Search
Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
Browse Posters
Search Tips
Register for SRCD23
Personal Schedule
Welcome Letter
Program Guide
Change Preferences / Time Zone
Sign In
Imagine joining a friend for a drink, and meeting a mutual friend (Sam) for the first time. You mention that you recently saw a Scorsese movie, and your friend turns to Sam and says “Martin Scorsese is a director who did The Departed and a bunch of other films.” Before Sam has even spoken, you’re able to form an expectation about their underlying knowledge in this case-- namely that Sam is likely unfamiliar with Scorsese. The current research explores this ability and its development in early childhood.
The inference above is guided by the idea that how we speak is fundamentally shaped by our assumptions about a listener’s knowledge (i.e. listener design, Clark & Murphy, 1982). Listener design is predicated on the idea that you (the speaker) first have an understanding of the listener’s knowledge and then you can adjust your speech accordingly. But what about the reverse inference-- do we also use aspects of how someone speaks to infer the listener’s underlying knowledge state? And when in development do children do this?
In children’s own language production, recent work suggests that children by age 5 take into account their partner’s knowledge state to selectively talk about more general or specific characteristics of an object (Baer & Friedman, 2018). Additionally, children by age 5 attribute greater knowledge to speakers who offer generically presented information, rather than specific (Koenig et al., 2015). Our study expands these findings to ask how children use a speaker’s description to infer not the speaker's knowledge, but the listener’s knowledge.
We conducted a pre-registered study with children ages 4 to 9-year-olds (N = 60, tested remotely over zoom). Participants were shown a child who speaks to two different listeners-- offering one a low knowledge description (e.g. “This is an astronaut. Astronauts get to go to space”), and the other a baseline description (e.g. “This is a cute astronaut. I like playing space with this astronaut”). Children were then asked to make an ignorance evaluation (“Which person has never seen a toy like this before?”). 6- to 7-year-olds (proportion of trials = 0.71, p < 0.01) and 8- to 9-year-olds (proportion of trials = 0.97, p < 0.001) reliably selected the low knowledge listener as ignorant about the toy whereas 4- to 5-year-olds did not (proportion of trials = 0.56, p = 0.51). In other words, older children inferred that someone is less knowledgeable when that person is given a general description that implies lower knowledge, without ever hearing that person speak themselves. Because this question had not been probed in adults, we ran an analogous study with adults (N = 50) and confirmed that they too make these inferences based on listener design. We have long known that a speaker’s utterance tells you a lot about how knowledgeable the speaker is, but this work demonstrates that such utterances can also imply things about a listener's knowledge. This work opens up exciting new questions about how children (and adults) make nuanced inferences about others based on listener design.