About me

Hello!! My name is Xinyu /ɕɪn.ɥy/ (like “Shin-yu” in English) Listen to my name in Mandarin Chinese

I’m a second-year PhD student in Linguistics at the University of Toronto. I’m a phonetician and computational psycholinguist by training, focusing on speech production and perception. The central focus of my research is how linguistic, cognitive, and social factors shape the speech signal variability in production and perception during speech communication. I enjoy combining methods from different subfields, including psycholinguistic experiments, computational language models, large-scale corpus analysis, cross-linguistic phonetic universals, and theories of linguistic change.

my CV is here: last updated in Oct 2025

Research

Surprisal effects in the information packaging of speech variation

Supervisors: Dr. Barend Beekhuizen (primary advisor) and Dr. Jessamyn Schertz

This is my ongoing project for the second Generals Paper at U of T. We are investigating how surprisal derived from language models predicts variation in speech signals in large-scale corpora.

Perceptual Recalibration and Sound Change

Supervisor: Dr. Yoonjung Kang

Collaborator: Dr. Hironori Katsuda

By exposing listeners to innovative phonetic input, we explored biases in the perceptual recalibration of phonemic categorization, with implications for the role of phonetically grounded typological biases and word frequency encoding in exemplar phonology in diachronic sound change. The languages that we are currently working on include North American English and Japanese.

News

I will present my paper, co-authored with my advisor, Yoonjung Kang, at the LSA Annual Meeting 2026 in New Orleans!