Electronic Thesis and Dissertation Repository

Thesis Format

Integrated Article

Degree

Master of Science

Program

Psychology

Supervisor

Batterink, Laura

Abstract

Research in statistical learning using artificial languages has suggested that passive exposure to linguistic patterns can guide word segmentation and word mapping. It is unknown whether this type of unsupervised learning can scale up to support these aspects of learning in a natural language. Our study exposed monolingual English speakers to either English or Italian podcasts for an hour daily for 21 days, collecting behavioural and EEG data during both the pre- and post-exposure period. Our behavioural measure tested if L2 exposure would lead to improved word mapping for words with high phonotactic probabilities. Our EEG measure tested whether the L2 exposure would improve word segmentation, as indexed by the TRF to word onsets. Neither of these measures detected significant changes in L2 processing as a function of passive exposure. Accounting for additional baseline acoustic features in the future may clarify the effects of L2 exposure on moment-by-moment L2 processing.

Summary for Lay Audience

It is common for new language learners to take second language classes, learn a large number of vocabulary items, and still be unable to pick out the words in fluent, seemingly rapid-fire speech produced by a native speaker. However, as many language learning apps encourage, listening to a language in the form of podcasts or audiobooks can boost comprehension. The goal of this research was to test if passively listening to a second language is sufficient to help monolingual English speakers find word boundaries and whether familiarity with the sound patterns of the second language makes it easier to treat common words from that language as labels for new objects. We asked participants to listen to either English or Italian podcasts for an hour a day and compared their performance on two distinct measures, a Word Mapping task and a Continuous Listening task. The Word Mapping task compared how well participants could make word-object associations for three word types – common words (typical sound patterns of Italian), rare words (atypical sound patterns for Italian) and non-words (atypical sound patterns for Italian). The Continuous Listening task recorded participants’ neural response to word onsets in Italian speech. In contrast to our expectations, we did not find a measurable impact of the three weeks of exposure to a second language on either the Word Mapping or Continuous Listening task. However, on the word mapping measure, we found some evidence that all participants became more sensitive to sound patterns in the second language from session 1 to session 2. On the Continuous Listening task, it is possible that additional analyses which model other features of speech may reveal more subtle changes occurring during online speech processing. In the future, this line of research could uncover new ways to facilitate second language learning in adults by exploring methods other than traditional classroom teaching.

Download

Share

COinS