Detecting errors in American English /ɹ/ along a normalized acoustic threshold

Perception of speech sounds has been characterized as largely categorical, such that listeners experience a shift in perception from one sound to a different sound at a certain point on a continuum. Children with speech errors, however, have been found to have atypical category boundaries. Following an observation by Hagiwara [6], who noted that typical speakers show third formant (F3) values for /ɹ/ between 80% and 60% of their average vowel F3, Hamilton et al [7] found that this value replicates a categorical boundary for /ɹ/ for adult listeners: Productions above and below the 80% threshold sounded consistently “incorrect” or “correct”, but productions closest to the threshold were given more ambiguous judgments. In this study, we apply this notion of an F3 threshold to investigate whether children with residual sound errors (RSE) respond like expert adult listeners (speech-language pathologists) when presented with natural-speech stimuli along a continuum of F3 distances.