​​Impact of Passive Second Language Exposure on Speech Segmentation and Word Learning

Poster Session A - Saturday, April 13, 2024, 2:30 – 4:30 pm EDT, Sheraton Hall ABC

Amiya S. Aggarwal¹ (aaggar7@uwo.ca), E. Rae Hoeppner¹, Laura J. Batterink¹; ¹University of Western Ontario

Previous studies have shown that statistical learning (SL), the process of becoming sensitive to patterns in the environment, occurs simply through exposure to input, without explicit training, effort, or intention. Studies using miniature artificial languages have shown that the ability to detect syllable co-occurrence patterns through SL supports wordform learning. However, it is not yet clear whether SL can scale up to support word learning in a fully natural language. In the current study, we tested whether passive, multi-week exposure to a novel spoken second language facilitates both word segmentation and the subsequent ability to map words to meaning in monolingual English-speaking adults. EEG responses to natural, continuous Italian speech were recorded before and after a three-week exposure period, during which they listened to either Italian podcasts (experimental group) or English podcasts (control group) for an hour a day. We quantified participants’ sensitivity to word boundaries by deriving the multivariate temporal response function (mTRF) to word onsets, and their word mapping abilities for Italian words and non-words using an associative learning task. We hypothesize that passive exposure will facilitate second language word learning, resulting in an enhanced EEG word onset response to continuous Italian speech and differential word mapping rates for the Italian words and non-words after the exposure period for participants in the experimental group. This result would provide evidence that SL facilitates the discovery of word boundaries in continuous speech in a fully natural language and further promotes object label mapping.

Topic Area: LANGUAGE: Other