has looked at children’s ability to infer and predict emotions from beliefs, desires, and the broader context (e.g., Harris et al., 1989; Wellman & Banerjee, 1991). Here we are interested in the reverse inference problem, investigating whether emotional expressions provide cues to knowledge sufficient for predicting others’ behaviors based on their true and false beliefs.

To test this, we adapted the classic Sally-Anne task (Baron-Cohen, Leslie, & Frith, 1985). Although whether the task is a fair assessment of children’s false belief attribution is controversial (Scott & Baillargeon, 2017), here we are not interested in establishing the onset of children’s false belief understanding. Instead, as our initial step, we are interested in looking at children’s ability to use emotional cues to infer others’ beliefs. Thus in our study, we do not tell children whether Sally sees Anne move her toy. However, there is a window in the room so that Sally might have seen what Anne has done. Critically, when Sally returns, she looks either Angry or Happy. We predicted that children could use Sally’s emotional expression to decide if she saw the transfer (Epistemic Access Question) and then predict where she would look for her toy (Behavior Question).

We tested 4-to-6-year-olds in Experiment 1. For the Epistemic Access Question, there was no age effect (p=.461) and children at all ages succeeded (ps<.002). For the Behavioral Question however, there was an age effect (p=.048); four-year-olds performed at chance (p=.727), six-year-olds performed above chance (p=.002), and five-year-olds were in between (p=.016).

Why did four-year-olds succeed in recovering Sally’s epistemic access but fail to use this to predict her action? Experiment 2 suggests that four-year-olds might not have been simply confused by the question because while they performed at chance, three-year-olds predicted that Sally would look for her toy in its current location regardless of what Sally’s expression was (p=.004). Experiment 3 found that four-year-olds passed the original Sally-Anne task (with no emotions but given epistemic access directly; ps=.003). These suggest that four-year-olds have the conceptual knowledge to pass the classic false belief task, and can recover epistemic access from emotional expressions, but they fail to chain the two inferences together. Whether this is due to processing demands or the relative fragility of the representations remains a question for future research.

This study suggests that 4-to-6-year-olds can use someone’s emotional expression to recover her epistemic access, and older, but not younger, ones can use this information to decide if she will act consistently with true or false beliefs. These results are consistent with a small set of studies (Repacholi & Gopnik, 1997; Rieffe, Terwogt, & Cowan, 2005; Wellman, Phillips & Rodriguez, 2000) suggesting that emotional expressions provide an important entrée into others’ minds and offer a novel contribution to the literature on how children integrate emotion understanding with other mental state inferences.

Authors

• Yang Wu, MIT
  Presenting Author

• Anya Keomurjian, Wellesley College
  Non-Presenting Author

• Jennah Haque, MIT
  Non-Presenting Author

• Laura E Schulz, Department of Brain and Cognitive Sciences, MIT
  Non-Presenting Author