The role of speaking rates in High Variability Phonetic Training

Abstract

This study examined whether incorporating multiple speaking rates into High Variability Phonetic Training (HVPT) facilitates second language (L2) learners’ acquisition of a non-native phonological contrast. Native English-speaking naïve learners of Korean were trained to identify the Korean three-way stop contrast, which differs from English stop voicing contrasts in the relative weighting of voice onset time (VOT) and fundamental frequency (f0). Participants were assigned to a Multi-group, trained with stimuli from multiple talkers at three speaking rates (slow, normal, fast), or a Single group, trained with the same talkers only at the slow rate. HVPT sessions spanned three days, followed by everyday identification tests, a new talker generalization test, and a retention test one week later. Test stimuli included only slow-rate items. Acoustic analysis showed that faster speaking rates increased VOT overlap between lenis and aspirated stops while f0 distinctions remained relatively stable. Preliminary results showed that the Multi-group consistently outperformed the Single group in training, generalization, and retention. We discuss speaking rate as a potentially meaningful dimension of stimulus variability in HVPT and highlight its promise for future large-scale research to examine the extent to which incorporating multiple speaking rates promotes more robust L2 perceptual learning.