Research

Below are brief summaries of some of the various projects I’m currently working on.

Cross-linguistic and inter-speaker variation in sibilant acoustics

I’m working on developing a new speech database for comparing patterns of cross-linguistic and inter-speaker variation. I’m adapting a number of large read speech corpora originally created for automatic speech recognition technology development, taken from a variety of sources, for phonetic research. To date, data for 32 languages and more than 4,600 speakers have been collected. I use this database in a submitted paper examining the spectral dynamics of /s/ (preprint here), and in ongoing work investigating how much of apparent gender differences in sibilant acoustics is physiologically-grounded (due to variation in vocal tract length) versus performed across languages.

Variation and change in Quebec French

I maintain the ParlBleu corpus of spoken Quebec French, a corpus of parliamentary speech created from proceedings of the National Assembly of Quebec, which can be accessed here. This is an expanded version of Peter Milne’s AssNat corpus (described in his thesis). I’ve lead work (with help from Laurie Dumont) to bring the corpus up to a total size of more than 31 hours of data across 156 speakers and to collect as much speaker metadata as possible to allow for sociolinguistic analyses. In a 2025 paper in Laboratory Phonology, I use the corpus to investigate the development of a rhotic quality in the vowels /ø, œ, œ̃/, examining change in F3 trajectory shape in apparent time and describing the social and regional patterning of rhoticity in Quebec French.

Articulation and acoustics in English rhotics

I’m also collaborating with Eleanor Lawson (University of Strathclyde) and Jane Stuart-Smith (University of Glasgow) on the relationship between tongue shape (bunched versus retroflex articulation) and formant (F1–F4) trajectories in rhotic sounds across varieties of English (Scottish, Irish, and North American). We’re investigating the extent to which formant patterns systematically relate to tongue shape, a point on which evidence from previous work is mixed.