My main research interest, and the focus of my PhD, is the effect of old age on sentence production. Some aspects of language processing seem to be more negatively affected by ageing that others and I am interested in whether the way we plan and produce sentences changes with old age. To examine this, I have conducted multiple studies in which I measure speech onset latencies in young and older adults when producing different sentence types: the time a speaker takes to begin sentence is informative about underlying sentence planning processes. In this study, we found that while syntactic planning scope appears to be relatively unaffected by old age, there is an age-related decline in the integration of lexical items into syntactic structures.
Syntactic priming is a behavioural phenomenon by which speakers are more likely to repeat a syntactic, or grammatical, structure if they have recently processed a sentence of the same structure. The magnitude of syntactic priming can be used to gain insight into nature of a speaker’s linguistic representations. We are using this technique to examine syntactic representations in young children, whose language skills are still developing, and older adults, who are experienced language users but who may be experiencing changes in way syntax is represented within the lexicon. In this study, we found that syntactic priming and the related phenomenon of lexical boost (repetition of verb enhances the priming effect) were preserved in old age.
Successful language processing involves the binding of multiple words into larger syntactic structures (or sentences) with more complex meaning. This is essential to the expressive power of human language, and we are interested in the neural processes that support syntactic binding. To examine this, we are conducting a studying using magnetoencephalography (MEG) in which we measure participants' neural responses (specifically, the magnetic fields produced by electric signals in the brain) when processing sentences that require syntactic binding vs. wordlists that do not require binding.