Tracking perception of the sounds of English
Warner, N., McQueen, J. M., & Cutler, A.
Tracking perception of the sounds of English. The Journal of the Acoustical Society of America, 135
, 2295-3006. doi:10.1121/1.4870486.
Twenty American English listeners identified gated fragments of all 2288 possible English within-word
and cross-word diphones, providing a total of 538 560 phoneme categorizations. The results show
orderly uptake of acoustic information in the signal and provide a view of where information about
segments occurs in time. Information locus depends on each speech sound’s identity and phonological
features. Affricates and diphthongs have highly localized information so that listeners’ perceptual
accuracy rises during a confined time range. Stops and sonorants have more distributed and gradually
appearing information. The identity and phonological features (e.g., vowel vs consonant) of the
neighboring segment also influences when acoustic information about a segment is available. Stressed
vowels are perceived significantly more accurately than unstressed vowels, but this effect is greater for
lax vowels than for tense vowels or diphthongs. The dataset charts the availability of perceptual cues to
segment identity across time for the full phoneme repertoire of English in all attested phonetic contexts.