How do children read minds in conversation?
To navigate the complex social world, children must learn to figure out what other poeple know, want and believe not just from
what they say, but
how they say it. In everyday conversation, our inner lives leak out in all manner of cues—written on our faces, brimming in our silences, hinted at in what we leave unsaid. My research investigates how children draw rich mental inferences from these subtle conversational cues. While even infants readily recruit mental representations to navigate conversation, my work asks how conversational cues help children build and revise those representations in the first place. These skills are crucial not only for children to become smooth conversationalists, but also for social learning, theory of mind, and emotion understanding.
Many developmental studies assume that the relevant mental state is already available (for instance, we establish that the child knows “this person cannot see X”) and tests whether children can use that representation appropriately. But in real-world communication, we often face a harder problem: figuring out what others know, believe, or feel based on subtle and noisy cues. My work asks how children build these representations in the first place, using information embedded in language, timing, disfluency, and conversational context. Below are examples of several lines of my ongoing research:
Delays and disfluencies (“um”) are ubiquitous in everyday conversation and often signal rich information about what’s going on inside a speaker’s mind. How—and how quickly—something is said can be as meaningful as what is said. In this line of work, I investigate how children (and adults) use disfluencies to infer a speaker’s underlying mental processes. My research explores how children use these cues to reason about just how much someone knows and what their true underlying feelings are. In this work, I argue that children’s reasoning about delays and disfluencies is inferential (vs. heuristic), broad (vs. limited), and flexible (vs. fixed).
Imagine a young boy expressing a gender counter-stereotypical preference (e.g., wanting to buy a Barbie doll) and his caregiver provides a permissive, gender egalitarian response. However, imagine that response comes slowly, with markers of surprise and production difficulty (e.g., “Oh! Um. . . Sure”). What message does that young boy really receive?
While people may be reluctant to explicitly state social stereotypes, their underlying beliefs nevertheless emerge through subtler conversational cues, such as surprisal reactions that reveal expectations. In this line of work, I investigate how messages that are explicitly permissive and outwardly egalitarian (e.g., “Sure, you can have that one”) might nevertheless be interpreted very differently based on presence of surprisal cues like interjections (“oh!”) and disfluencies (“um”). This work provides emerging evidence that conversational feedback may play a critical and underappreciated role in the transmission of social beliefs ( Morris & Shaw, 2024).
