Poster #73 - Where Do Differences in Theory-of-Mind Development Come from? An Agent-Based Model of Social Interaction and Theory-of-Mind

• In Event: Poster Session 05

Abstract

Theory-of-Mind (ToM), the ability to understand others’ internal states, emerges early in life and develops in childhood (Wellman, 2014). Although all humans develop ToM (Yu & Wellman, 2022), they do so at different paces: autistic and deaf and hard-of-hearing children have delayed ToM development (Peterson et al., 2019), whereas bilingual children develop faster (Yu et al., 2021). Even within typically-developing children, some develop ToM faster and some slower.
Where do these differences come from? Social interaction may well be a cause (Astington, 2001): more social interaction means increased opportunities to think about others’ internal states, thus enhancing ToM; less social interaction, in contrast, delays ToM development.
Nevertheless, the causal link between social interaction and ToM is tricky to establish because of several challenges. First, children’s social interactions are too complex to precisely measure as they are confounded with many hidden home and schooling factors. Second, given the limits of existing ToM assessments (e.g., false belief tasks) and the fact that ToM grows slowly over childhood (Wellman, 2014), measuring ToM growth day-by-day or week-by-week over extended periods is impractical, making it difficult to track how social interaction provokes ToM development. Third, causal experiments that manipulate children’s social interactions over such extended periods are likewise impractical. As a result, the causal link between social interaction and ToM remains only indirectly and patchily demonstrated at best.
Here, we take a different approach and use computational modeling, in the form of agent-based models (ABMs), to demonstrate how social interaction causally influences ToM development. ABMs are simulations where cognitive “agents” interact over time and space under specified manipulations, offering opportunities to do what is infeasible in the real world (Madsen et al., 2019).
Figure 1 shows the graphical demonstration of our ABM that simulated a group of children in a neighborhood setting. The children randomly moved in an environment, causing them to encounter and interact with one another. If these interactions continued long enough, their ToM probabilistically improved. Focally, we manipulated the amount of social interaction in a time interval (e.g., a tick of a computer’s clock). Using this ABM, we observed the total time (ticks) needed for a group of children to fully develop their ToM and the effects of different amounts of social interaction.
Our results, summarized in Figure 2, showed that the amount of social interaction strongly influenced how fast a group of children fully developed ToM, where more social interaction led to faster development. These results hold regardless of the group sizes and also regardless of whether children began with an existing social network or not (e.g., children who had or had not met each other before).
In sum, in an ABM, we empirically demonstrate a causal link between social interaction and ToM and show how social interaction levels account for differences in the pace of ToM development. This demonstration not only helps show how differences in ToM development emerge but also helps specify an underlying mechanism for how children develop ToM and vary in that development.

Authors

• Chi-Lin Yu, University of Michigan - Ann Arbor
  Presenting Author

• Henry M. Wellman, University of Michigan