Have you ever found yourself taking on aspects of someone else’s pronunciation after listening to them speak? This phenomenon, known as phonetic imitation, is one of the main areas of research in our lab.
People vary considerably in how much they imitate. They also vary in how readily they can hear imitation in other people’s speech. In this study, we asked whether these two abilities are connected: Are the people who produce the most imitation also the ones who are best at perceiving it?
Twenty-seven native speakers of Canadian English took part in two experimental sessions. In the first, we measured how much they imitated. Participants initially read a list of words aloud to establish their usual pronunciation. They then immediately repeated recordings of a model talker saying the same words. In the model’s speech, words beginning with sounds such as p, t, and k had an extended puff of air, giving participants a clear pronunciation feature they could potentially reproduce.
To determine how much each participant imitated, 200 additional listeners compared their recordings from before and during the imitation task with the model talker’s speech. The more often listeners judged a participant’s imitation recording as similar to the model, the higher that participant’s production score.
Two to three weeks later, the original participants completed a listening task measuring how well they could detect imitation produced by other speakers. Both measurements used a task called the 4IAX, in which listeners heard two pairs of recordings and decided which pair sounded more similar.
Participants varied substantially in how much imitation they produced: individual scores ranged from 45% to 76%. They also varied in how much imitation they perceived, although most perception scores were clustered closer to 50%.
There was a small positive relationship between the two measures: people who produced more imitation tended, on average, to perceive slightly more. However, this relationship was not statistically reliable. Some of the participants whose speech showed the most imitation differed considerably in how well they detected imitation in others.
For now, then, the connection between perceiving and producing phonetic imitation remains uncertain. Scores near 50% are especially difficult to interpret: they could indicate that imitation was hard to hear, but they could also reflect guessing or lapses in attention. This clustering may conceal meaningful differences in perceptual ability and make the relationship harder to detect.
Our next step is to recruit more participants to determine whether the positive trend becomes clearer with additional data. We also plan to compare these listener judgments with direct acoustic measurements of how participants’ pronunciation changed. We presented these preliminary findings at LabPhon 20 in June 2026.