Passive distributional learning of non-native vowel contrasts does not work for all listeners

Distributional learning studies investigating the acquisition of the Dutch contrast /ɑ/-/a:/ by nonnative Dutch learners have reported mixed results. The present study extends the literature by examining whether (i) naïve listeners are able to extract the distribution structure of a sequence of/ɑ//aː/ tokens drawn from a continuum,; and (ii) differential effects exist between naturalistic vs. exaggerated distributions. Australian-English listeners were randomly assigned to a flat, unimodal, bimodal or enhanced distribution training condition. Their performance was assessed using a categorisation task before and after training. Our findings showed that while categorisation accuracy was higher at post-test vs. pre-test (perhaps due to task learning), naïve learners did not show the predicted distributional learning effects: the bimodal and enhanced groups did not outperform the flat and unimodal groups. The results could be attributed to individual differences in the ability to sustain attention throughout the training phase, which may be necessary for highly variable speech sounds such as vowels.