Research
Some of my research foci include:
Language variation and change in Beijing dialect
Rhotacization
My PhD research focuses on rhotacization (erhua), a salient feature of the Beijing dialect. It addresses two central questions:
How can we define the variable context of rhotacization?
Identifying where rhotacization can occur is essential to understanding its variation. Although every syllable in the Beijing dialect is phonologically compatible with rhotacization, this does not mean that every morpheme or word allows it. Even words that permit rhotacization do not always occur with it. My research examines these distinctions to establish the contexts in which rhotacized and non-rhotacized forms can alternate, providing a basis for quantitative variation analysis.
Who contributes to rhotacization’s status as a defining feature of Beijing speech?
The association between rhotacization and Beijing is widely recognized, but how can we investigate that relationship quantitatively? I approach this question by examining speakers’ relationships with the city. Using a place attachment score informed by local knowledge, I investigate how speakers’ connections to Beijing relate to their use of rhotacization, moving beyond broad demographic categories to explore the relationship between language and place.
You can find my PhD thesis here.
Classifier omission
Beyond rhotacization, I’m currently extending my research to classifier omission, a distinctly local vernacular feature that, unlike rhotacization, is not a widely recognized linguistic stereotype. Investigating classifier omission will allow me to extend the analysis to variation at the grammatical level while also exploring co-variation between features with different degrees of social salience. By examining the two variables together, I aim to investigate whether individual speakers’ use of rhotacization aligns with their use of classifier omission, and how these patterns relate to broader community-level tendencies.
Beijing Dialect Corpus
A central foundation of my research on the Beijing dialect is the dataset I collected between September 2021 and June 2022. Working with 58 participants, I conducted sociolinguistic interviews and several elicitation tasks to capture language use across different speaking contexts. These data allow me to examine variation both across speakers and within individual speakers, exploring how linguistic behaviour and patterns vary with context.
I welcome opportunities to collaborate with others working on the Beijing dialect and am happy to discuss sharing these data for comparative research across participants, linguistic features, and speaking contexts.
Chinese syntax
Beyond my work on the Beijing dialect, I am interested in grammatical variation in Mandarin Chinese, particularly how syntactic structure, meaning, and discourse context interact to shape speakers’ choices.
In Contextualizing the variation in causal clause ordering in Mandarin Chinese: a multifactorial analysis, co-authored with Qiao Gan and published in Corpus Linguistics and Linguistic Theory (2025), we investigate why speakers place causal clauses before or after the main clause. Drawing on a corpus of spoken Mandarin and using mixed-effects regression models, we examine how multiple contextual factors jointly condition this variation. Our findings show that clause ordering is shaped by clause length and complexity, emotional valence, and the use of connectives such as yīnwèi (‘because’) and suǒyǐ (‘so’), with interactions between these factors and discourse genre.
I am currently developing research on conditional clauses in Mandarin Chinese, extending my broader interest in how speakers express relationships between clauses and how grammatical choices are shaped by the contexts in which they occur.
Production and perception
Beyond language production, I am interested in how people perceive linguistic variation: how listeners recognize accents, distinguish between varieties, and form impressions of speakers. Which linguistic cues do listeners attend to, and how do these cues work together to shape perception? How are these interpretations influenced by listeners’ own linguistic backgrounds, experiences, and expectations?
Rhotacization provides one example of these broader questions. Although it occurs in Beijing, Tianjin, and northeastern Mandarin varieties, it may evoke different regional and social associations depending on how it is used and the other features with which it co-occurs.
By connecting production and perception, my work explores the relationship between the patterns speakers produce and those listeners notice, and how linguistic differences become socially meaningful.
Human and AI voices
Building on my background in human speech production and perception, I am developing research on how listeners evaluate AI-generated voices. Naturalness is a central criterion in voice AI evaluation, but what makes a voice sound natural, robotic, or human-like, and how do listeners arrive at these judgments?
I am interested in how features such as intonation, rhythm, voice quality, pauses, and pronunciation variation combine to shape these perceptions. A key focus is determining when conversational features, such as hesitation, contribute to a convincing voice, and when they sound exaggerated or inappropriate. I also investigate how listeners’ linguistic backgrounds, expectations, and the interactional context shape what counts as natural or “human enough”.