It has been proposed that readers and listeners generate predictions about what message is likely to be conveyed and, based on internal models, translate high-level expectations into progressively lower-level and form-based expectations about what features upcoming words will contain. When approaching such questions using materials from alphabetic languages, it is oftentimes unclear which level of representation is being predicted since phonological and orthographical features are correlated. I am currently developing a study using materials from Chinese, a language with opaque orthography, to inspect whether people generate expectations about what visual features the upcoming langauge stimuli likely contain.
Read more:
Listeners and readers have been shown to rapidly incorporate knowledge about the current communicative context to modify their predictions in the same context, i.e. adaptation effect. Such ability renders the processing system flexibility when encountering unfamiliar situations or fulfilling specific task demands. However, in a hierarchical predictive processing system, how prediction errors are interpreted at one level will be subject to how prediction errors are distributed at other levels. I am working on a few different projects to address how adaptation will happen in the face of different distributions of prediction errors across the processing hierarchy.
Read more:
Prosody, i.e. the pauses, metrics and intonation changes in speech, includes rich information about lexical identity, syntactic phrasing, as well as the intention of the speaker. I am interested in how prosodic information interacts with other linguistic cues to affect people's interpretation of the conveyed message.
Read more: