Testing the limits of accent adaptation in 15-month-old infants

Abstract

Studies examining accent adaptation in infancy repeatedly find it emerges only around 19 months (e.g., Best et al., 2009; Mulak et al., 2013). Yet standing in contrast to this body of work—and its support for theories that early phonological representations are underspecified and/or exemplar-based—is a single study finding 15-month-old infants demonstrate readily-observable accent adaptation in some contexts (van Heugten and Johnson, 2014). Here, we perform a much-needed replication and extend their task to probe the limitations of these early abilities. Following van Heugten and Johnson (2014), a group of 15-month-old Canadian English-learning infants heard a female Australian-accented talker read a story and were then tested for adaptation by their recognition of new words produced by the same talker. For another group of infants, we alter the identity of the talker reading the story (a new male talker) but not the test talker. Results from infants in the first group replicate the original study, supporting alternative views that abstract phonological representations are early-emerging. However, circumstances under which infants can leverage learning of accent information across talkers remain unclear. Further work in cross-talker adaptation has the potential to shed light on processes involved in accent adaptation, which remain poorly understood into adulthood.