UC San Diego

In this dissertation we investigate the relationships between syntactic structures and the meanings of the events they are used to describe, how such relationships may develop within language, and how they may be acquired by language learners. We employ a novel paradigm to assess these relationships through an effect termed syntactic entrainment. Over the course of 9 experiments we explore the possibility that syntactic entrainment reflects a process by which these relationships may be introduced into natural languages, and acquired by children learning language.

Study 1 demonstrates syntactic entrainment. In Experiment 1, we show that when a speaker hears an event described with a particular syntactic structure they will tend to use the same structure when subsequently describing the same event. In Experiment 2, we demonstrate that this effect is equally likely to be present when speaking to the same or a different interlocutor. However, in Experiment 3 we demonstrate a small but significant partner specific component of the syntactic entrainment effect, but only when subjects are given four identical descriptions of the same picture. This result suggests that speakers are creating enduring, primarily partner independent associations between syntactic structures and event content.

In Study 2 we show that the associations reflected in syntactic entrainment apply not just to the particular depictions of events, but also to visually distinct depictions of the same events (Experiment 2), and even to larger categories of events defined by specific event semantic features (Experiment 3). This suggests that syntactic entrainment could reflect the mechanism by which language users learn about the associations between syntactic structures and particular event meanings, which we find in natural language.

Finally, in Study 3 we show that for children – but not adults – the magnitude of the syntactic entrainment effect is sensitive to the main verb used in the encoding sentences and their target descriptions. These results suggest that 4–6 year old children may still be using the identity of verbs to learn about associations between syntactic structures and event meaning features found in natural language, whereas adults may no longer rely on this information for grammatical language use.