Expectations for upcoming content:Do children reason about speakers' informativity goals?

Research on typicality has disclosed seemingly contradictory patterns in comprehenders and speakers: typical content is easier for comprehenders to process (Kamide et al. 2003) but is less frequently produced by speakers (Brown & Dell 1987). However, recent work shows that comprehenders also pay attention to speakers’ production preferences (e.g. Rohde et al. 2021, 2022) – the more the speaker's role as an intentional communicator is emphasised, the more comprehenders expect contributions about non-typical content, i.e., content that is cooperatively informative (Authors, submitted). Here, we ask whether children reason about speakers’ informativity goals in adult-like ways. If child-directed speech favours mention of typical events and properties, particularly at younger ages (Bergey et al. 2021), children may expect speech to transparently reflect what the world is like. In ongoing data collection, participants (ages 5-11) complete sentences about what is found at different locations. A bare condition only mentions the location (e.g. “At the beach, there’s _____”); a visible speaker condition uses first person and adds a photo of a speaker (“I’m at the beach, and there’s _____”). So far, results (N=111) show that children expect more informative completions with the visible speaker, with a trend towards more adult-like behaviour as age increases.