Understanding the social world is one of the most important challenges of children’s early development, as social cognition forms a foundation for everything from language learning to relationship building, and later academic success. The current project enhanced understanding of a novel aspect of social reasoning: How people learn about others from the objects they create or choose, and how this reasoning develops in childhood. In doing so, this project increased understanding of the way that social cognition operates in the mind, and how it forms a foundation for positive outcomes in child development. In addition, the project also involved training graduate and undergraduate students on all aspects of research, including experimental and computational modeling techniques.

Over the course of the project, we constructed computational models of adults’ and children’s reasoning, and conducted behavioral experiments with both adults and children to examine the developmental trajectory and mature state of social reasoning about objects’ designs. Overall our data show that when thinking about objects, children as young as 4 years of age can think about the design process. This means that when seeing an inanimate object, children can draw richly social conclusions: What shaped the creator’s choices (their goals, constraints), and where their idea came from (whether it was copied, or an original idea). We find that with age (over early-to-middle childhood), children become more adept in reasoning about the design process. For example, by age 6-7 years, children can think about the design process even when they have not experienced creating a similar object themselves. Our computational models provide a formal account of this reasoning as inverse planning, allowing for specific, quantitative predictions of children’s and adults’ judgements.

Several related lines of experiments also provide evidence that this kind of complex, nuanced reasoning explains several related aspects of children’s and adults’ social thinking about objects. For example, we find that children reason about the social meaning of others’ shared preferences for objects in a similar way, by considering others’ underlying goals and social intentions. In addition, we find that adults use social reasoning about others’ goals and beliefs when reasoning about others’ descriptions of objects, and whether they are deceptive, or true.

In each of these cases, we construct and test formal computational models of the social reasoning processes, allowing this work to inform both (1) our understanding of social cognition and how it develops, and (2) our understanding of how to construct artificial intelligences (AI) that reason about the world in a human-like way.

The work supported by this project led to the publication of 2 peer-reviewed papers, with 2 other papers currently under review; 4 peer-reviewed conference papers; and 12 presentations at scientific conferences.

The project has also made substantial contributions to training graduate and undergraduate students, including several members of underrepresented groups. The lab has been engaged in outreach throughout the project, interacting with over 500 children and families across the U.S. via in-person and live virtual interactions with researchers. We have also disseminated our findings to many additional members of the public by sharing our findings via parent-focused posts on social media, and by sharing open-access materials in line with best practices of transparency and open science.

Last Modified: 11/23/2022
Modified by: Adena Schachner