
Listening in more than one language is not only about recognizing words; it is also about how much effort the brain invests to make sense of speech. This study examined multilingual speech perception and listening effort across spoken languages in quiet and noisy conditions. Participants completed a listening task with single words and sentences presented in quiet and in noise across their spoken languages. Listening effort was measured with pupillometry, allowing us to capture processing demands that may remain hidden when accuracy alone appears high.
Results showed that quiet listening was easier than listening in noise, with higher accuracy and reduced pupil dilation. Importantly, listening effort increased in less experienced or less proficient languages, even when listeners were able to understand the speech accurately. Greater language experience, including proficiency, exposure, and daily use, was associated with better speech perception and reduced effort, but the strength of this association depended on stimulus type and listening condition. For single words, language experience was especially important for reducing effort in noise. For sentences, language experience played a stronger role in quiet, particularly when sentences contained meaningful contextual cues. Contextual predictability supported understanding in some conditions, but it did not always reduce effort. In less experienced languages, highly predictable sentences sometimes elicited greater pupil dilation, suggesting that using context may itself require cognitive resources.
These findings show that multilingual listening effort is shaped by both the environment and the listener’s language experience, and that pupillometry can reveal hidden effort beyond accuracy.