Children’s vocabulary acquisition starts with the challenge of speech segmentation: dividing continuous speech into words. One way of finding the boundaries between words is by tracking the transitional probabilities between syllables, a process named Statistical Learning (SL). These probabilities reflect the chance that one syllable is followed by a certain other syllable, and a low chance indicates a likely word boundary. SL ability varies across individuals, and this variation is related to differences in language attainment.
We investigated individual differences in SL and tested a new hypothesis that musical rhythmic ability relates to SL ability. We evaluated this in experimental studies with infants, neurotypical adults, adults with developmental dyslexia, and adults with attention-deficit/hyperactivity disorder (AD(H)D). Our adult studies additionally included measures of working memory and vocabulary, as they were previously hypothesized to relate to SL.
Across the studies comprising this dissertation, we found evidence that our participants successfully segmented unfamiliar continuous speech into words through SL. In infants, individual differences in SL indeed relate to rhythmic ability. However, we did not find this effect with adult participants. Results regarding adults’ working memory influencing SL were inconclusive, but we found small positive correlations between SL ability and vocabulary size. Finally, we found that individuals with dyslexia or AD(H)D did not differ from neurotypical participants in SL abilities. Rather, individual differences outweighed group effects.
These results imply that rhythmic ability supports language acquisition at an early age and highlight the importance of studying individual differences in language acquisition across development and neurodiversity.
