How does our language processing ability enable us to rapidly perceive, produce and understand language given the massive diversity observed across both speakers and languages?
We examine language processing in a breadth and depth previously unmatched in any one project. We will map processing at multiple levels of description and observation, in monolingual and multilingual individuals, and in typical and impaired populations. All these investigations will take place across a range of languages and dialects representing the unrivalled diversity in the Indo-Pacific region.
Two kinds of studies underpin our research into language processing:
The study of language processing naturally links with all other Centre programs. Links to Shape underlie our focus on cross-language comparison, and on the processing challenges posed by diversity in language structure. We link to Learning in examining processing across the lifespan, assuming that language processing is dynamic and subject to developmental change. Individual processing differences reveal sources of variation in individual speakers, and relates to variation amongst speakers which leads to language change and Evolution.
The Processing program makes significant use of research technologies and is actively developing new tools. The challenge of taking our basic research into remote field sites in Australia and beyond necessitates methodological innovation (e.g. eye-tracking technology in the field). We will apply new generation data collection techniques using smart phone apps (e.g. sound recordings, BabyTalk app for collecting longitudinal acquisition data) thus generating new forms of archival data.
We will take our research with clinical populations out of the laboratory and clinic, by using sophisticated recording devices (combined with GPS and movement recording technology) to study language in situ, providing a never-before-captured, objective and contextualised measure of language use. We will use the software tool Discursis to analyse conversations between people with neurodegenerative diseases (primarily Alzheimer‘s), to gauge how neurocognitive impairments affect language and vice versa.
These unique, in-context datasets will serve as customised databases for machine learning, delivering predictive language scripts to repair troubled language.