Preschoolers Consider Probability Information When Inferring People’s Surprise

• In Event: Probabilistic Emotion Signals Guide Inference and Behavior

Abstract

How do young children understand surprise? The predominant view is that children conceptualize it in terms of beliefs (MacLaren & Olson, 1993; Ruffman & Keenan, 1996; Wellman & Bartsch, 1988). However, children also understand surprise in terms of probability, though this understanding appears relatively late in development. It is not until age 7 that children spontaneously infer that people are surprised by improbable events but not probable ones (Doan et al., 2018). Here we investigate whether preschoolers recognize this if probability information was made more salient. Specifically, we examined whether they recognize that a low probability event is surprising, but a deterministic one is not.
In Experiment 1 (n=120), 3-4-year-olds saw stories about two characters who are at different gumball machines; Figure 1. In the Probabilistic condition, one character uses a machine with mostly red gumballs and few purple ones, and the other character uses a machine with mostly purple gumballs and few red ones. In the Deterministic condition, one character uses a machine with only red gumballs, and the other character uses a machine with mostly purple gumballs and few red ones. In both conditions, both characters receive a red gumball. Children judged which character was surprised to get a red gumball. We only found a significant main effect of condition, p=.017, where children chose the character who was at the machine with fewer red gumballs more so in the Deterministic condition than the Probabilistic condition; Figure 1. Children in the Deterministic condition mostly chose the correct character, p=.038, whereas children in the Probabilistic condition performed at chance, p=.168. These findings show that preschoolers only recognize that an improbable outcome is surprising if it was contrasted with a deterministic outcome. In the next experiment, we wanted to see if we could replicate these findings with a slightly different task.
In Experiment 2 (n=120), 3-4-year-olds saw stories that were identical to the Deterministic condition of Experiment 1. That is, all children saw stories where one character uses a gumball machine with all red gumballs, and one character uses a machine with many purple gumballs and just a few red ones. Both characters receive a red gumball. However, what children were asked differed by condition. In the Surprise condition, children were asked to judge which character was surprised to get a red gumball. In the Knowledge condition, children were asked to judge which character knew they would get a red gumball. We found only a significant main effect of condition, p=.003, where children chose the character who was at the machine with fewer red gumballs more so in the Surprise condition than the Knowledge condition; Figure 2. Children in the Surprise condition mostly chose the character at the machine with fewer red gumballs, p=.036, whereas children in the Knowledge condition mostly chose the other character, p=.038.
Together, these findings provide initial evidence that even preschoolers consider probability information when inferring people’s surprise.

Authors

• Tiffany Doan, University of Toronto

• Ori Friedman, University of Waterloo

• Stephanie Denison, University of Waterloo
  Presenting Author