Abstract
Reaction-time (RT) investigations connected with the basic unit of speech perception are reinterpreted with a view to specifying the attentional constraints operating during speeded speech-unit identification tasks. Within the explanatory framework offered a test paradigm is presented that allows RT data to be interpreted as evidence for the segment or the syllable as the basic perceptual unit. Nonsense syllables of the structure /b/ or /d/ + /i/ or /u/ with devoiced, correctly voiced and incorrectly voiced initial plosives are played to subjects for either consonant or vowel identification. The effects of incorrect voicing on the RTs for consonant and vowel decisions are examined with a view to deciding whether or not the voicing information is an integral part of the perceptual unit. Results point to the possibility of segmental processing but indicate also that strategies differ across subjects.
Get full access to this article
View all access options for this article.
