Perceptual processing time for brief CV syllables and steady‐state vowels was examined in a backward recognition making paradigm. Subjects were required to identify a 40‐msec target sound selected from either a consonant set (/ba/, /da/, /ga/), or a vowel set (/i/, /I/, /ε/). The target sound was followed by a different sound drawn from the same set after a variable silent interstimulus interval. The second sound interrupted the perceptual processing of the target stimulus at short interstimulus intervals. Recognition performance improved with increases in the silent interstimulus interval. One experiment examined processing time for consonants and vowels under binaural presentation. A second experiment compared consonant and vowel recognition under both binaural and dichotic presentation. The results indicated that (1) consonants require more processing time for recognition than vowels, and (2) binaural and dichotic presentation conditions produce differential effects on consonant and vowel recognition. These findings have several implications. First, speech perception is not immediate, but is the result of several distinct operations that are distributed over time. Second, speech perception involves various memorial processes and mechanisms which recode and store information at different stages of perceptual analysis. [Research supported in part by grant to Haskins Laboratories from NICHD and a PHS Bio‐Medical Sciences Grant to Indiana University.]