The impact of talker variability and individual differences on word learning in adults

Abstract

Studies have shown that exposure to multiple talkers during learning is beneficial in a variety of spoken language tasks, such as learning speech sounds in a second language and learning novel words in a lab context. However, not all studies find the multiple talker benefit. Some studies have found that processing benefits from exposure to multiple talkers depend on factors related to the linguistic profile of the listeners and to the cognitive demands during learning (blocked versus randomized talkers). The current study examines whether scaffolding talker variability (blocked versus randomized) supports word-learning and whether individual differences in language ability, reading ability, and phonological working memory influence word-learning in adults. One hundred and fifty-two listeners were randomly assigned to four conditions: (1) single talker, (2) maximal scaffolding (blocked two-then-two talkers), (3) minimal scaffolding (blocked by four-talkers), and (4) multiple-talker mixed (four-talker randomized). Listeners completed a word-learning task in which they learned to associate nonsense words with novel objects, and were then tested on their ability to name the objects. Our results showed that listeners performed similarly across all talker conditions, with no evidence for a benefit of talker variability. In addition, participants with better language and phonological working memory skills performed better on the word-learning task. These results suggest that blocking and manipulating the presentation of talkers may not support word-learning in adults and that variability benefits may depend on a variety of experimental factors.