About this srcd poster session

---

Panel information
Panel 20. Social Cognition

Abstract

---

Understanding the relationship between seeing and knowing is crucial for social interactions. While research shows that even infants understand this basic relation, past work treats perceptual access and knowledge as equivalent (if you see it, you know about it; if you don’t see it, you don’t know about it). In reality however, the relationship is more nuanced. More complex visual scenes usually require longer time to encode and are more interesting to look at. Across two experiments, we investigated whether children intuitively understand the complex relationship between seeing and knowing.

In study 1 (Figure 1A; pre-registered), we explored whether 5- to 6-year-olds (N = 46) believe that more interesting visual scenes usually make agents look longer. In each trial, children were shown two sets of objects, each placed inside an opaque box. One set was always more interesting than the other and children did not know which box had which set. An ignorant agent then peeked into one box for 1.5s (short-attention box) and another ignorant agent peeked into the other box for 5s (long-attention box; order and looking time counterbalanced). Children inferred which box contained which set of objects. Children completed three trials shown in Figure 1B. Older children were more likely to say the long-attention box had the more interesting set (i.e., ‘many’ in trial 1, ‘complex’ in trial 2, and ‘variable’ in trial 3; β=1.25, p<.05 by mixed-effects logistic regression with maximal random effects structure for participant and trial type), and their performance overall was significant in the `one vs. many` trial (66%; p<.05), but not in the `simple vs. complex` trial or the `same vs. variable` trial (48% and 53%, respectively; p>.43 in both cases).

Study 2 (Figure 2A; pre-registered) examined whether 5- to 6-year-olds (N = 46) understand that people who look longer have a more accurate representation of what they saw.Children completed three trials, in each trial, children were introduced to a box that had a toy inside: 1) a children’s book, 2) a fruit, 3) a stuffed animal. Two ignorant agents peeked into the same box for different amounts of time (same as study 1), and they reported a different answer about the content inside: 1) the book was about trains or cars, 2) an apple or a peach, 3) a tiger or a monkey (Figure 2B). Children were asked to endorse one of the answers. Children were significantly more likely to endorse the statement from the agent that looked longer (69% success, p < .001, by mixed-effects logistic regression with maximal random effects structure for participant, query type, and trial type) and their performance improved with age (β = 1.02, p < .05).

Overall, by age six, children expect that more complex visual scenes will reflect longer looking times, and that agents who look longer are more likely to have an accurate representation of what they saw. This reveals an early intuitive theory of attention that children can use to make rich inferences about other people.

Author information

Author	Role
Dr. Julian Jara-Ettinger, Ph.D., Yale University	Presenting author
Rui Zhang, Yale University	Non-presenting author
Marlene Berke, Yale University	Non-presenting author

⇦ Back to session