UC San Diego

Language learning processes are often examined by learning miniature languages in the lab, where controlling the learning of a full language is infeasible. Most work on learning words assumes these meaningful chunks of sound are composed of smaller, equally meaningful phonemes. However, recent evidence suggests crystallized phoneme categories do not fully explain how words are learned but instead to flexible representation of phonemes in terms of gradient similarity in word learning by adults and children.

I propose a shift in the understanding of the role of phoneme categories in word learning by presenting a set of three studies consisting of 6 experiments that challenge our understanding of phoneme representation. The first study assesses whether the adult word learning system is more flexible in its word representations than previously described and finds repeated examples that within certain boundaries such flexibility is easily achieved. The second study relies on eye-tracking measures, a more sensitive measure of recognition difficulty, to extend the behavioral results from the first and demonstrate that recognition is similarly rapid for phonemically-consistent and phonemically-variable words. It illuminates parts of the underlying responses of the adult word-learning system. The third study explores how flexibility changes across developmental time, testing 3-5-year-old children on their ability to learn flexible vocabularies as well as distinguish minimal pairs. This final study suggests that though children are capable of learning phonemically-variable words, the flexibility experienced by adults is due to extended time, practice, or experience with a language, and not available immediately to learners.

These findings contribute to the general field of word learning across three main avenues: first, by demonstrating incredible flexibility in sound category interpretation across adult learners of different linguistic backgrounds, counter to prevailing assumptions, and adding nuance to the research on bilingual advantages in word learning. Second, contributing new accuracy and time-course data on positional and segmental importance to speech sound flexibility in word learning, adding to the general understanding of word learning. And finally, by providing data to extrapolate a developmental time course for these abilities, contributing to our picture of spoken language development over a developmental time scale.