Mothers do not enhance phonemic contrasts of Cantonese lexical tones with didactic intent in infant-directed speech

Abstract

Infant-directed speech (IDS) is suggested to facilitate infants' speech sound acquisition. The present study examined the perceptual and acoustic differences of Cantonese mothers' monosyllabic tone production in IDS and adult-directed speech (ADS) to determine whether mothers didactically exaggerated tonal contrasts of the six Cantonese tones when addressing infants. Nineteen Cantonese-speaking mothers produced 805 tones in ADS and 804 tones in IDS in a picture-naming task. Five Cantonese-native speakers identified the tones in the mothers' productions that were low-pass filtered to eliminate lexical information. Overall, no statistical significant difference of judgment accuracy was found between ADS and IDS, despite a general trend of lower accuracy in IDS. Acoustic analysis showed that syllable duration and fundamental frequencies (F0) of the six tones were significantly higher in IDS than in ADS. Differential changes of F0 slope were found for the six tones in IDS, with changes not always in the direction of enhancing the tonal targets or the tonal contrasts among the tones. Furthermore, acoustic differences were greater in less confused tone pairs than in easily confused tone pairs in IDS. Taken together, mothers acoustically modified tones in IDS but there was little evidence of mothers exaggerating tonal contrasts for facilitating purposes.