Child-directed speech enhances preschoolers’ speech perception in noise

Abstract

A dominant idea in research on infant-directed speech is that caregivers hyperarticulate their speech and exaggerate speech clarity as a means of facilitating language learning for infants. Much less is known about the acoustic-phonetic properties of speech to young children who are more amenable to perceptual testing. This study tests whether child-directed speech (CDS) is more intelligible by testing 4- to 6-year-old listeners in a speech-in-noise task in which they selected the corresponding picture on a touchscreen. Target words were two-way minimal pairs that contrasted in terms consonant voicing (e.g., “back” versus “pack"), vowel (e.g., “back” versus “beak"), or consonant and vowel (e.g., “back” versus “peak"). At a constant signal-to-noise ratio of -6 dB, speech recognition performance was significantly greater for CDS than for adult-directed speech (ADS). However, significant differences were also found across stimulus talkers, with some mothers providing a greater overall CDS benefit than others. Furthermore, CDS was related to greater enhancement of vowel contrasts in some mothers, and greater enhancement of consonant voicing in others. These perceptual results are discussed in terms of the acoustic-phonetic properties of these speech productions.